Journal of Atmospheric and Solar-Terrestrial Physics最新文献

{"title":"Estimation, prediction, and forecasting of urban solar brightness: A comprehensive benchmarking of empirical, hybrid AI, and Deep-NARMAX models","authors":"Youness EL Mghouchi ,&nbsp;Mihaela Tinca Udristioiu","doi":"10.1016/j.jastp.2026.106761","DOIUrl":"10.1016/j.jastp.2026.106761","url":null,"abstract":"<div><div>Understanding the variability of solar brightness is essential for optimising solar energy systems, improving urban air quality assessments, and enhancing environmental forecasting. The aim of this study is to investigate the influence of meteorological and atmospheric pollutant variables—including temperature, relative humidity, precipitation, wind direction, wind speed, PM₂.₅, PM₁₀, ozone (O₃), carbon monoxide (CO), SO₂, O₃, NO, NO₂, NO_x and others—on incoming solar radiation, using solar brightness as a proxy. A comprehensive dataset spanning five years of hourly observations was analysed. Open-source data from four air quality monitoring stations in Craiova, was provided by the Romanian Environmental Agency. The study followed a five-stage approach. First, data preprocessing was conducted to identify and address anomalies, outliers, and missing values, while trends for solar brightness and other studied variables were analysed. In the second stage, the best global solar radiation (GSR) model among 10 GSR models is selected. In the third stage, correlations between solar brightness and other variables, including data provided by the best GSR model, based on exploratory data analysis, were examined. A deep AI-based hybrid approach was applied in the fourth stage to discover the optimal AI predictive model for solar brightness based on related variables. Finally, a deep NARMAX model was elaborated and applied to model and anticipate next hourly solar brightness in Craiova. A set of statistical metrics was employed to assess the results of the models.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"280 ","pages":"Article 106761"},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147420971","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Using physical parameters in a deep neural network for regional VTEC prediction","authors":"Mohammad Alikhani ,&nbsp;Yazdan Amerian ,&nbsp;Hany Mahbuby","doi":"10.1016/j.jastp.2026.106737","DOIUrl":"10.1016/j.jastp.2026.106737","url":null,"abstract":"<div><div>Accurate prediction of Total Electron Content (TEC) is essential for satellite navigation and communication systems, as ionospheric disturbances affect signal accuracy. TEC forecasting supports the stable operation of GNSS technologies. Artificial intelligence (AI), particularly deep learning, has shown strong potential in TEC prediction by modeling complex spatiotemporal patterns. This study presents a hybrid deep learning model combining BiLSTM, CNN, and GRU to predict TEC over Iran. The training dataset consisted of 365 days of Global Ionospheric Maps (GIM) VTEC data and ionospheric parameters from the International Reference Ionosphere (IRI) during 2017, while validation used the first quarter of 2018. The year 2017 was selected to cover a wide range of geomagnetic and solar conditions, thus providing comprehensive input for model development. Input parameters included electron temperature, critical frequencies, geomagnetic indices, and solar activity indicators, enabling the model to capture both spatial and temporal dependencies. During validation on 2018 data, the model achieved an RMSE of 1.63 TECU, MAE of 1.19 TECU, NRMSE of 0.04, MAPE of 11.80%, and a correlation coefficient of 92.15%, clearly outperforming the IRI-2016 model and closely matching the GIM values. Evaluation using 2019 data, covering dates with high and low geomagnetic activity across multiple locations in Iran, demonstrated strong predictive accuracy, with RMSE of 1.35 TECU, MAE of 1.12 TECU, NRMSE of 0.10, MAPE of 9.55%, and R of 97.62%. The improvement is largely due to the inclusion of physical parameters, highlighting the robustness of the proposed model under diverse ionospheric conditions.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"280 ","pages":"Article 106737"},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147420976","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Storm-time geomagnetic variations and Pc5 activity at mid- and high-latitudes: Insights from the 20 November 2003 storm","authors":"E.M. Takla ,&nbsp;A. Yoshikawa ,&nbsp;T. Uozumi","doi":"10.1016/j.jastp.2026.106763","DOIUrl":"10.1016/j.jastp.2026.106763","url":null,"abstract":"<div><div>Geomagnetic field measurements from four mid- and high-latitude stations of the MAGDAS/CPMN network were analyzed to investigate the effects of the intense Geomagnetic Storm (GS) of November 20, 2003 on geomagnetic field variations and Pc5 pulsation activity. The results show that the geomagnetic field measurements at high-latitude stations were strongly influenced by variations in the interplanetary magnetic field (IMF) and solar wind parameters during this storm. Both the amplitude of geomagnetic fluctuations and the intensity of Pc5 activity were found to be significantly larger at polar and auroral latitudes than at mid-latitude. In particular, the occurrence and amplitude of Pc5 pulsations were greatest within the auroral zone and decreased slightly with increasing latitude poleward. The findings, consistent with previous results reported in the literature, emphasize the central role of auroral-latitude ionospheric processes in controlling storm-time geomagnetic variability.