• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.31-31
• /
• 2023

During December 2022, the northeast monsoon, which dominates the south and the Gulf of Thailand, had significant rainfall that impacted the lower southern region, causing flash floods, landslides, blustery winds, and the river exceeding its bank. The Golok River, located in Narathiwat, divides the border between Thailand and Malaysia was also affected by rainfall. In flood management, instruments for measuring precipitation and water level have become important for assessing and forecasting the trend of situations and areas of risk. However, such regions are international borders, so the installed measuring telemetry system cannot measure the rainfall and water level of the entire area. This study aims to predict 72 hours of water level and evaluate the situation as information to support the government in making water management decisions, publicizing them to relevant agencies, and warning citizens during crisis events. This research is applied to machine learning (ML) for water level prediction of the Golok River, Lan Tu Bridge area, Sungai Golok Subdistrict, Su-ngai Golok District, Narathiwat Province, which is one of the major monitored rivers. The eXtreme Gradient Boosting (XGBoost) algorithm, a tree-based ensemble machine learning algorithm, was exploited to predict hourly water levels through the R programming language. Model training and testing were carried out utilizing observed hourly rainfall from the STH010 station and hourly water level data from the X.119A station between 2020 and 2022 as main prediction inputs. Furthermore, this model applies hourly spatial rainfall forecasting data from Weather Research and Forecasting and Regional Ocean Model System models (WRF-ROMs) provided by Hydro-Informatics Institute (HII) as input, allowing the model to predict the hourly water level in the Golok River. The evaluation of the predicted performances using the statistical performance metrics, delivering an R-square of 0.96 can validate the results as robust forecasting outcomes. The result shows that the predicted water level at the X.119A telemetry station (Golok River) is in a steady decline, which relates to the input data of predicted 72-hour rainfall from WRF-ROMs having decreased. In short, the relationship between input and result can be used to evaluate flood situations. Here, the data is contributed to the Operational support to the Special Water Resources Management Operation Center in Southern Thailand for flood preparedness and response to make intelligent decisions on water management during crisis occurrences, as well as to be prepared and prevent loss and harm to citizens.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.7
• /
• pp.29-37
• /
• 2023

Due to climate change, interest in crop production and distribution is increasing, and attempts are being made to use bigdata and AI to predict production volume and control shipments and distribution stages. Prediction of agricultural product imports not only affects prices, but also controls shipments of farms and distributions of distribution companies, so it is important information for establishing marketing strategies. In this paper, we create an artificial intelligence prediction model that predicts the future import volume based on the wholesale market melon import volume data disclosed by the agricultural statistics information system and evaluate its accuracy. We create prediction models using three models: the Neural Prophet technique, the Ensembled Neural Prophet model, and the GRU model. As a result of evaluating the performance of the model by comparing two major indicators, MAE and RMSE, the Ensembled Neural Prophet model predicted the most accurately, and the GRU model also showed similar performance to the ensemble model. The model developed in this study is published on the web and used in the field for 1 year and 6 months, and is used to predict melon production in the near future and to establish marketing and distribution strategies.

• International Journal of Aeronautical and Space Sciences
• /
• v.11 no.4
• /
• pp.326-337
• /
• 2010

This work addresses problems regarding trajectory planning for unmanned aerial vehicle sensors. Such sensors are used for taking measurements of large nonlinear systems. The sensor investigations presented here entails methods for improving estimations and predictions of large nonlinear systems. Thoroughly understanding the global system state typically requires probabilistic state estimation. Thus, in order to meet this requirement, the goal is to find trajectories such that the measurements along each trajectory minimize the expected error of the predicted state of the system. The considerable nonlinearity of the dynamics governing these systems necessitates the use of computationally costly Monte-Carlo estimation techniques, which are needed to update the state distribution over time. This computational burden renders planning to be infeasible since the search process must calculate the covariance of the posterior state estimate for each candidate path. To resolve this challenge, this work proposes to replace the computationally intensive numerical prediction process with an approximate covariance dynamics model learned using a nonlinear time-series regression. The use of autoregressive time-series featuring a regularized least squares algorithm facilitates the learning of accurate and efficient parametric models. The learned covariance dynamics are demonstrated to outperform other approximation strategies, such as linearization and partial ensemble propagation, when used for trajectory optimization, in terms of accuracy and speed, with examples of simplified weather forecasting.

