• Proceedings of The Korean Society of Agricultural and Forest Meteorology Conference
• /
• 2018.06a
• /
• pp.129-130
• /
• 2018

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.10
• /
• pp.1-9
• /
• 2024

In this paper, we propose a comparative analysis to evaluate the impact of activation functions and attention mechanisms on the performance of time-series models for Mars meteorological data. Mars meteorological data are nonlinear and irregular due to low atmospheric density, rapid temperature variations, and complex terrain. We use long short-term memory (LSTM), bidirectional LSTM (BiLSTM), gated recurrent unit (GRU), and bidirectional GRU (BiGRU) architectures to evaluate the effectiveness of different activation functions and attention mechanisms. The activation functions tested include rectified linear unit (ReLU), leaky ReLU, exponential linear unit (ELU), Gaussian error linear unit (GELU), Swish, and scaled ELU (SELU), and model performance was measured using mean absolute error (MAE) and root mean square error (RMSE) metrics. Our results show that the integration of attentional mechanisms improves both MAE and RMSE, with Swish and ReLU achieving the best performance for minimum temperature prediction. Conversely, GELU and ELU were less effective for pressure prediction. These results highlight the critical role of selecting appropriate activation functions and attention mechanisms in improving model accuracy for complex time-series forecasting.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.4 no.3
• /
• pp.164-168
• /
• 2002

Yield prediction model of rice based on the boundary line analysis of the relationships between rice yield and meteorological conditions during rice growing period was reported in the previous report (Kim et al, 2001). Using the 15-year data of the 20 locations used for the model formulation and of the 12 locations not used, the model was tested for its predictability of location to location, year to year, and variety to variety variation of rice yield. The model predicted reliably the mean yield differences among locations, the yearly yield variation in each location, and the yield variation by variety. However, the model showed relatively lower predictability for the years of cool weather injury especially in mountainous locations. In conclusion, the model using boundary line analysis could be used to predict the yield responses to meteorological conditions during rice growth period and the locational, yearly, and varietal variations of rice yield. And the predictability of the present yield prediction model might be improved by including the boundary line analysis for the other factors such as soil characteristics, fertilization levels, etc.

• Journal of Korean Society of Forest Science
• /
• v.111 no.3
• /
• pp.357-364
• /
• 2022

We investigated the early growth characteristics of Quercus rubra planted in six regions (Hwaseong, Yangpyeong, Pyeongchang, Samcheok, Chungju, and Gimje) in South Korea in relation to soil physicochemical properties and meteorological factors. Q. rubra (1-0) were planted at a density of 3,000 trees ha^-1. The average height, root collar diameter (RCD), and volume of 8-year-old Q. rubra planted in 2014 were 3.52 m, 3.84 cm, and 0.0023 m³, respectively. The growth parameters of Q. rubra were the highest and lowest in Hwaseong and Pyeongchang, respectively. Correlation analysis among the soil physicochemical properties, meteorological factors, and plantation growth characteristics found that pH was the only soil factor negatively correlated with RCD, and the other soil factors were not significantly correlated with the growth characteristics. However, growth characteristics were positively correlated to average temperature from March to October and daily maximum temperature; and they were negatively correlated to altitude, topology, and the number of rainy days from March to October. In particular, the trees planted in Hwaseong area showed the best early growth characteristics because this area had the highest daily maximum temperature, the x average temperature from March to October, the low altitude, and it is located close to the foot of a mountain. In Pyeongchang, the early growth characteristics were negatively affected by winter cold damage because of the high altitude, low daily minimum temperature, and damage by wild animals. In Hwaseong, meteorological factors such as temperature and altitude were more highly correlated to growth characteristics of Q. rubra than the physicochemical soil properties. These results will provide useful information for determining suitable sites for Q. rubra plantations and for predicting early growth characteristics in response to environmental factors.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.17 no.2
• /
• pp.156-164
• /
• 2015

