VSTAR-G: Cloud-Based Tool for Global Detection and Quantification of Volcanic SO2 Plumes Using TROPOMI and Machine Learning Approach

SO2 emissions are among the primary indicators of magmatic activity and play a fundamental role both in local-scale risk management, including air quality and aviation safety, and in regional-to-global climate processes through the formation of sulfate aerosols capable of altering the Earth’s radiative balance. In recent years, satellite observations have revolutionized the monitoring of volcanic emissions. In particular, the Sentinel-5P satellite, part of the Copernicus programme, carries the TROPOMI spectrometer, which provides high-resolution observations for monitoring of atmospheric trace gases (e.g., SO2). Here, we present VSTAR-G, a Google Earth Engine (GEE) application that operationalizes an algorithm previously developed. The application is able to visualize and analyse, using an algorithm previously developed, the automatic detection of volcanic SO2 plumes and their mass quantification according to plume altitude. Through an intuitive interface, users can interactively explore SO2 emissions from more than 200 volcanoes worldwide, define spatial and temporal regions of interest (ROIs), and compare different dates or plume-height products using a synchronized split-window visualization. The application also generates time series of the SO2 mass within the selected area, supports the simultaneous analysis of multiple plumes, and allows users to export plume-detection maps in GeoTIFF format and the associated satellite data as CSV files. The developed application transforms an existing detection algorithm into an accessible operational tool, combining satellite observations, machine learning, and cloud computing to support scalable and near-real-time monitoring of volcanic SO2 emissions.