Preliminary Insights into Kīlauea 2024–2025 Lava Fountains from Satellite Observations, Machine Learning Approaches, and Ground-Based Validation

Volcanic eruptions are major natural drivers of climate variability and influence atmospheric composition, radiative balance, and regional climate systems. While large explosive eruptions dominate global climatic signals, smaller and more frequent basaltic events can produce regional-scale perturbations, especially where topography and geographic isolation modulate the circulation and residence time of volcanic pollutants. Hawai‘i represents a uniquely suitable natural laboratory for this purpose: its combination of steep relief, persistent trade winds, and insular setting creates conditions in which volcanic SO₂, aerosols, and wind-advected tephra are easy to detect and quantify.

The 2024–2025 Kīlauea lava fountaining episodes offer an opportunity to investigate how moderate explosive basaltic eruptions affect atmospheric composition and short-term regional climate dynamics. We combine multisensor satellite observations (GOES, TROPOMI, Sentinel-2 MSI) with artificial intelligence algorithms to retrieve key eruptive parameters, including Volcanic Radiative Power (VRP), Time-Averaged Discharge Rate (TADR), erupted volume, cloud height, SO₂ mass, and ash dispersion. These satellite-derived measurements are contextualized with local environmental data, including temperature and precipitation, to explore potential short-term atmospheric impacts.

Ground-based observations provide additional constraints, including high-resolution video analysis capturing eruptive precursors, fountain height, plume rise, and ash fall patterns. Video sequences are processed with artificial intelligence-based algorithms to extract time-resolved metrics of eruptive dynamics, generating robust datasets that complement and calibrate satellite measurements.

Here, we present the preliminary dataset and initial observations from this integrated monitoring. The approach yields a high-resolution assessment of lava fountain dynamics, associated gas and aerosol emissions, and localized atmospheric impacts, highlighting the potential of combining satellite data, video analysis, and machine learning-driven processing to improve the monitoring and understanding of volcanic processes in topographically complex island environments. The results lay the groundwork for more comprehensive analyses of future eruptive dynamics and the influence of lava fountaining eruptions on local and regional atmospheric conditions.