국제학술지 게시판목록 | 이화여자대학교 기후물리실험실

Publications

국제학술지

공지 (2025. 10. 28.)

현재까지 출판된 논문 목록입니다. 아직 준비 중인 논문은 포함하지 않은 점 참고 바랍니다.

Untangling the contribution of input parameters to an artificial intelligence PM2.5 forecast model using the LRP method

Abstract: The recurrent neural network (RNN), an artificial intelligence algorithm, applied to the predictions based on the Community Multiscale Air Quality operational model has significantly improved the forecast accuracy of the concentrations of particulate matter with a diameter of ≤2.5 μm (PM2.5) in the Seoul metropolitan area of the Republic of Korea. It is challenging to interpret the prediction results and identify the related error sources because the decision-making process of the RNN model is inaccessible. This study evaluated the relevance score of the RNN input variables using the layer-wise relevance propagation (LRP) at 6-hourly forecasts over the winters of 2015–2021 (December through February). The relevance score magnitudes summed over the period from the target prediction time to 2–5 and 4–7 time-steps before it (i.e., the most recent 12–30 h and 24–42 h, respectively) show ∼80% of the total relevance score for one- and two-day forecasts, respectively. The input variables were originally selected by their correlation coefficients with the observed PM2.5 concentration; however, the order of input variable contributions measured by the LRP differs from the order of the correlation coefficients, implying inconsistency between the linear and nonlinear methods. Retraining the RNN model using a subset of variables of high relevance scores is found to yield prediction skills comparable to the original set of input variables. This study can contribute to the improvement of the RNN prediction model by decoding the black box of an artificial intelligence model using the LRP method. Full title: Untangling the contribution of input parameters to an artificial intelligence PM2.5 forecast model using the layer-wise relevance propagation method
- 작성자Kim et al.
- 작성일2024.09.04
- 조회수64
- 2022
Assessing the influence of large-scale environmental conditions on rainfall structure of Atlantic tropical cyclones: An

Abstract: Understanding the mechanisms related to the variations in the rainfall structure of tropical cyclones (TCs) is crucial in improving forecasting systems of TC rainfall and its impact. Using satellite precipitation and reanalysis data, we examined the influence of along-track large-scale environmental conditions on inner-core rainfall strength (RS) and total rainfall area (RA) for Atlantic TCs during the TC season (July–November) from 1998 to 2019. Factor analysis revealed three major factors associated with variations in RS and RA: large-scale low and high pressure systems [factor 1 (F1)]; environmental flows, sea surface temperature, and humidity [factor 2 (F2)]; and maximum wind speed of TCs [factor 3 (F3)]. Results from our study indicate that RS increases with an increase in the inherent primary circulation of TCs (i.e., F3) but is less affected by large-scale environmental conditions (i.e., F1 and F2), whereas RA is primarily influenced by large-scale low and high pressure systems (i.e., F1) over the entire North Atlantic and partially influenced by environmental flows, sea surface temperature, humidity, and maximum wind speed (i.e., F2 and F3). A multivariable regression model based on the three factors accounted for the variations of RS and RA across the entire basin. In addition, regional distributions of mean RS and RA from the model significantly resembled those from observations. Therefore, our study suggests that large-scale environmental conditions over the North Atlantic Ocean are important predictors for TC rainfall forecasts, particularly with regard to RA. Full title: Assessing the influence of large-scale environmental conditions on rainfall structure of Atlantic tropical cyclones: An observational study
- 작성자Kim et al.
- 작성일2024.09.04
- 조회수59
- 2021
Assessment and valuation of health impacts of fine PM during COVID-19 lockdown: A comprehensive study...

