The assumption about registration of gamma-ray fluxes from several large groups (“swarms”) of ball lightning by M. Marisaldi et al. [Nature 634, 57 (2024)] during the ALOFT aircraft campaign is presented. This assumption is compatible, in particular, with the report about observation of the large number of fiery balls during the time of about 20 minutes from the airplane flying in the thundercloud with the velocity of about 600 km hour-1. Several parameters of one of the gamma-ray fluxes from tropical thunderclouds detected during ALOFT are explained as those of gamma-ray flux emitted by ball lightning. The ball lightning model explaining these parameters also explains several parameters of two other gamma-ray fluxes, ball lightning lifetimes up to a few minutes, the volume densities of energy of ball lightning up to a few tens kJ cm-3 and some other observational data. The problems related to the possibility and expedience of identification of mechanism of generation of gamma rays in thunderclouds as associated or not associated with ball lightning are considered.
Air pollution is a major environmental and public health concern worldwide, and India is particularly affected, with rapid urbanization, industrial growth, and increasing transportation activities contributing to the complexity of urban air quality challenges. This study evaluates the temporal variability of ambient air pollution and Air Quality Index (AQI) across four major Indian metropolitan cities - Delhi, Patna, Ahmedabad, and Hyderabad during the period 2020 - 2024. Daily air quality data for PM2.5, NO2, SO2, and AQI were obtained from the Central Pollution Control Board (CPCB) and analyzed using Python-based statistical and visualization tools. Seasonal and annual variations were examined to identify temporal patterns and inter-city differences in air quality. The results revealed substantial spatial and temporal variability among the selected cities. Delhi consistently recorded the highest pollutant concentrations and AQI values, followed by Patna, whereas Ahmedabad and Hyderabad generally exhibited comparatively lower pollution levels. PM2.5 concentrations frequently exceeded the National Ambient Air Quality Standards (NAAQS), particularly during winter months. Seasonal analyses showed higher pollutant concentrations during winter and lower concentrations during summer and monsoon periods across all cities. Correlation analysis demonstrated that PM2.5 exhibited the strongest positive association with AQI, while NO2 and SO2 showed comparatively weaker relationships. Unlike previous city-specific studies, this research provides a comparative assessment of pollutant behaviour and AQI variability across four geographically distinct Indian metropolitan regions using a consistent five-year dataset. The findings contribute to a better understanding of urban air quality dynamics and may support evidence-based environmental management, urban planning, and air pollution mitigation strategies.
This study examines the intraseasonal modulation of June–September (JJAS) rainfall over Northeast India (NEI) by the Madden–Julian Oscillation (MJO) and evaluates its possible conditioning by the El Niño–Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD) during the common complete analysis period 1979–2023. Phase-dependent variations in mean rainfall and extreme-rainfall probability, defined using the pooled JJAS 95th-percentile threshold of 29.33 mm day-1, are quantified for dynamically active MJO days with RMM amplitude ≥1. Mean rainfall and extreme-rainfall probability are highest during MJO Phases 1–2 and lowest primarily during Phases 4–5, with suppression extending into Phase 6 in some diagnostics. Lagged composites indicate that extreme-rainfall probability reaches its maximum during Phase 2 at a lag of approximately three days, suggesting delayed moisture accumulation and convective organization. Continuous wavelet power, cross-wavelet power and wavelet coherence reveal intermittent rather than persistent variability within the canonical 30–60-day MJO band. ENSO- and IOD-conditioned composites show descriptive differences in the magnitude and phase-wise expression of rainfall; however, no phase-specific climate-state contrast remains statistically significant after false-discovery-rate correction. Annual MJO-band coherence is descriptively higher during La Niña than during Neutral and El Niño conditions, although the overall ENSO comparison is not statistically significant. Combined ENSO–IOD distributions are interpreted descriptively because the categories are highly unequal and some contain only one or two years, preventing reliable statistical inference. The results demonstrate a statistically supported phase- and lag-dependent association between the MJO and NEI rainfall while indicating that the apparent conditioning by ENSO and IOD remains descriptive. These findings may support subseasonal identification of rainfall-risk windows over Northeast India.
