• The Korean Journal of Ecology
• /
• v.28 no.6
• /
• pp.353-363
• /
• 2005

We developed a local-scale ecophysiological model, Sim-CYCLE Fine by modifying Sim-CYCLE which was developed for a global scale simulation. Sim-CYCLE fine is able to simulate not only carbon fluxes but also plant growth with various time-steps from an hour to a month. The model outputs of $CO_2$ flux and biomass/LAI were highly reliable; we validated the model results with measurements from the eddy covariance technique and the harvest method ($R^2$ values of around 0.9 for both). The results suggested that the phonology and the seasonal dynamics of the $C_3/C4$ plant communities affected significantly the carbon fluxes and the plant growth during the plant growing season.

• Proceedings of The Korean Society of Agricultural and Forest Meteorology Conference
• /
• 2014.10a
• /
• pp.146-150
• /
• 2014

• Proceedings of The Korean Society of Agricultural and Forest Meteorology Conference
• /
• 2012.07a
• /
• pp.41-43
• /
• 2012

• Journal of Environmental Impact Assessment
• /
• v.26 no.2
• /
• pp.127-139
• /
• 2017

Measurements of net ecosystem exchange (NEE) of $CO_2$ based on the eddy covariance technique provide reasonable carbon balance estimates in response to local environmental conditions. In South Korea, the forest ecosystems cover approximately 64% of the total area, thereby strongly affecting regional carbon balances. Cultivated croplands that cover about 17% of the total area should also be considered when calculating the carbon balance of the country. In this study, our objectives were (a) to quantify the range and seasonal variation of NEE at forest ecosystems, including deciduous, coniferous, and mixed forests, and agricultural ecosystems, including rice paddies and a potato field, in South Korea and (b) to compare NEE at ten Fluxnet sites that have the same or similar ecosystems as found in South Korea. The results showed that the forest and agricultural ecosystems were carbon sinks. In Korea, NEE at the forest ecosystems varied between -31 and $-362gC/m^2/yr$, and NEE at the croplands ranged from -210 to $-248gC/m^2/growing$ season. At the deciduous forest, NEE reached low values in late spring, early summer, and early autumn, while at the coniferous forest, it reached low values in spring, early summer, and mid autumn. The young mixed forest was a much stronger carbon sink than the old-growth deciduous and coniferous forests. During each crop growing season, beet had the lowest NEE value within six crops, followed by wither wheat, maize, rice, potato, and soybean. These results will be useful for designing and applying management strategies for the reduction of $CO_2$ emissions.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.20 no.1
• /
• pp.5-17
• /
• 2018

The standardized eddy covariance flux data processing in KoFlux has been updated, and its database has been amended accordingly. KoFlux data users have not been informed properly regarding these changes and the likely impacts on their analyses. In this paper, we have documented how the current structure of data processing in KoFlux has been established through the changes and improvements to ensure transparency, reliability and usability of the KoFlux database. Due to increasing diversity and complexity of flux site instrumentation and organization, we have re-implemented the previously ignored or simplified procedures in data processing (e.g., frequency response correction, stationarity test), and added new methods for $CH_4$ flux gap-filling and $CO_2$ flux correction and partitioning. To evaluate the effects of the changes, we processed the data measured at a flat and homogeneous paddy field (i.e., HPK) and a deciduous forest in complex and heterogeneous topography (i.e., GDK), and quantified the differences. Based on the results from our overall assessment, it is confirmed that (1) the frequency response correction (HPK: 11~18% of biases for annually integrated values, GDK: 6~10%) and the stationarity test (HPK: 4~19% of biases for annually integrated values, GDK: 9~23%) are important for quality control and (2) the minimization of the missing data and the choice of the appropriate driver (rather than the choice of the gap-filling method) are important to reduce the uncertainty in gap-filled fluxes. These results suggest the future directions for the data processing technology development to ensure the continuity of the long-term KoFlux database.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.5 no.2
• /
• pp.61-69
• /
• 2003

Surface energy and $CO_2$ fluxes have been measured over a farmland in Haenam, Korea since July 2002. Eddy covariance technique, which is the only direct flux measurement method, was employed to quantitatively understand the interaction between the farmland ecosystem and the atmospheric boundary layer. Maintenance of eddy covariance system was the main concern during the early stage of measurement to minimize gaps and uncertainties in the dataset. Half-hourly averaged $CO_2$ concentration showed distinct diurnal and seasonal variations, which were closely related to changes in net ecosystem exchange (NEE) of $CO_2$. Daytime maximum $CO_2$ uptake was about -1.0 mg $CO_2$ m$^{-2}$ s$^{-1}$ in August whereas nighttime $CO_2$ release was up to 0.3 mg $CO_2$ m$^{-2}$ s$^{-1}$ during the summer. Both daytime $CO_2$ uptake and nighttime release decreased gradually with season. During the winter season, NEE was from near zero to 0.05 mg $CO_2$ m$^{-2}$ s$^{-1}$ . FK site was a moderate sink of atmospheric $CO_2$ until September with daily NEE of 22 g $CO_2$ m$^{-2}$ d$^{-1}$ . In October, it became a weak source of $CO_2$ with an emission rate of 2 g $CO_2$ m$^{-2}$ d$^{-1}$ . Long-term flux measurements will continue at FK site to further investigate inter-annual variability in NEE. to better understand these exchange mechanism and in-depth analysis, process-level field experiments and intensive short-term intercomparisons are also expected to be followed.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.16 no.3
• /
• pp.233-245
• /
• 2014

