• Korean Journal of Agricultural and Forest Meteorology
• /
• v.21 no.1
• /
• pp.29-41
• /
• 2019

Terrestrial Gross Primary Production (GPP) is the largest global carbon flux, and forest ecosystems are important because of the ability to store much more significant amounts of carbon than other terrestrial ecosystems. There have been several attempts to estimate GPP using mechanism-based models. However, mechanism-based models including biological, chemical, and physical processes are limited due to a lack of flexibility in predicting non-stationary ecological processes, which are caused by a local and global change. Instead mechanism-free methods are strongly recommended to estimate nonlinear dynamics that occur in nature like GPP. Therefore, we used the mechanism-free machine learning techniques to estimate the daily GPP. In this study, support vector machine (SVM), random forest (RF) and artificial neural network (ANN) were used and compared with the traditional multiple linear regression model (LM). MODIS products and meteorological parameters from eddy covariance data were employed to train the machine learning and LM models from 2006 to 2013. GPP prediction models were compared with daily GPP from eddy covariance measurement in a deciduous forest in South Korea in 2014 and 2015. Statistical analysis including correlation coefficient (R), root mean square error (RMSE) and mean squared error (MSE) were used to evaluate the performance of models. In general, the models from machine-learning algorithms (R = 0.85 - 0.93, MSE = 1.00 - 2.05, p < 0.001) showed better performance than linear regression model (R = 0.82 - 0.92, MSE = 1.24 - 2.45, p < 0.001). These results provide insight into high predictability and the possibility of expansion through the use of the mechanism-free machine-learning models and remote sensing for predicting non-stationary ecological processes such as seasonal GPP.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.57 no.5
• /
• pp.101-111
• /
• 2015

Establishment of appropriate data in certain formats is essential for agricultural water cycle analysis, which involves complex interactions and uncertainties such as climate change, social & economic change, and watershed environmental change. The main objective of this study was to develop web-based Data processing and Model linkage Techniques for Agricultural Water-Resource analysis (AWR-DMT). The developed techniques consisted of database development, data processing technique, and model linkage technique. The watershed of this study was the upper Cheongmi stream and Geunsam-Ri. The database was constructed using MS SQL with data code, watershed characteristics, reservoir information, weather station information, meteorological data, processed data, hydrological data, and paddy field information. The AWR-DMT was developed using Python. Processing technique generated probable rainfall data using non-stationary frequency analysis and evapotranspiration data. Model linkage technique built input data for agricultural watershed models, such as the TANK and Agricultural Watershed Supply (AWS). This study might be considered to contribute to the development of intelligent watercycle analysis by developing data processing and model linkage techniques for agricultural water-resource analysis.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2019.05a
• /
• pp.17-17
• /
• 2019

IPCC (Intergovernmental Panel on Climate Change) 기후변화 전망보고서에 따르면 RCP 4.5 시나리오 기준, 21세기 전 지구 평균기온은 $2.5^{\circ}C$ 상승(한반도 $+3.0^{\circ}C$)하며, 전 지구 평균강수량은 4.1% 증가(한반도 +16.0%)할 것이라 전망하고 있다(기상청, 2012). 최근 기후변화와 기상이변에 따른 도심지 폭우특성이 변화하고 있음을 많은 연구결과에서 말해주고 있으며, 그 발생 빈도와 강도가 점차 증가하고 있는 추세이다. 특히, 서울시의 경우 인구와 재산이 밀집해 있어 폭우 발생에 의한 시민의 인명과 재산 피해 우려가 크다. 따라서 본 연구에서는 서울시를 대상으로 근미래(~2050년) 기후변화 하에서의 재현기간에 따른 확률강우량 변화 특성을 분석하여 비교 평가한 후 설계 강우량 산정에 활용하고자 하였다. 관측자료 기반 강수량의 변동 특성 분석과 Non-stationary GEV방법을 이용한 비정상성 빈도해석을 수행하였으며, 근미래 폭우특성 변화분석을 위하여 CMIP5 (Coupled Model Intercomparison Project 5)에 참여한 GCMs(General Circulation Models)을 활용한 강우빈도해석을 수행하였다. Mann-Kendall Test와 Quantile Regression을 통한 서울지점 여름철 강수량(June to September)과 기준강수량 초과 강수(30, 50, 80, 100mm/hr), 연간 10th 최대 강수량(Annual Top 10th Precipitation) 등을 분석한 결과 최근 증가 경향이 뚜렷하게 나타났으며, 비정상성 빈도해석에 의한 확률강우량 분석의 가능성과 신뢰성을 확인하였다. 또한 19-GCMs을 통하여 모의된 일(Daily) 단위 강수량자료를 비모수통계적 상세화(Nonparametric Temporal Downscaling) 기법을 적용하여 시간(Hourly) 강우로 다운스케일링하였으며, 서울시 미래 확률강우량에 대한 IDF 곡선(Intensity-Duration-Frequency Curve)을 작성하여 비교?분석한 결과 지속시간 1시간 강우에 대하여 재현기간 30년, 100년 조건에서 확률강우량이 약 4%~11% 수준에서 증가하고 있음을 확인하였다. 본 연구의 결과는 도심지 수공구조물의 설계빈도 영향을 진단하고, 근미래 발생가능한 확률강우량 변화에 따른 시간당 목표 강우량설정의 방법론을 제시하였다는데 의의가 있으며, 서울시의 방재성능목표 설정과 침수취약지역 해소를 위한 기후변화에 따른 수공구조물 설계 시 활용이 가능할 것으로 기대된다.