The 2017 Great American Solar Eclipse on 21 August provided a unique opportunity for citizen science meteorological research. The Global Learning and Observations to Benefit the Environment (GLOBE) Program leveraged this occasion to initiate the Eclipse Across America (EAA) campaign to collect meteorological data through citizen science. This research analyzed the comprehensive dataset garnered by the EAA, exploring the correlation between the eclipse-induced temperature depression and prevailing cloud cover conditions.The EAA campaign under GLOBE collected over 80,000 temperature observations and 20,000 cloud cover data points, facilitating an in-depth evaluation of the eclipse's influence on local meteorological conditions. Comprehensive data validation confirmed the robustness of the citizen-science data. Specifically, air temperature measurements from GLOBE's collection closely aligned with National Weather Service station figures and GLOBE expert data, with correlation values of 0.78 and 0.73, respectively, and the majority of discrepancies ranged from 0 to 2°C. However, cloud data interpretation posed some difficulties, particularly for newer GLOBE contributors. When assessing cloud data, these newer participants more accurately identified low and high cloud types but had some challenges discerning specific cloud cover classifications.Moreover, this research explored changes in air temperature and cloud patterns during the eclipse. A notable temperature depression was found and was more pronounced in areas with greater eclipse coverage. Detailed analysis also showcased that regions with less cloud cover underwent more pronounced temperature reductions, corroborating multiple prior studies. The investigation observed temperature decreases of up to 5°C, contingent on cloud cover and geographical location. This study underscored the efficacy of citizen science in capturing meteorological variations during unique astronomical events. Insights from this study and the GLOBE 2017 EAA campaign establish a robust methodological framework for future astronomical events, demonstrating how high-density, citizen-collected data can effectively capture localized meteorological phenomena.
Cold air outbreaks over north India exert profound impacts on human health, agriculture, and energy demand. This study investigates the characteristics, duration-dependent dynamics, and trends of cold wave (CW) events across different sub-regions of north India using India's first high-resolution regional reanalysis, the Indian Monsoon Data Assimilation and Analysis (IMDAA), for the winters of 1979–2019. The analysis represents one of the first attempts to systematically examine CW behaviour across sub-regions of north India using the high-resolution IMDAA dataset (0.12° × 0.12°). To better understand the evolution of CWs, events are classified based on their duration into short-duration (SCW), medium-duration (MCW) and long-duration (LCW) categories. The results reveal distinct dynamical characteristics associated with each category. SCWs are primarily confined to northwestern India and are linked to transient intrusions of cold northwesterly winds following the passage of western disturbances. MCWs exhibit a broader spatial influence, extending eastward into central India, indicating sustained cold air advection and continued surface cooling. In contrast, LCWs are characterized by widespread temperature anomalies across the Indo-Gangetic Plain and are associated with persistent northwesterly flow, strengthened surface anticyclones, and enhanced baroclinic instability that favour prolonged cold air advection and radiative cooling under clear-sky conditions. The results indicate that consistent patterns are observed across the ERA5, IMD, and IMDAA. The high-resolution IMDAA dataset provides an improved representation of regional temperature variability and helps delineate the spatial evolution of CW events across India. These findings highlight the importance of duration-dependent dynamics in controlling the evolution and spatial extent of CW events and demonstrate the potential of IMDAA for improving regional cold wave monitoring and early-warning systems over the Indian subcontinent.