High temperature stress would affect rice production in the future as heat wave is expected to occur frequently under climate change conditions. The objective of this study was to obtain rudimentary information to assess the impact of heat stress on rice yield and its yield component in Korea. Two rice cultivars "Hwaseongbyeo" (Japonica) and "Dasanbyeo" (Tongil-type) were grown at different nitrogen fertilization levels in two seasons. These cultivars were grown in 1/5000a Wagner pot placed within four plastic houses where temperature was controlled at ambient, ambient$+1.5^{\circ}C$, ambient$+3^{\circ}C$ and ambient$+5^{\circ}C$ throughout the rice growing season in Suwon ($37^{\circ}16^{\prime}N$, $128^{\circ}59^{\prime}E$), Korea. The degree of temperature change affected grain yield whereas the level of nitrogen had little impact on grain yield. The number of panicle per pot and spikelet per panicle were not significantly different among temperature treatments in both cultivars tested. In contrast, 1000-grain weight and ripened grain ratio were decreased significantly under the treatments raising the air temperature to the level of $5.0^{\circ}C$ and $1.5^{\circ}C$ above the ambient air temperature in Dasanbyeo and Hwaseongbyeo, respectively. Reduction of 1000-grain weight and ripened grain ratio under the temperature treatments of $3.0^{\circ}C$ and $5.0^{\circ}C$ above the ambient air temperature resulted in significantly less grain yield for Dasanbyeo and Hwaseongbyeo, respectively. The greater sensitivity of grain yield to temperature increase in Dasanbyeo was attributable to the sharp decrease of 1000-grain weight and ripened grain ratio with the temperature rise above $23^{\circ}C$ during ripening period. On the other hand, Hwaseongbyeo had little variation of them in the temperature range of $23-27^{\circ}C$. These results suggested that grain yield would decrease under future climate conditions due to grain weight decreased by shorter grain filling period as well as the ripened grain ratio reduced by spikelet sterility and early abortion of rice kernel development. Thus, it would be essential to use cultivars tolerant to heat stress for climate change adaptation, which merits further studies for developing varieties that have traits to avoid spikelet sterility and early abortion of rice kernel, e.g., early morning flowering, under heat wave.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.8 no.4
• /
• pp.229-241
• /
• 2006

Global air temperature has risen by $0.6^{\circ}C$ over the last one hundred years due to increased atmospheric greenhouse gases. Moreover, this global warming trend is projected to continue in the future. This study was carried out to evaluate spatial variations in rice production areas by simulating rice-growth and development with projected high resolution climate data in Korea far 2011-2100, which was geospatially interpolated from the 25 km gridded data based on the IPCC SRES A2 emission scenario. Satellite remote sensing data were used to pinpoint the rice-growing areas, and corresponding climate data were aggregated to represent the official 'crop reporting county'. For the simulation experiment, we used a CERES-Rice model modified by introducing two equations to calculate the leaf appearance rate based on the effective temperature and existing leaf number and the final number of leaves based on day-length in the photoperiod sensitive phase of rice. We tested the performance of this model using data-sets obtained from transplanting dates and nitrogen fertilization rates experiments over three years (2002 to 2004). The simulation results showed a good performance of this model in heading date prediction [$R^2$=0.9586 for early (Odaebyeo), $R^2$=0.9681 for medium (Hwasungbyeo), and $R^2$=0.9477 for late (Dongjinbyeo) maturity cultivars]. A modified version of CERES-Rice was used to simulate the growth and development of three Japonica varieties, representing early, medium, and late maturity classes, to project crop status for climatological normal years between 2011 and 2100. In order to compare the temporal changes, three sets of data representing 3 climatological years (2011-2040, 2041-2070, and 2071-2100) were successively used to run the model. Simulated growth and yield data of the three Japonica cultivars under the observed climate for 1971-2000 was set as a reference. Compared with the current normal, heading date was accelerated by 7 days for 2011-2040 and 20 days for 2071-2100. Physiological maturity was accelerated by 15 days for 2011-2040 and 30 days for 2071-2100. Rice yield was in general reduced by 6-25%, 3-26%, and 3-25% per 10a in early, medium, and late maturity classes, respectively. However, mid to late maturing varieties showed an increased yield in northern Gyeonggi Province and in most of Kwangwon Province in 2071-2100.