• Atmosphere
• /
• v.21 no.4
• /
• pp.391-404
• /
• 2011

This study investigates the impact of cumulus parameterization scheme (CPS) with different horizontal grid sizes on the simulation of the local heavy rainfall case over the Korean Peninsula. The Weather Research and Forecasting (WRF)-based real-time forecast system of the Joint Center for High-impact Weather and Climate Research (JHWC) is used. Three CPSs are used for sensitivity experiments: the BMJ (Betts-Miller-Janjic), GD (Grell-Devenyi ensemble), and KF (Kain-Fritsch) CPSs. The heavy rainfall case selected in this study is characterized by low-level jet and low-level transport of warm and moist air. In 27-km simulations (DM1), simulated precipitation is overestimated in the experiment with BMJ scheme, and it is underestimated with GD scheme. The experiment with KF scheme shows well-developed precipitation cells in the southern and the central region of the Korean Peninsula, which are similar to the observations. All schemes show wet bias and cold bias in the lower troposphere. The simulated rainfall in 27-km horizontal resolution has influence on rainfall forecast in 9-km horizontal resolution, so the statements on 27-km horizontal resolution can be applied to 9-km horizontal resolution. In the sensitivity experiments of CPS for DM3 (3-km resolution), the experiment with BMJ scheme shows better heavy rainfall forecast than the other experiments. The experiments with CPS in 3-km horizontal resolution improve rainfall forecasts compared to the experiments without CPS, especially in rainfall distribution. The experiments with CPS show lower LCL(Lifted Condensation Level) than those without CPS at the maximum rainfall point, and weaker vertical velocity is simulated in the experiments with CPS compared to the experiments without CPS. It means that CPS suppresses convective instability and influences mainly convective rainfall. Consequently, heavy rainfall simulation with BMJ CPS is better than the other CPSs, and even in 3-km horizontal resolution, CPS should be applied to control convective instability. This conclusion can be generalized by conducting more experiments for a variety of cases over the Korean Peninsula.

• Atmosphere
• /
• v.31 no.3
• /
• pp.327-340
• /
• 2021

In this study, the impact of soil moisture initialization in GloSea5, the operational climate prediction system of the Korea Meteorological Administration (KMA), has been investigated for the period of 1991~2010. To overcome the large uncertainties of soil moisture in the reanalysis, JRA55 reanalysis and CMAP precipitation were used as input of JULES land surface model and produced soil moisture initial field. Overall, both mean and variability were initialized drier and smaller than before, and the changes in the surface temperature and pressure in boreal summer and winter were examined using ensemble prediction data. More realistic soil moisture had a significant impact, especially within 2 months. The decreasing (increasing) soil moisture induced increases (decreases) of temperature and decreases (increases) of sea-level pressure in boreal summer and its impacts were maintained for 3~4 months. During the boreal winter, its effect was less significant than in boreal summer and maintained for about 2 months. On the other hand, the changes of surface temperature were more noticeable in the southern hemisphere, and the relationship between temperature and soil moisture was the same as the boreal summer. It has been noted that the impact of land initialization is more evident in the summer hemispheres, and this is expected to improve the simulation of summer heat wave in the KMA's operational climate prediction system.