• Journal of Korea Water Resources Association
• /
• v.54 no.12
• /
• pp.1317-1328
• /
• 2021

This study aims to provide a predictive model based on climate models for simulating continuous daily rainfall sequences by combining bias-correction and spatio-temporal downscaling approaches. For these purposes, this study proposes a combined modeling system by applying conditional Copula and Multisite Non-stationary Hidden Markov Model (MNHMM). The GloSea5 system releases the monthly rainfall prediction on the same day every week, however, there are noticeable differences in the updated prediction. It was confirmed that the monthly rainfall forecasts are effectively updated with the use of the Copula-based bias-correction approach. More specifically, the proposed bias-correction approach was validated for the period from 1991 to 2010 under the LOOCV scheme. Several rainfall statistics, such as rainfall amounts, consecutive rainfall frequency, consecutive zero rainfall frequency, and wet days, are well reproduced, which is expected to be highly effective as input data of the hydrological model. The difference in spatial coherence between the observed and simulated rainfall sequences over the entire weather stations was estimated in the range of -0.02~0.10, and the interdependence between rainfall stations in the watershed was effectively reproduced. Therefore, it is expected that the hydrological response of the watershed will be more realistically simulated when used as input data for the hydrological model.

• Korean Journal of Remote Sensing
• /
• v.35 no.6_2
• /
• pp.1117-1132
• /
• 2019

Forest covers 30% of the Earth's land area and plays an important role in global carbon flux through its ability to store much greater amounts of carbon than other terrestrial ecosystems. The Gross Primary Production (GPP) represents the productivity of forest ecosystems according to climate change and its effect on the phenology, health, and carbon cycle. In this study, we estimated the daily GPP for a forest ecosystem using remote-sensed data from Moderate Resolution Imaging Spectroradiometer (MODIS) and machine learning algorithms Support Vector Machine (SVM). MODIS products were employed to train the SVM model from 75% to 80% data of the total study period and validated using eddy covariance measurement (EC) data at the six flux tower sites. We also compare the GPP derived from EC and MODIS (MYD17). The MODIS products made use of two data sets: one for Processed MODIS that included calculated by combined products (e.g., Vapor Pressure Deficit), another one for Unprocessed MODIS that used MODIS products without any combined calculation. Statistical analyses, including Pearson correlation coefficient (R), mean squared error (MSE), and root mean square error (RMSE) were used to evaluate the outcomes of the model. In general, the SVM model trained by the Unprocessed MODIS (R = 0.77 - 0.94, p < 0.001) derived from the multi-sites outperformed those trained at a single-site (R = 0.75 - 0.95, p < 0.001). These results show better performance trained by the data including various events and suggest the possibility of using remote-sensed data without complex processes to estimate GPP such as non-stationary ecological processes.