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"280 ","pages":"Article 106763"},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147422310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A simulation study on the key cloud microphysical processes in an extreme warm-sector heavy rainfall over the south China mountains","authors":"Xuehan Dong ,&nbsp;Jiangnan Li ,&nbsp;Zhourong Liu ,&nbsp;Jianfei Chen ,&nbsp;Risheng Liu","doi":"10.1016/j.jastp.2026.106740","DOIUrl":"10.1016/j.jastp.2026.106740","url":null,"abstract":"<div><div>This study employs the Weather Research and Forecasting (WRF) mesoscale model to simulate an extreme warm-sector heavy rainfall event that occurred on 17–18 June 2022, over Yuanbao Mountain in Liuzhou, Guangxi. The research focuses on the cloud microphysical characteristics and latent heat budget during the heavy precipitation stage, aiming to clarify the key physical mechanisms driving the intensification of the heavy rainfall. The simulation successfully reproduces the spatio-temporal evolution of this extreme precipitation. Solid-phase hydrometeors, namely snow and graupel, are found to dominate the precipitation process, exhibiting complementary vertical distributions that formed an efficient hydrometeor conversion chain and served as the primary source of rainwater. The intense release of condensation latent heat near the 0 °C level acted as the core energy source for precipitation, while the depositional latent heat release from ice-phase particles in the mid-upper levels served as a “leading indicator” for the extreme intensification of rainfall. The center of extreme precipitation was located on the southern windward slope of the mountains. There, warm and moist air parcels underwent adiabatic cooling during upslope ascent. Upon reaching saturation, water vapor condensed, releasing substantial latent heat and establishing a typical positive feedback mechanism of “orographic lifting–condensational heating.” This process significantly altered the local thermal structure and vertical motion of the atmosphere, representing the direct cause for the triggering of the extreme heavy rainfall.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"280 ","pages":"Article 106740"},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146077017","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Solar-geomagnetic activity effect on NmE over Ouagadougou, an equatorial anomaly station in the African sector","authors":"A.K. Kazeem ,&nbsp;A.B. Rabiu ,&nbsp;C. Amory-Mazaudier","doi":"10.1016/j.jastp.2026.106756","DOIUrl":"10.1016/j.jastp.2026.106756","url":null,"abstract":"<div><div>The variability of the ionospheric E-region peak electron density (NmE) and its response to solar, solar wind, and geomagnetic forcing are investigated at Ouagadougou (12.4° N, 358.5° E; dip latitude +1.45°), an equatorial station in the African sector. Daytime hourly measurements of the E-layer critical frequency (foE) are analysed in conjunction with selected solar, interplanetary, and geomagnetic indices over the period 1976-1997, spanning solar cycles 21 and 22. The results show a pronounced diurnal variation in NmE and a strong dependence on solar activity. Seasonal variability is weak and not consistently observed across years, as represented by high, moderate, and low solar activity (1991, 1993, and 1996). No clear linear long-term relationship is found between NmE and sunspot number. Geomagnetic activity exerts a measurable influence on NmE, particularly during periods of high and low solar activity. Among the parameters examined, the interplanetary magnetic field magnitude exhibits the strongest association with NmE, while the planetary amplitude index, disturbance storm time, solar wind dynamic pressure, and the southward interplanetary magnetic field component show weaker but detectable effects. Solar wind speed and sunspot number contribute only marginally to NmE variability. These results emphasise the coupled roles of solar and magnetospheric processes in controlling the equatorial E-region ionosphere over the African sector.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"280 ","pages":"Article 106756"},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147420970","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Statistical analysis of lightning activity in recurrent lightning locations in northeastern Colombia","authors":"S. Ardila,&nbsp;E. Soto,&nbsp;K. Ríos,&nbsp;M. Romero","doi":"10.1016/j.jastp.2026.106745","DOIUrl":"10.1016/j.jastp.2026.106745","url":null,"abstract":"<div><div>This study presents a temporal-based approach for identifying Recurrent Lightning Locations (RLLs) in the municipalities of Barrancabermeja and Yondó, in northeastern Colombia, using lightning data from 2014 to 2021. Results reveal that RLLs are predominantly situated near the Magdalena River. Temporally, lightning activity in RLSs shows a bimodal seasonal distribution with early morning peaks in stroke occurrence. The analysis of flash multiplicity and peak current distributions indicates that