• Journal of The Korean Society of Grassland and Forage Science
• /
• v.38 no.1
• /
• pp.53-60
• /
• 2018

Yield prediction model for mixed pasture was developed with a shortage that the relationship between dry matter yield (DMY) and days of summer depression (DSD) was not properly reflected in the model in the previous research. Therefore, this study was designed to eliminate the data of the regions with distinctly different climatic conditions and then investigate their relationships DMY and DSD using the data in each region separately of regions with distinct climatic characteristics and classify the data based on regions for further analysis based on the previous mixed pasture prediction model. The data set used in the research kept 582 data points from 11 regions and 41 mixed pasture types. The relationship between DMY and DSD in each region were analyzed through scatter plot, correlation analysis and multiple regression analysis in each region separately. In the statistical analysis, DMY was taken as the response variable and 5 climatic variables including DSD were taken as explanatory variables. The results of scatter plot showed that negative correlations between DMY and DSD were observed in 7 out of 9 regions. Therefore, it was confirmed that analyzing the relationship between DMY and DSD based on each region is necessary and 5 regions were selected (Hwaseong, Suwon, Daejeon, Siheung and Gwangju) since the data size in these regions is large enough to perform the further statistical analysis based on large sample approximation theory. Correlation analysis showed that negative correlations were found between DMY and DSD in 3 (Hwaseong, Suwon and Siheung) out of the 5 regions, meanwhile the negative relationship in Hwaseong was confirmed through multiple regression analysis. Therefore, it was concluded that the interpretability of the yield prediction model for mixed pasture could be improved based on constructing the models using the data from each region separately instead of using the pooled data from different regions.

• Korean Journal of Remote Sensing
• /
• v.36 no.5_4
• /
• pp.1267-1276
• /
• 2020

To quantitatively predict the impacts of large-scale volcanic eruptions of Mt. Baekdu on air quality and damage around the Korean Peninsula, a three-dimensional chemistry-transport modeling system (Weather Research & Forecasting - Sparse Matrix Operation Kernel Emission - Comunity Multi-scale Air Quality) was adopted. A worst-case meteorology scenario was selected to estimate the direct impact on Korea. This study applied the typical worst-case scenarios that are likely to cause significant damage to Korea among worst-case volcanic eruptions of Mt. Baekdu in the past decade (2005~2014) and assumed a massive VEI 4 volcanic eruption on May 16, 2012, to analyze the concentration of PM_2.5 caused by the volcanic eruption. The effects of air quality in each region-cities, counties, boroughs-were estimated, and vulnerable areas were derived by conducting an exposure assessment reflecting vulnerable groups. Moreover, the effects of cities, counties, and boroughs were analyzed with a high-resolution scale (9 km × 9 km) to derive vulnerable areas within the regions. As a result of analyzing the typical worst-case volcanic eruptions of Mt. Baekdu, a discrepancy was shown in areas between high PM_2.5 concentration, high population density, and where vulnerable groups are concentrated. From the result, PM_2.5 peak concentration was about 24,547 ㎍/㎥, which is estimated to be a more serious situation than the eruption of Mt. St. Helensin 1980, which is known for 540 million tons of volcanic ash. Paju, Gimpo, Goyang, Ganghwa, Sancheong, Hadong showed to have a high PM_2.5 concentration. Paju appeared to be the most vulnerable area from the exposure assessment. While areas estimated with a high concentration of air pollutants are important, it is also necessary to develop plans and measures considering densely populated areas or areas with high concentrations of susceptible population or vulnerable groups. Also, establishing measures for each vulnerable area by selecting high concentration areas within cities, counties, and boroughs rather than establishing uniform measures for all regions is needed. This study will provide the foundation for developing the standards for disaster declaration and preemptive response systems for volcanic eruptions.