Nighttime correction of $CO_2$ flux is one of the most important and challenging tasks in eddy covariance measurements over a complex mountainous terrain. In this study, we have scrutinized the quality and the credibility of the $CO_2$ flux datasets which were produced by employing three different methods of nighttime correction, i.e., (1) friction velocity ($u^*$) correction, (2) light response curve (LRC) correction, and (3) advection-based van Gorsel (VG) correction. The whole year datasets used in our analysis were collected at the two KoFlux tower sites (i.e., GDK deciduous forest site at the upper hill and GCK coniferous forest site at the lower hill) located in the valley of Gwangneung National Arboretum in central Korea. The resultant magnitudes and patterns of ecosystem respiration ($R_E$), gross primary productivity (GPP), and net ecosystem exchange (NEE) of $CO_2$ showed marked differences among the datasets produced with three different correction methods, which were also site-specific. The examination from micrometeorological and ecological perspectives suggests that the major cause of some inconsistency seems to be associated with the advection of $CO_2$ along the sloping terrain and the inappropriate selection of the correction data that might have been already affected by advective flows. The comparison with the results from other studies indicated that the overall characteristics of the corrected $CO_2$ fluxes at GDK and GCK (except those with LRC correction) were well within the ranges reported in the literature for various ecosystems in East Asia in similar latitudes. However, our study also implies that there will be always a room for further improvement in the present datasets. Therefore, caution must be exercised for the data users in order to properly use the updated version of datasets through transparent, open and participatory communication with data producers.

• 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.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2009.05a
• /
• pp.1112-1116
• /
• 2009

The importance of securing water resources and their efficient management has attracted more attention recently due to water deficit. In water budget analysis, however, evapotranspiration (${\lambda}E$) has been approximated as the residual in the water balance equation or estimated from empirical equations and assumptions. To minimize the uncertainties in these estimates, it is necessary to directly measure ${\lambda}E$. In this study, using the eddy covariance technique, we have measured ${\lambda}E$ in a mixed forest in the Seolmacheon catchment in Korea from September 2007 to December 2008. During the growing season (May - July), ${\lambda}E$ in this mixed forest averaged about 2.2 mm $d^{-1}$, whereas it was on average 0.5 mm $d^{-1}$ during the non-growing season in winter. The annual total ${\lambda}E$ in 2008 was 581 mm $y^{-1}$, which is about 1/3 of the annual precipitation of 1997 mm. Despite the differences in the amount and frequency of precipitation, the accumulated ${\lambda}E$ during the overlapping period (i.e., September to December) for 2007 and 2008 was both ${\sim}$ 110 mm, showing virtually no difference. The omega factor, which is a measure of decoupling between forest and the atmosphere, was on average 0.5, indicating that the contributions of equilibrium ${\lambda}E$ and imposed ${\lambda}E$ to the total ${\lambda}E$ were about the same. The results suggest that ${\lambda}E$in this mixed forest was controlled by various factors such as net radiation, vapor pressure deficit, and canopy conductance. In this study, based on the direct measurements of ${\lambda}E$, we have quantified the relative contribution of ${\lambda}E$in the water balance of a mixed forest in the Seolmacheon catchment. In combination with runoff data, the information on ${\lambda}E$ would greatly enhance the reliability of water budget analysis in this catchment.

• Journal of Korea Water Resources Association
• /
• v.53 no.10
• /
• pp.889-902
• /
• 2020

This study was performed in Seolmacheon basin to evaluate the adequacy of GLDAS (Global Land Data Assimilation System) and GLEAM (Global Land Evaporation Amsterdam Model) evapotranspiration data. The verification data necessary for the evaluation of adequacy were calculated after processing the latent heat flux data produced in the Seolmacheon basin with the Koflux program. In order to gap-fill the empty period, alternative evapotranspiration was calculated in three ways: FAO-PM (Food and Agriculture Organization-Penman Monteith), MDV (Mean Diurnal Variation) and Kalman Filter. This study selected Kalman Filter method as the data gap-filling method because it showed the best Bias and RMSE among the three methods. The amount of GLDAS spatial evapotranspiration was calculated as Noah (version 2.1) with a time interval of 3 hours and a spatial resolution of 0.25°. The amount of GLEAM spatial evapotranspiration was calculated using GLEAM (version 3.1a). This study evaluated the spatial evapotranspiration of GLDAS and GLEAM as the evapotranspiration based on eddy covariance. As a result of evaluation, GLDAS spatial evapotranspiration showed better results than GLEAM. Accordingly, in this study, the GLDAS method was proposed as a method for calculating the amount of spatial evapotranspiration in the Seolmacheon basin.