most events are single-stroke flashes with right-skewed current profiles, occasionally exceeding 75 kA. Notably, over 42% of RLLs are located outside high Ground Flash Density (GFD) zones, highlighting that lightning recurrence is not necessarily correlated with stroke density but is likely influenced by localized geographic and atmospheric factors. The study proposes a novel methodological framework centered on temporal recurrence rather than spatial density, offering practical implications for lightning risk assessment in infrastructure-critical regions. Results highlight the importance of considering both spatial and temporal dimensions in lightning risk assessment and offer a practical framework for improving lightning protection strategies in vulnerable regions. The application of temporal methodologies, as implemented in this study, constitutes a distinctive criterion for identifying sites with recurrent lightning activity, with a focus on the design of optimal lightning protection systems.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"280 ","pages":"Article 106745"},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147420972","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparison of precipitation-based drought indices: Drought assessment and monitoring in a TR82 region in Türkiye","authors":"Utku Zeybekoglu","doi":"10.1016/j.jastp.2026.106753","DOIUrl":"10.1016/j.jastp.2026.106753","url":null,"abstract":"<div><div>Drought is a multifaceted natural hazard with profound impacts on water resources, agriculture, ecosystems, and socio-economic systems. This study investigates meteorological drought in the TR82 Region of northern Türkiye, analysing precipitation data from ten meteorological stations spanning 1981–2023. Three precipitation-based drought indices—the Standardized Precipitation Index (SPI), China-Z Index (CZI), and Z-Score Index (ZSI)—were applied to assess drought characteristics, including frequency, severity, and spatial variability. Analysis revealed recurrent drought periods in 1981–1982, 1985–1986, 1989, 1992–1994, 1996, 2003–2004, 2006–2008, 2013, 2017, 2020, and 2022, indicating that drought occurrence in the region is neither sporadic nor isolated. Moderately dry and wet conditions were more frequent than extreme events, while prolonged droughts after 2000 suggest increasing persistence and intensification. Correlation analysis demonstrated very strong agreement among the indices, with mean correlation coefficients of 0.9956 (SPI–ZSI), 0.9953 (SPI–CZI), and 0.9943 (ZSI–CZI), confirming ZSI as a viable alternative to SPI for regional drought monitoring. In addition, the results are consistent with the presence of scale-invariant, nonlinear dynamics in precipitation variability, as indicated by observed power-law behavior and the framework of self-organized criticality (SOC) in rainfall systems, suggesting that short-term fluctuations are statistically related to longer-term patterns. The use of precipitation-based drought indices showed that the driest years align with the years of lowest annual precipitation, and similar correspondence is observed for wet years. These findings provide robust evidence of the TR82 Region's vulnerability to drought and underscore the importance of multi-index approaches and consideration of nonlinear climate dynamics for effective drought assessment and proactive water resource management.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"280 ","pages":"Article 106753"},"PeriodicalIF":1.9,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147421987","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Long-term spatiotemporal analysis of variation in soil moisture over Iran","authors":"Mohammad Darand ,&nbsp;Ramtin Tashan","doi":"10.1016/j.jastp.2026.106727","DOIUrl":"10.1016/j.jastp.2026.106727","url":null,"abstract":"<div><div>Soil moisture, as an important hydrological component, plays a crucial role in land–atmosphere interactions. Understanding variations in soil moisture content is highly valuable for effective water resource management, agricultural activities, and climate adaptation. The objective of this study is to analyze the spatiotemporal variations of surface soil moisture (0–7 cm) across Iran during the period 1979–2024. To achieve this, daily gridded data with a spatial resolution of 0.1° from the ERA5-Land dataset provided by ECMWF were used. The results showed that the spatial distribution pattern of soil moisture content follows the spatial patterns of precipitation, rainy days, and temperature across Iran. The modified Mann–Kendall test and Sen's slope estimator were applied to detect trends and their magnitudes at a 95 % confidence level. The findings indicated that soil moisture content across Iran has shown a decreasing trend, with an average reduction of 0.0032 m³ m⁻³ per decade. Temporally, the greatest reduction in soil moisture occurred during the cold and rainy seasons. Spatially, the decrease in soil moisture volume during winter was significantly higher in the northeastern part of the country compared to other regions. In some areas of the Alborz and Zagros highlands, however, soil moisture content exhibited an increasing trend. The findings of this study suggest that changes in soil moisture can be a potential predictor of climate change and can be applied to the fields of water resource management, agriculture, flood management, and hydrology.