Abstract: A novel coronavirus disease (COVID-19) continues to challenge the whole world. The disease has claimed many fatalities as it has transcended from one country to another since it was first discovered in China in late 2019. To prevent further morbidity and mortality associated with COVID-19, most of the countries initiated a countrywide lockdown. While physical distancing and lockdowns helped in curbing the spread of this novel coronavirus, it led to massive economic losses for the nations. Positive impacts have been observed due to lockdown in terms of improved air quality of the nations. In the current research, ten tropical and subtropical countries have been analysed from multiple angles, including air pollution, assessment and valuation of health impacts and economic loss of countries during COVID-19 lockdown. Countries include Brazil, India, Iran, Kenya, Malaysia, Mexico, Pakistan, Peru, Sri Lanka, and Thailand. Validated Simplified Aerosol Retrieval Algorithm (SARA) binning model is used on data collated from moderate resolution imaging spectroradiometer (MODIS) for particulate matters with a diameter of less than 2.5 μm (PM2.5) for all the countries for the month of January to May 2019 and 2020. The concentration results of PM2.5 show that air pollution has drastically reduced in 2020 post lockdown for all countries. The highest average concentration obtained by converting aerosol optical depth (AOD) for 2020 is observed for Thailand as 121.9 μg/m3 and the lowest for Mexico as 36.27 μg/m3. As air pollution is found to decrease in the April and May months of 2020 for nearly all countries, they are compared with respective previous year values for the same duration to calculate the reduced health burden due to lockdown. The present study estimates that cumulative about 100.9 Billion US$ are saved due to reduced air pollution externalities, which are about 25% of the cumulative economic loss of 435.9 Billion US$. Full title: Assessment and valuation of health impacts of fine particulate matter during COVID-19 lockdown: A comprehensive study of tropical and sub tropical countries
- 작성자Bherwani et al.
- 작성일2024.09.04
- 조회수72
- 2021
Asymmetric expansion of summer season on May and September in Korea

Abstract: Global warming and its associated changes in the timing of seasonal progression may produce substantial ripple effects on the regional climate and ecosystem. This study analyzes the surface air temperature recorded during the period 1919–2017 at seven stations in the Republic of Korea to investigate the long-term changes at the beginning and ending of the summer season and their relationship with the warming trends of spring and autumn. The temperatures at the starting (June 1) and ending (August 31) dates of the past period (1919–1948) advanced by 13 days and delayed by 4 days, respectively, for the recent period (1988–2017). This asymmetric change was caused by continuous warming in May for the entire period of analysis and an abrupt warming in September in the recent decades. Different amplitudes of the expansion of the western North Pacific subtropical high in May and September are responsible for the asymmetric expansion of the summer season. The projections of surface warming for spring and autumn in Korea used the downscaled grid data of a regional climate model, which were obtained by the Representative Concentration Pathway 8.5 scenario of a general circulation model, and indicated a continuous positive trend until 2100. Larger interannual variability of blooming timing of early autumn flowers than that of late spring flowers may represent the response of the ecosystem to the seasonally asymmetric surface warming. Results suggest that the shift of seasons and associated warming trend have a disturbing effect on an ecosystem, and this trend will intensify in the future.
- 작성자Ho et al.
- 작성일2024.09.04
- 조회수132
- 2021
Development of a PM2.5 prediction model using a recurrent neural network algorithm for the Seoul metropolitan area, ROK

Abstract: The National Institute of Environmental Research, the Ministry of Environment, has been forecasting the concentrations of particulate matter (PM) with a diameter ≤ 2.5 μm (PM2.5) over Seoul, Republic of Korea, in terms of four PM2.5 concentration categories (low, moderate, high, and very high) since August 31, 2013. The current model, the Community Multiscale Air Quality (CMAQ) model, is run four times a day to forecast air quality for up to two days in 6-h intervals. In 2018, the hit ratio (i.e., accuracy) of the model was 60%, with an additional increase of 10% with the involvement of a forecaster. The CMAQ was improved in this study by incorporating a recurrent neural network (RNN) algorithm for the Seoul Metropolitan Area. Input datasets to the RNN model—PM values, meteorological parameters, and back-trajectory tracks obtained from both observations and model forecasts—were sorted according to time as the RNN algorithm learns time sequence series information, unlike typical neural network algorithms. To reflect the seasonality of the meteorological parameters that influence the PM2.5 concentrations in the region, one year was divided into 36 sets of three-month periods (i.e., there are three sets for July: May–June–July, June–July–August, and July–August–September). Several indices representing the accuracy of the forecast were calculated based on the RNN model results for 2018 after training the model for the previous three years (2015–2017). The accuracy of the RNN model is 74–81% for forecast lead times up to two days, about 20% higher than the CMAQ-only forecasts and ~10% higher than the combined CMAQ-forecaster forecast. The RNN model probabilities of detection for both high and very high PM2.5 events are comparable to those of the CMAQ model; however, the RNN model notably reduces the false alarm rate. Overall, the RNN model yields higher performance than the current forecast methods. Hence, this model can be adopted as an operational forecast model in Korea. Full title: Development of a PM2.5 prediction model using a recurrent neural network algorithm for the Seoul metropolitan area, Republic of Korea
- 작성자Ho et al.
- 작성일2024.09.04
- 조회수94
- 2021
Development of a track-pattern-based medium-range tropical cyclone forecasting system for the western North Pacific

