Solar flares are often associated with coronal mass ejections (CMEs) that, when directed toward Earth, can disrupt the magnetosphere and impact space weather. In this study, we investigate three long‐duration solar flares observed in 2012 by the POlarization Emission of Millimeter Activity at the Sun (POEMAS) telescopes at 45 and 90 GHz. We analyze the correlation between their radio spectral properties, associated CMEs, and geomagnetic and ionospheric responses, to assess the possibility of space weather prediction. Spectral fitting of the radio data using a gyrosynchrotron model reveals accelerated electron populations with a hard energy distribution (spectral index ) that hardens with time, and magnetic fields of G. The most energetic flare involved a source region 3–6 times larger than the others and produced 10 times more energetic electrons. In all three events, CMEs were launched simultaneously with the flares. The July 12 X1.4 flare was associated with a fast (884 km/s) halo CME that led to a moderate geomagnetic storm (Dst = −139 nT, Kp = 6). VLF radio data show that all three flares increased ionospheric electron density, lowering its boundary from 70 km to 58–67 km. These results suggest that long‐duration flares with high flux at 45 GHz and hard radio spectra are associated with more energetic events characterized by higher GOES class, greater SXR energy, faster and more energetic CMEs, and stronger geomagnetic disturbances, highlighting the potential of high‐frequency radio observations as a tool for space weather prediction.
Ground-based measurements of the atmospheric potential gradient (PG) are sensitive to near-surface electrical conductivity and can reflect anthropogenic perturbations under carefully defined fair-weather (FW) conditions. This study investigates the coupling between PG and major air pollutants in São Paulo, Brazil, using a long-term dataset (2018–2024) combining rooftop electric field measurements with hourly concentrations of CO, SO2, NO, NO2, NOx, and PM10. FW periods were identified using surface meteorological screening and satellite-based cloud masking.The analysis reveals a reproducible hierarchy in PG–pollutant coupling controlled by the dynamical regime of the atmospheric boundary layer. During nocturnal stable conditions, primary combustion-related gases (NOx, NO, CO) exhibit the strongest associations with PG, with median Pearson correlation coefficients (r) on the order of r ≈ 0.6, representing the central tendency of a regime-dependent distribution. In contrast, PM10 shows weaker and more variable coupling. Across all seasons, PG–pollutant correlations strengthen at night and weaken markedly during the afternoon mixing maximum, with enhanced wind ventilation further suppressing coupling strength.Large-scale emission perturbations support the anthropogenic origin of these relationships. During the 2020 COVID-19 lockdown, FW PG exhibits a persistent shift toward lower values relative to the multi-year baseline while preserving the diurnal phase structure. In contrast, a short (11-day) truck drivers’ strike in 2018 produces no statistically distinguishable PG response.These results demonstrate that PG is not a universal proxy for urban air pollutant concentrations, but a regime-dependent electrical indicator whose sensitivity to anthropogenic emissions emerges preferentially under stable boundary-layer conditions in a megacity environment.