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"279 ","pages":"Article 106727"},"PeriodicalIF":1.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145940742","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Research on rainfall prediction near Qinyang-Yichuan expressway based on deep learning","authors":"Mingfei Zhang ,&nbsp;Jiaheng Wang ,&nbsp;Xiaogang Wei ,&nbsp;Guotao Dou ,&nbsp;Xiaorui Wang","doi":"10.1016/j.jastp.2025.106716","DOIUrl":"10.1016/j.jastp.2025.106716","url":null,"abstract":"<div><div>In response to the challenges posed by complex terrain and the strong non-stationarity of rainfall—factors that complicate prediction efforts near the Qinyang-Yichuan Expressway and can easily trigger engineering disasters such as tunnel water ingress and slope landslides—this study aims to develop a high-precision rainfall prediction model for disaster prevention in engineering applications. Utilizing historical rainfall data from the region, we propose a hybrid model that integrates Empirical Mode Decomposition (EMD) and the Coyote Optimization Algorithm (COA) with Long Short-Term Memory (LSTM) networks (EMD-COA-LSTM), alongside a Comprehensive Index (CI). A comparative study of 13 deep learning models reveals that traditional models generally underestimate rainfall amounts in complex terrain predictions. The EMD-COA-LSTM model (CI = 5.78) effectively mitigates this bias and demonstrates significantly superior overall performance compared to the worst-performing baseline LSTM model (CI = 209.11). Further analysis shows that the ranking of overall model prediction performance and most evaluation metrics does not fully correspond with the coefficient of determination (R²), indicating that R² alone is insufficient for assessing model performance and should be combined with other metrics. Evaluation based on the CI indicator provides a more accurate reflection of the model's overall effectiveness. The findings of this study offer a more reliable scientific basis for geological disaster risk early warning, tunnel flood-season construction, and engineering safety management along expressway corridors, demonstrating substantial practical engineering value.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"279 ","pages":"Article 106716"},"PeriodicalIF":1.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145940743","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrating LGBM machine learning methods with SHAP model to explain the impact of different environmental factors on precipitation in China","authors":"Yong Liu ,&nbsp;Qingzu Luan ,&nbsp;Pengguo Zhao","doi":"10.1016/j.jastp.2025.106715","DOIUrl":"10.1016/j.jastp.2025.106715","url":null,"abstract":"<div><div>Precipitation processes are influenced by a combination of environmental variables, and it is crucial to identify the main contributing factors to rainfall. China has vast geographical diversity and complex terrain, with significant differences in precipitation formation mechanisms across climatic regions. Although machine learning models are efficient, they lack interpretability. Therefore, this study utilizes ground-based meteorological observation data, satellite remote sensing data, and atmospheric reanalysis data in conjunction with the explainable artificial intelligence (XAI) tool SHAP model combined with the Light Gradient Boosting Machine (LGBM) machine learning framework to investigate the impact of various environmental variables on precipitation in China and different climate regions, quantifying their contributions to precipitation. The results reveal that relative humidity (RH), Convective Available Potential Energy (CAPE), K index (KI), and ice water path (IWP) are the most critical factors influencing precipitation within China, ranking among the top four in terms of average SHAP values and significantly higher than other factors. Due to regional variations across different climate zones in China, land surface temperature (LST), wind direction (WD), evaporation (E), liquid water path (LWP), surface pressure (SP), cloud cover (CFC), aerosol optical depth (AOD), and relative humidity (RH) exert the most pronounced positive effects on precipitation at the national scale, while ice water path (IWP), K index (KI), cloud top height (CTH), and Convective Available Potential Energy (CAPE) demonstrate more significant negative impacts—stemming from the varying influences of each variable across different climate regions. Significant regional variations exist in precipitation drivers. CAPE shows stronger influence on precipitation in North Subtropical Humid climatic regions, while RH dominates in Marginal Tropical Humid and Mid-temperate Semi-humid zones. CTH is more pronounced in Plateau Temperate Semi-arid areas, IWP stands out in Mid-temperate Semi-arid climatic regions, KI predominates in Mid-temperate Arid areas, and LST plays a more significant role in Warm Temperate Semi-humid climatic regions.</div></div>","PeriodicalId":15096,"journal":{"name":"Journal of Atmospheric and Solar-Terrestrial Physics","volume":"279 ","pages":"Article 106715"},"PeriodicalIF":1.9,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145897867","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}