• Proceedings of the Korea Water Resources Association Conference
• /
• 2008.05a
• /
• pp.2207-2211
• /
• 2008

한반도 및 동아시아의 여름철은 장마와 태풍으로 인한 집중호우의 발생으로 많은 피해를 입는다. 따라서 여름철에 나타나는 이러한 집중호우가 나타나는 지역, 시기, 기간, 그리고 강수량 등을 예측하는 것은 매우 중요하다. 특히, 효율적인 수자원 관리를 위하여 이러한 예측은 매우 중요한데, 단기적으로 정확하고 신속하게 강수를 예측하는 것도 중요하지만, 장기적으로 계절 강수, 특히 여름철의 장마 또는 우기의 시기와 강수량과 태풍 발생의 시기 등을 미리 예측하여 이에 따른 집중 호우의 발생 지역, 기간, 강수량을 예측하여 사전에 대비하는 것도 매우 중요하다. 특히, 최근에는 6,7월 장마에 의한 집중 호우의 영향보다도 8월에 강수량이 높아지고 있는 경향을 보이므로 강수량의 장기적 경향의 파악이 매우 중요하다. 장기 기후를 예측하는 데는 과거 자료를 이용한 통계 방법도 유용하지만 최근에는 AOGCM (Atmospheric Oceanic General Circulation Model)을 이용한 연구가 활발하게 이루어지고 있다. 하지만 강수와 같이 지역적으로 나타나는 현상은 저해상도의 AOGCM으로는 유용한 정보를 제공하기가 어려움이 따른다. 따라서 본 연구에서는 전구를 삼각형으로 된 20면체로 격자화 시켜 모든 격자의 크기가 거의 동일하고, 해상도 조절이 가능한 Geodesic 격자를 활용한 GME 모델을 사용하였다. GME 모델은 icosahedral-hexagonal grid 격자 체계를 가진 독일 기상청(Deutscher Wetterdient)에서 현업으로 사용 중인 모델이다. 본 연구에서는 수직/수평 해상도를 40km/40layers로 하여 GME 모델을 수행하였으며, 일간격의 장기 기후 자료를 생산하였다. 사용된 초기자료로는 ECMWF (European Centre for Medium Range Weather Forecasts) 자료이며, 경계 자료로는 ERA Climatology의 최근 30년간의 SST (Sea Surface Temperature) 평균 자료를 이용하여 규준 실험(Control Run), 즉, climatology 자료를 생산하였으며, persistent SST 아노말리와 ERA Climatology의 최근 30년간의 SST 자료를 이용하여 내삽 과정을 거친 SST forcing을 주어서 예측 실험(Prediction Run)을 통하여 모의 자료를 생산하였다. 특히, 규준 실험에서는 수치 모델이 가지는 불확실성을 줄이고 예보 정확도를 향상시키기 위하여 각각의 실험은 초기자료를 달리한 앙상블 모의실험을 수행하였다. 장기 모의 3개월을 위하여 모의 기간 1달 전부터 모의를 수행하여, 첫 1달은 모델의 spin-up 시간으로 분석에서 제외 하였다. 생산된 Climatology 자료와 Prediction 자료를 비교하여 아노말리와 Category 분석을 실시하여 한반도 및 동아시아 지역의 강수(Precipitation)를 중심으로 기압장(Pressure), 온도(2m Temperature) 위주로 분석하였다. 이러한 예측된 매 계절의 전망 자료 중에서도 수자원 분야에서 관심이 집중되는 여름철에 초점을 맞추어 실제 관측 자료와 비교하여 GME 모델의 계절 모의 예측성 성능을 분석하여 평가하고 다가올 여름철의 강수량의 장기 변화를 모의하고자 하였다.

• Ecology and Resilient Infrastructure
• /
• v.5 no.3
• /
• pp.125-133
• /
• 2018

Artificial neural network (ANN) is a powerful model to predict time series data and have been frequently adopted to predict groundwater level (GWL). Many researchers have also tried to improve the performance of ANN prediction for GWL in many ways. Dummies are usually used in ANN as input to reflect the seasonal effect on predicted results, which is necessary for improving the predicting performance of ANN. In this study, the effect of Dummy on the prediction performance was analyzed qualitatively and quantitatively using several graphical methods, correlation coefficient and performance index. It was observed that results predicted using dummies for ANN model indicated worse performance than those without dummies.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.25 no.2
• /
• pp.80-98
• /
• 2023

