• Proceedings of the Korean Society of Crop Science Conference
• /
• 2004.04a
• /
• pp.104-105
• /
• 2004

[ $\bigcirc$ ] In the phenology model of ORYZA2000, the effect of photoperiod on the developmental rate was a little ignored because most crop parameters were measured with IRRI varieties which are insensitive to photoperiod, therefore it is very difficult to apply this phenology model directly to Korean varieties which are usually sensitive to photoperiod. $\bigcirc$ After introducing PPFAC and PPSE to improve the phenology model, the precision of heading date prediction was improved but not satisfied. $\bigcirc$ In the growth simulation using data from several regions, yield tended to be overestimated under high nitrogen applicated condition. $\bigcirc$ The precision of yield was much improved by introducing nitrogen use efficiency, but still different between regions because of different soil fertility or property of irrigation water between regions

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.44 no.2
• /
• pp.149-153
• /
• 1999

Prediction of rice developmental stage is necessary for proper crop management and a prerequisite for growth simulation as well. The objectives of the present study were to find out the relationship between the plastochrone index(PI) and the developmental index(DVI) estimated by non-parametric phenology model which simulates the duration from seedling emergence(DVI=0) to heading(DVI=l) by employing daily mean air temperature and daylength as predictor variables, and to confirm the correspondency of developmental indice to panicle developmental stages based on this relationship. Four japonica rice cultivars, Kwanakbyeo, Sangpungbyeo, Dongjinbyeo, and Palgumbyeo which range from very early to very late in maturity, were grown by sowing directly in dry paddy field five times at an interval of two weeks. Data for seedling emergence, leaf appearance, differentiation stage of primary rachis branch and heading were collected. The non-parametric phenology model predicted well the duration from seedling emergence to heading with errors of less than three days in all sowings and cultivars. PI was calculated for every leaf appearance and related to the developmental index estimated for corresponding PI. The stepwise polynomial analysis produced highly significant square-rooted cubic or biquadratic equations depending on cultivars, and highly significant square-rooted biquadratic equation for pooled data across cultivars without any considerable reduction in accuracy compared to that for each cultivar. To confirm the applicability of this equation in predicting the panicle developmental stage, DVI at differentiation stage of primary rachis branch primordium was calculated by substituting PI with 82 corresponding to this stage, and the duration reaching this DVI from seedling emergence was estimated. The estimated duration revealed a good agreement with that observed in all sowings and cultivars. The deviations between the estimated and the observed were not greater than three days, and significant difference in accuracy was not found for predicting this developmental stage between those equations derived for each cultivar and for pooled data across all cultivars tested.

• Korean Journal of Remote Sensing
• /
• v.40 no.1
• /
• pp.57-69
• /
• 2024

Crop classification plays a vitalrole in monitoring agricultural landscapes and enhancing food production. In this study, we explore the effectiveness of Long Short-Term Memory (LSTM) models for crop classification, focusing on distinguishing between apple and rice crops. The aim wasto overcome the challenges associatedwith finding phenology-based classification thresholds by utilizing LSTM to capture the entire Normalized Difference Vegetation Index (NDVI)trend. Our methodology involvestraining the LSTM model using a reference site and applying it to three separate three test sites. Firstly, we generated 25 NDVI imagesfrom the Sentinel-2A data. Aftersegmenting study areas, we calculated the mean NDVI values for each segment. For the reference area, employed a training approach utilizing the NDVI trend line. This trend line served as the basis for training our crop classification model. Following the training phase, we applied the trained model to three separate test sites. The results demonstrated a high overall accuracy of 0.92 and a kappa coefficient of 0.85 for the reference site. The overall accuracies for the test sites were also favorable, ranging from 0.88 to 0.92, indicating successful classification outcomes. We also found that certain phenological metrics can be less effective in crop classification therefore limitations of relying solely on phenological map thresholds and emphasizes the challenges in detecting phenology in real-time, particularly in the early stages of crops. Our study demonstrates the potential of LSTM models in crop classification tasks, showcasing their ability to capture temporal dependencies and analyze timeseriesremote sensing data.While limitations exist in capturing specific phenological events, the integration of alternative approaches holds promise for enhancing classification accuracy. By leveraging advanced techniques and considering the specific challenges of agricultural landscapes, we can continue to refine crop classification models and support agricultural management practices.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.15 no.1
• /
• pp.40-49
• /
• 2013