• Korean Journal of Construction Engineering and Management
• /
• v.24 no.6
• /
• pp.66-73
• /
• 2023

The construction industry suffers losses due to failures in demand forecasting due to price fluctuations in construction raw materials, increased user costs due to project cost changes, and lack of forecasting system. Accordingly, it is necessary to improve the accuracy of construction raw material price forecasting. This study aims to predict the price of construction raw materials and verify applicability through the improvement of the Data Refactor technique. In order to improve the accuracy of price prediction of construction raw materials, the existing data refactor classification of low and high frequency and ARIMAX utilization method was improved to frequency-oriented and ARIMA method utilization, so that short-term (3 months in the future) six items such as construction raw materials lumber and cement were improved. ), mid-term (6 months in the future), and long-term (12 months in the future) price forecasts. As a result of the analysis, the predicted value based on the improved Data Refactor technique reduced the error and expanded the variability. Therefore, it is expected that the budget can be managed effectively by predicting the price of construction raw materials more accurately through the Data Refactor technique proposed in this study.

• Atmosphere
• /
• v.27 no.1
• /
• pp.105-118
• /
• 2017

This study assesses the performance of the Weather Research and Forecasting (WRF) model in reproducing regional climate over CORDEX-East Asia Phase 2 domain with different cumulus parameterization schemes [Kain-Fritch (KF), Betts-Miller-Janjic (BM), and Grell-Devenyi-Ensemble (GD)]. The model is integrated for 27 months from January 1979 to March 1981 and the initial and boundary conditions are derived from European Centre for Medium-Range Weather Forecast Interim Reanalysis (ERA-Interim). The WRF model reasonably reproduces the temperature and precipitation characteristics over East Asia, but the regional scale responses are very sensitive to cumulus parameterization schemes. In terms of mean bias, WRF model with BM scheme shows the best performance in terms of summer/winter mean precipitation as well as summer mean temperature throughout the North East Asia. In contrast, the seasonal mean precipitation is generally overestimated (underestimated) by KF (GD) scheme. In addition, the seasonal variation of the temperature and precipitation is well simulated by WRF model, but with an overestimation in summer precipitation derived from KF experiment and with an underestimation in wet season precipitation from BM and GD schemes. Also, the frequency distribution of daily precipitation derived from KF and BM experiments (GD experiment) is well reproduced, except for the overestimation (underestimation) in the intensity range above (less) then $2.5mm\;d^{-1}$. In the case of the amount of daily precipitation, all experiments tend to underestimate (overestimate) the amount of daily precipitation in the low-intensity range < $4mm\;d^{-1}$ (high-intensity range > $12mm\;d^{-1}$). This type of error is largest in the KF experiment.

• Atmosphere
• /
• v.31 no.5
• /
• pp.637-656
• /
• 2021

The United States has been known as the world's major producer of crops such as wheat, corn, and soybeans. Therefore, using meteorological long-term forecast data to project reliable crop yields in the United States is important for planning domestic food policies. The current study is part of an effort to improve the seasonal predictability of regional-scale precipitation across the United States for estimating crop production in the country. For the purpose, a dynamic downscaling method using Weather Research and Forecasting (WRF) model is utilized. The WRF simulation covers the crop-growing period (March to October) during 2000-2020. The initial and lateral boundary conditions of WRF are derived from the Pusan National University Coupled General Circulation Model (PNU CGCM), a participant model of Asia-Pacific Economic Cooperation Climate Center (APCC) Long-Term Multi-Model Ensemble Prediction System. For bias correction of downscaled daily precipitation, empirical quantile mapping (EQM) is applied. The downscaled data set without and with correction are called WRF_UC and WRF_C, respectively. In terms of mean precipitation, the EQM effectively reduces the wet biases over most of the United States and improves the spatial correlation coefficient with observation. The daily precipitation of WRF_C shows the better performance in terms of frequency and extreme precipitation intensity compared to WRF_UC. In addition, WRF_C shows a more reasonable performance in predicting drought frequency according to intensity than WRF_UC.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2021.06a
• /
• pp.392-392
• /
• 2021