Accurate wind speed forecasting remains a challenging task because atmospheric processes exhibit highly nonlinear, stochastic and region-dependent characteristics. Rather than proposing another increasingly complex deep learning (DL) architecture, this study presents a feature engineering-driven forecasting framework that systematically integrates cyclic encoding, differential features and moving average smoothing with deep learning models to improve forecasting accuracy while maintaining computational efficiency. Five representative models, namely CNN, LSTM, stacked BiLSTM, FE-GRU and a CNN–BiLSTM–Linear stacking ensemble, were comparatively evaluated under identical experimental conditions using a 13-year (2010–2022) hourly meteorological dataset collected from three climatically distinct regions of Türkiye (Çanakkale, Niğde and Van), each containing more than 100,000 observations. Experimental results consistently demonstrate that the proposed FE-GRU model achieved the lowest relative prediction errors across all study regions when evaluated using Mean Absolute Percentage Error (MAPE), which was adopted as the primary performance criterion. The average MAPE values obtained for Çanakkale, Niğde and Van were 1.20%, 1.31% and 1.72%, respectively, while maintaining high goodness-of-fit (R2 ≥ 0.995) and low Root Mean Square Error (RMSE) values. Although the CNN–BiLSTM–Linear stacking ensemble produced competitive RMSE values, it required substantially higher computational cost, whereas FE-GRU achieved a superior balance between forecasting accuracy and computational efficiency. The reliability of the proposed framework was further supported through paired statistical significance tests, confidence interval analysis, Taylor diagrams and feature importance evaluation, demonstrating consistent agreement with observed wind speed measurements under diverse geographical and climatic conditions. Overall, the results indicate that explicit feature engineering contributes more effectively to forecasting performance than increasing architectural complexity alone. The proposed framework therefore provides an accurate, computationally efficient and geographically robust solution for practical wind speed forecasting and renewable energy management applications.
Water vapor and ozone are crucial trace gases in the atmosphere, playing a vital role in climate change within the upper troposphere and lower stratosphere (UTLS). Understanding their behavior is particularly important for studying atmospheric dynamics and predicting tropical cyclones. In this study, we analyzed the vertical profiles of ozone and water vapor variation in the UTLS of both tropical Ethiopia and subtropical Egypt regions using AURA/MLS satellite data from 2006 to 2020. In general, with the exception of spring in the tropical region of Ethiopia, the ozone distribution in the UTLS tends to be quite consistent. In the subtropical region of Egypt, ozone concentrations remain steady during the autumn months. However, in winter, the levels of tropospheric ozone in tropical Egypt are typically lower than those observed in the summer and autumn. The fluctuation of the annual mean ozone volume mixing ratio (OVMR) in subtropical Egypt varies between 6 to 9 ppmv in the UTLS, within a pressure range from 50 to 0.001 hPa. In tropical Ethiopia, the variability is observed to be between 6 to 10 ppmv across the pressure levels 50 to 0.001 hPa. The annual fluctuations of average water vapor mixing ratio (WVMR) in tropical region Ethiopia reach peak values of 400 to 600 ppmv in the UTLS between the pressure levels of 200 to 300 hPa, typically occurring during May, Jun, July, and August. Conversely, subtropical Region Egypt records peak WVMR values of 100 to 200 ppmv in the UTLS, within the pressure range of 200 to 300 hPa, generally from March to June. Our study indicates that Egypt experiences a greater ozone mass flux at the tropopause than Ethiopia. In the tropical region Ethiopia, we detect the Semi Annual Oscillation (SAO) in the WVMR, while the Annual Oscillation (AO) is found in the OVMR. Conversely, in subtropical region Egypt, the Tri Annual Oscillation (TAO) is seen in the WVMR, with the AO present in the OMR. The transport of ozone from the stratosphere to the troposphere is controlled by the tropical CPTh, while the transport of water vapor across the tropopause is controlled by the tropical CPTt. Over a period of 15 years, the maximum mass flux of ozone crossing the tropopause is 5 kg/s (equivalent to 6.57 × 105 kg of ozone) in Egypt and 3.4 kg/s (equivalent to 4.47 × 105 kg of ozone) in Ethiopia. In terms of water vapor, the maximum mass flux crossing the tropopause is 68 kg/s (equivalent to 8.94 × 106 kg of water vapor) in tropical Ethiopia and 30 kg/s (equivalent to 3.94 × 106 kg of water vapor) in subtropical Egypt. Our study reveals that Egypt experiences a higher ozone mass flux in the tropopause compared to Ethiopia. Conversely, Ethiopia experiences a higher water vapor mass flux in the tropopause compared to Egypt.