Current service system of the Korea Meteorological Administration (KMA) for blooming date forecasting in spring depends on regression equations derived from long term observations in both temperature and phenology at a given station. This regression based system does not allow a timely correction or update of forecasts that are highly sensitive to fluctuating weather conditions. Furthermore, the system cannot afford plant responses to climate extremes which were not observed before. Most of all, this method may not be applicable to locations other than that which the regression equations were derived from. This note suggests a way to replace the location restricted regression equations with a thermal time based phenology model to complement the KMA blooming forecast system. Necessary parameters such as reference temperature, chilling requirement and heating requirement were derived from phenology data for forsythia, azaleas and Japanese cherry at 29 KMA stations for the 1951-1980 period to optimize spring phenology prediction model for each species. Best fit models for each species were used to predict blooming dates and the results were compared with the observed dates to produce a correction grid across the whole nation. The models were driven by the KMA's daily temperature data at a 5km grid spacing and subsequently adjusted by the correction grid to produce the blooming date maps. Validation with the 1971-2012 period data showed the RMSE of 2-3 days for Japanese cherry, showing a feasibility of operational service; whereas higher RMSE values were observed with forsythia and azaleas.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.14 no.1
• /
• pp.11-18
• /
• 2012

Daily temperature data produced by harmonic analysis of monthly climate summary have been used as an input to plant phenology model. This study was carried out to evaluate the performance of the harmonic based daily temperature data in prediction of major phenological developments and to apply the results in improving decision support for agricultural production in relation to the climate change scenarios. Daily maximum and minimum temperature data for a climatological normal year (Jan. 1 to Dec. 31, 1971-2000) were produced by harmonic analysis of the monthly climate means for Seoul weather station. The data were used as inputs to a thermal time - based phenology model to predict dormancy, budburst, and flowering of Japanese cherry in Seoul. Daily temperature measurements at Seoul station from 1971 to 2000 were used to run the same model and the results were compared with the harmonic data case. Leaving no information on annual variation aside, the harmonic based simulation showed 25 days earlier release from endodormancy, 57 days longer period for maximum cold tolerance, delayed budburst and flowering by 14 and 13 days, respectively, compared with the simulation based on the observed data. As an alternative to the harmonic data, 30 years daily temperature data were generated by a stochastic process (SIMMETEO + WGEN) using climatic summary of Seoul station for 1971-2000. When these data were used to simulate major phenology of Japanese cherry for 30 years, deviations from the results using observed data were much less than the harmonic data case: 6 days earlier dormancy release, 10 days reduction in maximum cold tolerance period, only 3 and 2 days delay in budburst and flowering, respectively. Inter-annual variation in phenological developments was also in accordance with the observed data. If stochastically generated temperature data could be used in agroclimatic mapping and zoning, more reliable and practical aids will be available to climate change adaptation policy or decision makers.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.24 no.3
• /
• pp.155-163
• /
• 2022

The optimization of long-range ensemble climate prediction for rice phenology model with advanced bias correction method is conducted. The daily long-range forecast(6-month) of mean/ minimum/maximum temperature and observation of January to October during 1991-2021 is collected for rice phenology prediction. In this study, the concept of "buffer period" is newly introduced to reduce the problem after bias correction by quantile mapping with constructing the transfer function by month, which evokes the discontinuity at the borders of each month. The four experiments with different lengths of buffer periods(5, 10, 15, 20 days) are implemented, and the best combinations of buffer periods are selected per month and variable. As a result, it is found that root mean square error(RMSE) of temperatures decreases in the range of 4.51 to 15.37%. Furthermore, this improvement of climatic variables quality is linked to the performance of the rice phenology model, thereby reducing RMSE in every rice phenology step at more than 75~100% of Automated Synoptic Observing System stations. Our results indicate the possibility and added values of interdisciplinary study between atmospheric and agriculture sciences.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.12 no.2
• /
• pp.95-106
• /
• 2010