자료동화(data assimilation) 기법은 관측 자료와 예측 모형의 정보를 동시에 활용, 모형의 상태량(state variables)이나 매개변수(model parameters)를 실시간으로 업데이트하는 Bayesian 필터링 이론에 근거한 방법으로, 최근 이를 활용한 수문 모의 정확도 향상 기술이 빠르게 발전하고 있다. 본 연구에서는 앙상블 자료동화의 정확성을 향상시키기 위한 세부 방법인 along-the-stream localization과 inflation 기법의 분포형 수문 모형에 대한 적용성을 대규모 지역 단위(regional-scale) 모의를 통해 검토한다. 분포형 수문모형과 자료동화 framework로는 WRF-Hydro(Weather Research and Forecasting Model Hydrological Modeling System)와 DART(Data Assimilation Research Testbed)를 각각 적용한다. WRF-Hydro는 미국의 전 대륙지역(CONUS; continental United States)에 대한 수문 모델링 framework인 National Water Model의 핵심엔진이고, DART는 미국 National Center for Atmospheric Research(NCAR) 연구소에서 개발한 범용 자료동화 도구이다. 본 연구에서는 지표수 수문모형의 자료동화를 위해 개발된 기법인 along-the-stream localization과 inflation 기법이 하도 추적에 미치는 영향을 분석한다. along-the stream localization 기법은 공간적 근접도 외에 하도의 수문학적 연관관계를 고려하는 localization 기법으로, 상대적으로 수문학적 상관도가 떨어지는 하도에 대한 과도한 자료동화를 줄여줄 수 있다. inflation 기법은 앙상블의 다양성을 증가시키는 기법으로, 칼만 필터(Kalman filter)에 의한 업데이트의 이전이나 이후 적용하여 앙상블 예측의 정확도를 추가적으로 향상시킬 수 있다. 본 고에서는 앙상블 자료동화 기법을 지표수 수문 모의에 적용할 경우 남아 있는 난제와 적용 가능한 방법에 대해 중점적으로 논의한다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.63-63
• /
• 2022

기상 예보자료는 발생 가능한 재난의 예방 및 대비 차원에서 매우 중요한 자료로 활용되고 있다. 우리나라 기상청에서는 동네예보를 통해 5km 공간해상도의 1시간 간격 초단기예보와, 6시간 간격 정량강우예보(Quantitative Precipitation Forecast, QPF)의 단기예보 정보를 제공하고 있다. 그러나 이와 같은 예보자료는 강우량의 시·공간변화가 큰 집중호우와 같은 기상자료를 활용한 수문학적인 해석에는 한계가 있다. 예보자료를 수문학에 활용하기 위한 시·공간적 해상도 개선뿐만 아니라 방대한 기상 및 기후 자료의 예측성능을 개선하기 위한 다양한 연구가 진행되고 있다. 본 연구에서는 기상청이 제공하는 지역 앙상블 예측 시스템(Local ENsemble prediction System, LENS)와 종관기상관측시스템(ASOS) 및 방재기상관측시스템(AWS) 관측 데이터 및 동네예보에 기계학습 방법을 적용하여 수문학적 정량적 강수량 예측(Hydrological Quantitative Precipitation Forecast, HQPF) 정보를 생산하였다. 전처리 과정을 통해 모든 데이터의 시간해상도와 공간해상도를 동일한 해상도로 변환하였으며, 예측 변수의 인자 분석을 통해 기계학습의 예측 변수를 도출하였다. 기계학습 방법으로는 처리속도와 확장성을 고려하여 XGBoost(eXtreme Gradient Boosting) 방식을 적용하였으며, 집중호우에서의 예측정확도를 높이기 위해 확률매칭(PM) 방식을 적용하였다. 생산된 HQPF의 성능을 평가하기 위해 2020년에 발생한 14건의 호우 사상을 대상으로 태풍형과 비태풍형으로 구분하여 검증을 수행하였다.