elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Imprint | Privacy Policy | Contact | Deutsch
Fontsize: [-] Text [+]

Volcanic SO2 Effective Layer Height Retrieval with OMI Using a Machine Learning Driven Approach

Fedkin, Nikita and Li, Can and Krotkov, Nickolay and Loyola, Diego and Hedelt, Pascal (2020) Volcanic SO2 Effective Layer Height Retrieval with OMI Using a Machine Learning Driven Approach. AGU Fall Meeting 2020, 2020-12-01 - 2020-12-17, San Francisco, USA / Online.

Full text not available from this repository.

Abstract

Information about the height and loading of aerosols and sulfur dioxide (SO2) plumes from volcanic eruptions is crucial for both aviation safety and for assessing the effect of volcanic sulfate aerosols on climate. Previously, a few retrieval techniques for the SO2 layer height using backscattered Earthshine ultraviolet (BUV) radiances have been demonstrated. However, these techniques often rely on time consuming direct spectral fitting methods and on-line radiative transfer calculations, and as a result are mostly not suited for near real time applications. Here, we introduce a new machine learning based algorithm for fast retrievals of effective volcanic SO2 layer height (SO2 LH) from the Ozone Monitoring Instrument (OMI) . The first part of this method is a training phase that utilizes extensive radiative transfer calculations to generate a large dataset of synthetic BUV spectra in the 310-330 nm spectral range. The principal components of this dataset, in addition to several key physical parameters, are used to train a feed-forward neural network to predict the SO2LH. This is followed by the application phase, where the trained inverse model is used on real OMI BUV measurements to retrieve the effective SO2 LH. The algorithm has been tested on four major explosive eruptions during the OMI data record. Results for the 2008 Kasatochi, 2019 Raikoke, 2015 Calbuco and 2014 Kelud eruptions are presented here and compared with volcanic plume heights estimated with other satellite sensors. For the most part, OMI-retrieved effective SO2 LHs agree well with the lidar measurements of co-located volcanic sulfate aerosol layer height from Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) and thermal infrared retrievals of SO2 LH from the infrared atmospheric sounding interferometer (IASI). The errors in OMI retrieved SO2 heights are estimated to be in the 1.5-2 km range for plumes with relatively large SO2 column densities exceeding ~40 Dobson Units (1DU= 2.69 x 1016 molecules SO2 cm-2). In the application phase the retrieval of plume height is highly efficient, and takes less than 3 minutes for a full OMI orbit. This approach and similar machine learning based applications can also be readily adapted for other satellite UV-Vis spectrometers.

Item URL in elib:https://elib.dlr.de/137718/
Document Type:Conference or Workshop Item (Poster)
Title:Volcanic SO2 Effective Layer Height Retrieval with OMI Using a Machine Learning Driven Approach
Authors:
AuthorsInstitution or Email of AuthorsAuthor's ORCID iDORCID Put Code
Fedkin, NikitaUniversity of Maryland College ParkUNSPECIFIEDUNSPECIFIED
Li, CanNASA Goddard Space Flight Center, Greenbelt, MD 20771, USAUNSPECIFIEDUNSPECIFIED
Krotkov, NickolayUniv. of Maryland at Baltimore County, Baltimore, MD USAhttps://orcid.org/0000-0001-6170-6750UNSPECIFIED
Loyola, DiegoUNSPECIFIEDhttps://orcid.org/0000-0002-8547-9350UNSPECIFIED
Hedelt, PascalUNSPECIFIEDhttps://orcid.org/0000-0002-1752-0040UNSPECIFIED
Date:2020
Refereed publication:No
Open Access:No
Gold Open Access:No
In SCOPUS:No
In ISI Web of Science:No
Status:Published
Keywords:SO2, Machine Learning, Volcanoes, OMI, SO2 layer height
Event Title:AGU Fall Meeting 2020
Event Location:San Francisco, USA / Online
Event Type:international Conference
Event Start Date:1 December 2020
Event End Date:17 December 2020
Organizer:AGU
HGF - Research field:Aeronautics, Space and Transport
HGF - Program:Space
HGF - Program Themes:Earth Observation
DLR - Research area:Raumfahrt
DLR - Program:R EO - Earth Observation
DLR - Research theme (Project):Vorhaben Spectroscopic Methods in Remote Sensing (old)
Location: Oberpfaffenhofen
Institutes and Institutions:Remote Sensing Technology Institute > Atmospheric Processors
Deposited By: Hedelt, Pascal
Deposited On:19 Nov 2020 09:16
Last Modified:24 Apr 2024 20:39

Repository Staff Only: item control page

Browse
Search
Help & Contact
Information
electronic library is running on EPrints 3.3.12
Website and database design: Copyright © German Aerospace Center (DLR). All rights reserved.