SO2 plume height retrieval: Applying inverse learning machines to S5p / TROPOMI SO2 data

Abstract

Precise knowledge of the location and height of the volcanic sulfur dioxide (SO2) plume is essential for accurate determination of SO2 emitted by volcanic eruptions. Current UV based SO2 plume height retrieval algorithms are very time-consuming and therefore not suitable for near-real-time applications. Here we present a novel method called the Full-Physics Inverse Learning Machine (FP-ILM) algorithm for extremely fast and accurate retrieval of the SO2 plume and apply it to the pre-operational Sentinel-5 Precursor (S5P). S5P was launched on October 13, 2017 carrying the TROPOspheric Monitoring Instrument (TROPOMI), which has a spatial resolution of 7x3.5 km², hence providing an unprecedented level of details. In this presentation, we introduce the FP-ILM algorithm, which is based on dimensionality reduction techniques and machine learning. We show the first results obtained with the FP-ILM algorithm applied to a selection of volcanic SO2 eruptions detected by TROPOMI and compare them to other plume height datasets available. The sensitivity of the plume height retrieval to various parameters is analyzed as well.