Abstract: Despite tremendous advancements in dynamical models for weather forecasting, statistical models continue to offer various possibilities for tropical cyclone (TC) track forecasting. Herein, a track-pattern-based approach was developed to predict a TC track for a lead time of 6–8 days over the western North Pacific (WNP), utilizing historical tracks in conjunction with dynamical forecasts. It is composed of four main steps: 1) clustering historical tracks similar to that of an operational 5-day forecast in their early phase into track patterns, and calculating the daily mean environmental fields (500-hPa geopotential height and steering flow) associated with each track; 2) deriving the two environmental variables forecasted by dynamical models; 3) evaluating pattern correlation coefficients between the two environmental fields from step 1 and those from dynamical model for a lead times of 6–8 days; and 4) producing the final track forecast based on relative frequency maps obtained from the historical tracks in step 1 and the pattern correlation coefficients obtained from step 3. TCs that formed in the WNP and lasted for at least 7 days, during the 9-yr period 2011–19 were selected to verify the resulting track-pattern-based forecasts. In addition to the performance comparable to dynamical models under certain conditions, the track-pattern-based model is inexpensive, and can consistently produce forecasts over large latitudinal or longitudinal ranges. Machine learning techniques can be implemented to incorporate nonlinearity in the present model for improving medium-range track forecasts.
- 작성자Cheung et al.
- 작성일2024.09.04
- 조회수82
- 2021
Large volcanic eruptions reduce landfalling tropical cyclone activity: Evidence from tree rings

Abstract: Tropical cyclones (TCs) are one of the most devastating storm systems with high socioeconomic impacts around the world. The drivers of long-term changes in TC frequency and intensity, including the recent global climatic changes, are, however, poorly understood due to short instrumental measurements and a lack of accurate proxy records. Here we present the long-term impacts of large volcanic eruptions on TC activity in northeast Asia. For this purpose, we performed a reconstruction of the frequency and intensity of landfalling TCs based on tree-ring oxygen isotope ratios over the past 350 years. Our results revealed markedly depleted δ18O values (P < 0.01) for TC years and confirm tree-ring δ18O as a strong proxy for the detection of past TCs. The agreement between the δ18O chronology and the corresponding TC record (1950–2005) was 96.4% and it was relatively high also for the preceding periods covered by less reliable TC records, specifically 76.1% (1904–1949) and 66.4% (1652–1903). In addition to the prominent TC frequency signal, we found a strong negative correlation (R = −0.65; P < 0.001) between the δ18O chronology and TC intensity expressed by the amount of TC precipitation. Our reconstruction revealed that the recent high frequency of landfalling TCs is distinct on a long-term scale. We provide the first long-term evidence of reduced TC activity for two years following large volcanic eruptions. Our results indicate that volcanic ash is a relevant driver of TC activity over northeast Asia via its role on radiative climate forcing. We suggest that large volcanic eruptions modulate large-scale atmospheric and oceanic circulation determining TC genesis and thus TC activity in the western North Pacific.
- 작성자Altman et al.
- 작성일2024.09.04
- 조회수81
- 2021
Possible cause of seasonal inhomogeneity in interdecadal changes of tropical cyclone genesis frequency over the WNP