This study explores the effectiveness of various data preprocessing algorithms for improving subseasonal to seasonal (S2S) climate predictions from six climate forecast models and their Multi-Model Ensemble (MME) using a deep learning-based postprocessing model. A pipeline of data transformation algorithms was constructed to convert raw S2S prediction data into the training data processed with several statistical distribution. A dimensionality reduction algorithm for selecting features through rankings of correlation coefficients between the observed and the input data. The training model in the study was designed with TimeDistributed wrapper applied to all convolutional layers of U-Net: The TimeDistributed wrapper allows a U-Net convolutional layer to be directly applied to 5-dimensional time series data while maintaining the time axis of data, but every input should be at least 3D in U-Net. We found that Robust and Standard transformation algorithms are most suitable for improving S2S predictions. The dimensionality reduction based on feature selections did not significantly improve predictions of daily precipitation for six climate models and even worsened predictions of daily maximum and minimum temperatures. While deep learning-based postprocessing was also improved MME S2S precipitation predictions, it did not have a significant effect on temperature predictions, particularly for the lead time of weeks 1 and 2. Further research is needed to develop an optimal deep learning model for improving S2S temperature predictions by testing various models and parameters.

• Korean Journal of Environmental Agriculture
• /
• v.32 no.4
• /
• pp.294-303
• /
• 2013

BACKGROUND: It is well known that rice-fields can provide excellent foraging places for birds including seasonal migrants, wintering, and breeding and hence the high biodiversity of rice-fields may be expected. However, how environmental change including climate-changes on life-history and population dynamics in birds on rice-fields has not been fully understood. In order to investigate how climate-change affects population migratory patterns and migration timing, I modeled a population dynamics of birds in rice-fields over a whole year. METHODS AND RESULTS: I applied the Lotka-Volterra equation to model the population dynamics of birds that have been foraging/visiting rice-fields in Korea. The simple model involves the number of interspecific individuals and temperature, and the model parameters are periodic in time as the biological activities related to the migration, wintering and reproduction are seasonal. As results, firstly there was a positive relationship between the variation of seasonal population sizes and temperature change. Secondly, the reduced lengths of season were negatively related to the population size. Overall, the effects of the difference of lengths of season on seasonal population dynamics were higher than the effects of seasonal temperature change. CONCLUSION(S): Climate change can alter population dynamics of birds in rice-fields and hence the variation may affect the fitness, such as reproduction, survival and migration. The unstable balances of population dynamics in birds using paddy rice field as affected by climate change can reduce the population growth and species diversity in rice fields. The results suggest that the agricultural production is partly affected by the unstable balance of population in birds using rice-fields.

• The Journal of Engineering Geology
• /
• v.33 no.3
• /
• pp.403-411
• /
• 2023

The possibility of utilizing radiant solar energy as a renewable energy resource in Tajikistan was investigated by assessing solar irradiance using XGBoost algorithm. Through training, validation, and testing, the seasonality of solar irradiance was clear in both actual and predicted values. Calculation of hourly values of solar irradiance on 1 July 2016, 2017, 2018, and 2019 indicated maximum actual and predicted values of 1,005 and 1,009 W/m², 939 and 997 W/m², 1,022 and 1,012 W/m², 1,055 and 1,019 W/m², respectively, with actual and predicted values being within 0.4~5.8%. XGBoost is thus a useful tool in predicting solar irradiance in Tajikistan and evaluating the possibility of utilizing radiant solar energy.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.25 no.8A
• /
• pp.1153-1159
• /
• 2000

This paper applies a seasonal ARIMA model to the timely forecasting in a line utilization and its confidence interval on the base of the past data of the lido utilization that QoS of the network is greatly influenced by and proposes the prediction algorithm of threshold violation in line utilization using the seasonal ARIMA model. We can predict the time of threshold violation in line utilization and provide the confidence based on probability. Also, we have evaluated the validity of the proposed model and estimated the value of a proper threshold and a detection probability, it thus appears that we have maximized the performance of this algorithm.

• Journal of Korea Water Resources Association
• /
• v.52 no.7
• /
• pp.475-482
• /
• 2019

Subseasonal-to-Seasonal (S2S) prediction information which have 2 weeks to 2 months lead time are expected to be used through many parts of industry fields, but utilizability is not reached to expectation because of lower predictability than weather forecast and mid- /long-term forecast. In this study, we used multi-layer perceptron (MLP) which is one of machine learning technique that was built for regression training in order to improve predictability of S2S precipitation data at South Korea through post-processing. Hindcast information of ECMWF was used for MLP training and the original data were compared with trained outputs based on dichotomous forecast technique. As a result, Bias score, accuracy, and Critical Success Index (CSI) of trained output were improved on average by 59.7%, 124.3% and 88.5%, respectively. Probability of detection (POD) score was decreased on average by 9.5% and the reason was analyzed that ECMWF's model excessively predicted precipitation days. In this study, we confirmed that predictability of ECMWF's S2S information can be improved by post-processing using MLP even the predictability of original data was low. The results of this study can be used to increase the capability of S2S information in water resource and agricultural fields.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2016.05a
• /
• pp.236-236
• /
• 2016