The performance of Community Land Model version 3.5 - Dynamic Global Vegetation Model (CLM-DGVM) was evaluated through a comparison with the observation over temperate deciduous forest in Gwangneung, Korea. Influence of plant phenology, composition of plant functional type, and climate variability on carbon exchanges was also examined through sensitivity test. To get equilibrium carbon storage, the model was run for 400 years driven by the observed atmospheric data at the deciduous forest of the year 2006. We run the model for 2006 with the equilibrium carbon storage at Gwangneung forest and compared the model output with the observation. A comparison of leaf area index (LAI) between the model and observation indicated that the simulated phenology poorly represented the timing of budburst, leaf-fall, and evolution of LAI. Senescence of the phenology was delayed about four weeks and the simulated maximum LAI (of 5.8 $m^2$ $m^{-2}$) was greater than the observed value (of 4.5 $m^2$ $m^{-2}$). The overestimated LAI contributed to overestimation of both gross primary productivity (GPP) and ecosystem respiration $(R_e)$ through increased photosynthesis and foliar autotropic respiration $(R_a)$, respectively. Despite the discrepancy between the simulated and observed LAI, the simulated tree carbon storage amounts were comparable with the reported values at the site. Change in plant phenology from the simulated to the observed reduced more than six weeks of the plant growth period, resulting in the decreased amount of GPP and $R_e$. These values, however, were still higher (~10% of GPP and 40% of $R_e$) than the observed values. The effect of change in plant functional type composition (from dominant temperate deciduous forest to the coexistence of temperate deciduous and needle leaf forests) on the estimated amount of GPP and $R_e$ was marginal. The influence of climate variability on carbon storage amounts was not significant. The simulated inter-annual variation of GPP and $R_e$ from 1994 to 2003 depended on annual mean air temperature and total radiation but not on precipitation. Other deficiencies of CLM3.5-DGVM have been discussed.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.16 no.4
• /
• pp.304-308
• /
• 2014

Accurate prediction of peak bloom dates (PBD) of flowering cherry trees is critical for organizing local cherry festivals and other associated cultural and economic activities. A two-step phenology model is commonly used for predicting flowering time depending on local temperatures as a result of two consecutive steps followed by chill and heat accumulations. However, an extensive computation requirement for parameter estimation has been a limitation for its practical use. We propose a sequential parameterization method by exploiting previously unused records of development stages. With an extra constraint formed by heat accumulation between two intervening stages, each parameter can then be solved sequentially in much shorter time than the brute-force method. The result was found to be almost identical to the previous solution known for cherry trees (Prunus ${\times}$ yedoensis) in the Tidal Basin, Washington D.C.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.24 no.2
• /
• pp.78-82
• /
• 2022

Phenology determines the timing of crop development, and the timing of phenological events is strongly influenced by the temperature during the growing season. In process-based model, leaf area is simulated dynamically by coupling of morphology and phenology module. Therefore, the prediction of leaf appearance rate and final leaf number affects the performance of whole crop model. The dataset for the model equation was collected from SPA R chambers with five different temperature treatments. Beta distribution function (proposed by Yan and Hunt (1999)) was used for describing the leaf appearance rate as a function of temperature. The optimum temperature and the critical value were estimated to be 26.0℃ and 35.3℃, respectively. For evaluation of the model, the accumulated number of onion leaves observed in a temperature gradient chamber was compared with model estimates. The model estimate is the result of accumulating the daily increase in the number of onion leaves obtained by inputting the daily mean temperature during the growing season into the temperature model. In this study, the coefficient of determination (R²) and RMSE value of the model were 0.95 and 0.89, respectively.

• Proceedings of the Korean Society of Applied Entomolohy Conference
• /
• 2000.10a
• /
• pp.36-36
• /
• 2000