Abstract: An abrupt decrease in annual tropical cyclone genesis frequency (TCGF), which is statistically significant only from October to December (OND), has been noticed over the western North Pacific Ocean. However, the seasonal inhomogeneity of interdecadal changes in TCGF between OND and the other seasons (from January to September) and the associated mechanisms are not clearly documented. This study examines and compares the different interdecadal changes in OND and in January–September from 1979 to 2018. According to our analysis, the TCGF decrease in OND (2.2) accounts for 79% of the total decrease (2.8) in annual TCGF after 1998, whereas the TCGF in January to September remains unchanged. The key differences in large-scale environment are found from the extension of equatorial easterly wind anomalies and attendant anticyclone anomalies in the subtropics. Under similar sea surface temperature (SST) warming pattern in the tropical Indo-Pacific region (i.e., the La Niña–like SST warming), tropical precipitation is significantly enhanced over the area where its seasonal peak occurs: the tropical Indian Ocean in OND and the tropical western Pacific in January–September. Thus, the equatorial easterly wind anomalies extend westward to 110°E in OND and to 145°E in January–September. Different extension of easterly wind anomalies results in different expansion of attendant large-scale anticyclone anomaly over the subtropical western Pacific, which dominates the entire main development region in OND but not in January–September. To summarize, the different extensions of easterly wind anomalies under similar La Niña–like SST warming are responsible for the seasonal inhomogeneity of interdecadal changes in TCGF. Full title: Possible cause of seasonal inhomogeneity in interdecadal changes of tropical cyclone genesis frequency over the western North Pacific
- 작성자Chang et al.
- 작성일2024.09.04
- 조회수58
- 2021
Quantifying the impact of synoptic weather systems on high PM2.5 episodes in the Seoul metropolitan area, Korea

Abstract: Variations in concentrations of PM2.5, particulate matters of diameters below 2.5 μm, vary following both meteorological conditions and emissions controls. Meteorological conditions particularly affect short-term high PM2.5 episodes through accumulations, transports, and secondary formations. This study quantifies the meteorological impacts on high PM2.5 episodes in the Seoul Metropolitan Area (SMA), Korea, for the period 2016–2018 using empirical and statistical methods. Synoptic weather maps of 77 high PM2.5 episodes in 2016 are grouped into two synoptic types: onshore winds associated with migratory pressure systems over the SMA and offshore winds from continental high pressure extending toward the SMA. We applied principal component analysis and regression to extract the dominant synoptic types controlling PM2.5 variability. It identifies two major principal components (PCs) from 12 surface and upper-air meteorological variables for 2017–2018. Weather patterns in 49 examples of the high-positive PCs show that the two PCs are capable of reproducing the synoptic weather patterns relevant for high PM2.5 episodes. To quantify the relationship between the synoptic weather patterns and PM2.5 levels, the two PCs are further classified into four groups according to their signs. Positive- and negative-PC groups are associated with about 82% and 73% of high- and low-PM2.5 episodes, respectively, suggesting that most of the high/low PM2.5 episodes in the SMA can occur under the two PCs-dominant weather conditions. The results can be utilized as a reference for daily predictions of high PM2.5 episodes, as well as for quantitative analysis of the climatic influence on the long-term PM2.5 variability.
- 작성자Chang et al.
- 작성일2024.09.04
- 조회수62
- 2021
Regulatory measures significantly reduced air-pollutant concentrations in Seoul, Korea

Abstract: The Government of the Republic of Korea has enforced strict regulations to improve air quality since the early 2000s. The regulations are mainly focused on reducing vehicle emissions in the Seoul metropolitan region by conforming to the European emissions standards, replacing diesel buses with compressed natural gas buses, incentives for installing diesel exhaust after-treatment systems and buying eco-friendly vehicles. There was a 20% reduction in 2010s compared to the 2000s in terms of the mean concentrations of particulate matter (PM) with mean aerodynamic diameters of ≤10 μm (PM10) and 2.5 μm (PM2.5) during cold seasons (October through following February) although the decrease may not be entirely attributable to the regulations. The influences of other external factors such as transboundary transport of air pollutants and regional meteorological conditions cannot be neglected. This study analyzes the change in the diurnal variations—two maxima at around 11 and 22 local time (LT) and two minima at around 6 and 16 LT—of air pollutant concentrations that may be closely related to the regulatory action in reducing local vehicle emissions. A reduction of over 40% for the amplitude of two PM concentrations at 11 LT was revealed when values from the 2010s were compared to those from the 2000s. There was a similar reduction for other vehicle exhaust gases including nitrogen dioxide, carbon monoxide, and sulfur dioxide. Additional analysis of the long-term trend in mixed layer height and surface wind speed showed that the change in environmental conditions in the diurnal time scale was either negligible or unfavorable for conditions which reduce PM concentrations. This study suggests that the mean concentration estimation may underestimate the regulatory effects, but the approach based on the diurnal variation may be a more accurate indicator.
- 작성자Ho et al.
- 작성일2024.09.04
- 조회수83
- 2021