This study is the first report of Doppler shifted hydrogen Lyman-alpha (Ly-α) emissions associated with proton precipitation observed by the global ultraviolet imager (GUVI) instrument on board NASA's Thermosphere, Ionosphere, Mesosphere Energetics and Dynamics (TIMED) mission. GUVI is a scanning-type imaging spectrograph and its main purpose is to provide disc (horizon-to-horizon) images of auroral intensity at 5 major far ultraviolet wavelengths along the spacecraft track. GUVI can also operate at the spectrograph mode, by which the scan mirror is held at a fixed viewing angle and the entire spectrum (∼1150–1800 Å in 176 wavelength bins) is down-linked, allowing detailed spectral analysis. Due to the wide spectral width (FWHM ∼13.9 Å) for the instrument response, attempts for spectral data analysis has not been reported. In this paper, we will address the proton aurora by describing a procedure for extracting the hydrogen Ly-α from the spectral data. We show evidence of proton aurora – redshift Ly-α spectrum, when TIMED traversed the auroral zone, in particular during geomagnetic storm times. We also derive an empirical formula that relates the wavelength shift to the average energy of precipitating ions observed by the SSJ sensor on board the Defense Meteorological Satellite Program.
Solar and wind energy are important renewable sources in the global energy mix and play a key role in the energy transition and in meeting international climate targets. The Brazilian semiarid region is a strategic area because of its abundant wind and solar resources and its high potential for sustainable energy generation. This study provides an integrated analysis of spatial trends, temporal dependence, and the complexity of mean wind-speed and mean global solar-radiation series in 27 municipalities in the Brazilian semiarid region. ERA5-Land reanalysis data for 1981–2024 were used for the regional trend analysis, whereas observational data from the Brazilian National Institute of Meteorology (INMET) for 2009–2024 supported local analyses of autocorrelation, entropy, and cross-correlation across multiple time scales. The results indicated statistically significant positive trends for both variables, with greater spatial homogeneity for solar radiation and greater heterogeneity for wind speed. The series exhibited long-range persistence, with αDFA exponents close to or greater than 1 in some municipalities, indicating temporal memory. The entropy measures revealed scale-dependent behavior: greater regularity was observed at short scales (n<30 days); irregularity progressively increased at intermediate scales (30250 days). The cross-correlation analysis, estimated using the ρDCCA coefficient, revealed a predominance of positive associations between wind speed and solar radiation, which intensified at larger time scales. The integration of 44 years of regional screening with local multiscale diagnoses based on 16 years of observations enabled the joint characterization of spatial trends, long-range persistence, temporal complexity, and cross-correlation between wind and solar resources. The results broaden the understanding of the spatiotemporal variability of these resources in the Brazilian semiarid region and provide support for regional climate studies and renewable-energy system planning.
As an integral constituent of Earth's hydrosphere, precipitable water vapor (PWV) intricately influences the dynamics of the planet's water cycle and energy exchange processes. The ERA5 provided by European Centre for Medium-Range Weather Forecasts (ECMWF), presenting a global meteorological reanalysis data with high temporal and spatial resolution, which is extensively employed for PWV retrieval. However, current research often lacks a detailed description of how to calculate PWV from ERA5, and even less discussion on the impact of different methods on the accuracy of ERA5-derived PWV. Therefore, this study systematically assesses the accuracy of ERA5-derived PWV obtained using different calculation pathways, which differed by different combinations of interpolation and methods, namely PWVa, PWVb, PWVc and PWVd. Hourly GNSS-derived PWV data obtained from 255 Crustal Movement Observation Network of China (CMONOC) stations spanning the period from 2016 to 2019 serves as the reference for the evaluation, and the root mean square error (RMSE) and mean absolute error (MAE) were selected as the criteria. The results show a robust correlation among the four types of PWV, particularly between PWVa and PWVc, as well as PWVb and PWVd, with correlation coefficients both reaching as high as 1. The RMSE for PWVb and PWVd concentrate between 1 and 3 mm, while the concentration is between 2 and 4 mm for PWVa and PWVc. Moreover, the PWVs exhibit pronounced regional disparities, particularly in northwest China, where the RMSE and MAE values at most stations are both below 2 mm. The distinctions among the four types of PWV exhibit an augmented trend with increasing latitudinal values. The accuracy of the four types of PWV improved as the elevation increases, and it is worth noting that the distinction in accuracy among these four types of PWV also decreases in this situation. These findings highlight the relatively minor influence of interpolation schemes on performance compared to the more pronounced impact attributed to calculation methods. This insight aids users in gaining understanding of the strengths and limitations in these methods, thereby advancing research in climate change.