현재 댐운영 계획 수립 시 매월 유지해야 하는 저수량의 범위를 나타낸 기준수위가 사용되고 있으며 매년 홍수기 말에 현재의 수문 상황과 장래의 전망을 통한 시기별 연간, 월간 댐운영 계획을 수립하고 있다. 물관리의 이수측면에서 댐수위 운영계획 수립과 홍수기 운영목표 수위를 결정하는데 활용하기 위해서는 계절단위, 연단위의 기상정보가 필요하다. 본 연구에서는 기상청에서 운영하고 제공하는 전지구 계절예측시스템 GloSea5(Global Seasonal forecasting system version 5)자료를 활용하여 금강유역에 적용하고자 하였다. GloSea5는 전지구계절예측시스템으로 대기(UM), 지면(JULES), 해양(NEMO), 해빙(CICE)모델이 서로 결합되어 하나의 시스템으로 구성되어 있으며 공간 수평해상도는 N216($0.83^{\circ}{\times}0.56^{\circ}$)으로 중위도에서 약60km이다. Hindcast자료는 유럽중기예보센터(ECMWF)에서 생산된 ERA-Interim 재분석장을 대기 모델의 초기장으로 사용하며 기간은 1996~2009년의 총 14년이다. 예보자료의 검증은 예보의 질을 결정하는 과정으로 Brier Skill Score (BSS), Reliability Diagrams, Relative Operating, Characteristics (ROC)등을 통해 정확성과 오차에 의한 예보의 성능을 검증하였다. 또한 Glosea5의 통계적 상세화를 수행하여 다양한 변수가 갖는 계통적인 지역 오차를 보정함으로써 자료의 신뢰도를 향상시키고자 하였으며 이는 이후 수문모델과의 연계 시 보다 정확하고 효율적인 댐운영에 활용할 수 있는 기후예측정보를 제공할 수 있을 것으로 판단된다.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2013.10a
• /
• pp.133-135
• /
• 2013

장래의 해상교통량에 대한 정확한 예측은 항로설계 및 해상교통의 안전성 평가 측면에서 중요한 요소이다. 본 연구는 신뢰성 있는 해상교통량을 추정하기 위해 시계열 모델의 지수평활법과 ARIMA 모형을 이용하여 모형의 식별 및 진단 방안을 제시하였다. 제시된 방법의 효과를 검증하기 위하여 주요항만인 부산항, 광양항, 인천항, 평택항의 해상교통량을 예측하였다. 그 결과로 부산항은 ARIMA 모형, 광양항은 Winters 승법 모형, 인천항은 단순계절 모형, 평택항은 ARIMA 모형이 더 적합한 모형으로 알 수 있었으며, 각 항만별 계절에 따라 월별 교통량의 차이를 보이는 것으로 분석되었다. 본 연구 결과는 향후 항로 및 항만설계 또는 해상교통 안전성 평가에 보다 신뢰성 있는 추정치를 제공할 수 있을 것으로 보인다.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.18 no.4
• /
• pp.253-263
• /
• 2016

Chrysanthemum production would benefit from crop growth simulations, which would support decision-making in crop management. Chrysanthemum is a typical short day plant of which floral initiation and development is sensitive to photoperiod. We developed a model to predict phenological development and leaf appearance of chrysanthemum (cv. Baekseon) using daylength (including civil twilight period), air temperature, and management options like light interruption and ethylene treatment as predictor variables. Chrysanthemum development stage (DVS) was divided into juvenile (DVS=1.0), juvenile to budding (DVS=1.33), and budding to flowering (DVS=2.0) phases for which different strategies and variables were used to predict the development toward the end of each phenophase. The juvenile phase was assumed to be completed at a certain leaf number which was estimated as 15.5 and increased by ethylene application to the mother plant before cutting and the transplanted plant after cutting. After juvenile phase, development rate (DVR) before budding and flowering were calculated from temperature and day length response functions, and budding and flowering were completed when the integrated DVR reached 1.33 and 2.0, respectively. In addition the model assumed that leaf appearance terminates just before budding. This model predicted budding date, flowering date, and leaf appearance with acceptable accuracy and precision not only for the calibration data set but also for the validation data set which are independent of the calibration data set.