A Scalable, Cloud‐Based Workflow for Spectrally‐Attributed ICESat‐2 Bathymetry With Application to Benthic Habitat Mapping Using Deep Learning

Abstract Since the 2018 launch of NASA's ICESat‐2 satellite, numerous studies have documented the bathymetric measurement capabilities of the space‐based laser altimeter. However, a commonly identified limitation of ICESat‐2 bathymetric point clouds is that they lack accompanying spectral refle...

Full description

Saved in:
Bibliographic Details
Main Authors: Forrest Corcoran, Christopher E. Parrish, Lori A. Magruder, J. P. Swinski
Format: Article
Language:English
Published: American Geophysical Union (AGU) 2024-11-01
Series:Earth and Space Science
Subjects:
Online Access:https://doi.org/10.1029/2024EA003735
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850260257506328576
author Forrest Corcoran
Christopher E. Parrish
Lori A. Magruder
J. P. Swinski
author_facet Forrest Corcoran
Christopher E. Parrish
Lori A. Magruder
J. P. Swinski
author_sort Forrest Corcoran
collection DOAJ
description Abstract Since the 2018 launch of NASA's ICESat‐2 satellite, numerous studies have documented the bathymetric measurement capabilities of the space‐based laser altimeter. However, a commonly identified limitation of ICESat‐2 bathymetric point clouds is that they lack accompanying spectral reflectance attributes, or even intensity values, which have been found useful for benthic habitat mapping with airborne bathymetric lidar. We present a novel method for extracting bathymetry from ICESat‐2 data and automatically adding spectral reflectance values from Sentinel‐2 imagery to each detected bathymetric point. This method, which leverages the cloud computing systems Google Earth Engine and NASA's SlideRule Earth, is ideally suited for “big data” projects with ICESat‐2 data products. To demonstrate the scalability of our workflow, we collected 3,500 ICESat‐2 segments containing approximately 1.4 million spectrally‐attributed bathymetric points. We then used this data set to facilitate training of a deep recurrent neural network for classifying benthic habitats at the ICESat‐2 photon level. We trained two identical models, one with and one without the spectral attributes, to investigate the benefits of fusing ICESat‐2 photons with Sentinel‐2. The results show an improvement in model performance of 18 percentage points, based on F1 score. The procedures and source code are publicly available and will enhance the value of the new ICESat‐2 bathymetry data product, ATL24, which is scheduled for release in Fall 2024. These procedures may also be applicable to data from NASA's upcoming CASALS mission.
format Article
id doaj-art-1f6f6889f08e4e5e9d9761bec0c0df39
institution OA Journals
issn 2333-5084
language English
publishDate 2024-11-01
publisher American Geophysical Union (AGU)
record_format Article
series Earth and Space Science
spelling doaj-art-1f6f6889f08e4e5e9d9761bec0c0df392025-08-20T01:55:41ZengAmerican Geophysical Union (AGU)Earth and Space Science2333-50842024-11-011111n/an/a10.1029/2024EA003735A Scalable, Cloud‐Based Workflow for Spectrally‐Attributed ICESat‐2 Bathymetry With Application to Benthic Habitat Mapping Using Deep LearningForrest Corcoran0Christopher E. Parrish1Lori A. Magruder2J. P. Swinski3School of Civl and Contruction Engineering Oregon State University Corvallis OR USASchool of Civl and Contruction Engineering Oregon State University Corvallis OR USACenter for Space Research University of Texas at Austin Austin TX USANASA Goddard Space Flight Center Greenbelt MD USAAbstract Since the 2018 launch of NASA's ICESat‐2 satellite, numerous studies have documented the bathymetric measurement capabilities of the space‐based laser altimeter. However, a commonly identified limitation of ICESat‐2 bathymetric point clouds is that they lack accompanying spectral reflectance attributes, or even intensity values, which have been found useful for benthic habitat mapping with airborne bathymetric lidar. We present a novel method for extracting bathymetry from ICESat‐2 data and automatically adding spectral reflectance values from Sentinel‐2 imagery to each detected bathymetric point. This method, which leverages the cloud computing systems Google Earth Engine and NASA's SlideRule Earth, is ideally suited for “big data” projects with ICESat‐2 data products. To demonstrate the scalability of our workflow, we collected 3,500 ICESat‐2 segments containing approximately 1.4 million spectrally‐attributed bathymetric points. We then used this data set to facilitate training of a deep recurrent neural network for classifying benthic habitats at the ICESat‐2 photon level. We trained two identical models, one with and one without the spectral attributes, to investigate the benefits of fusing ICESat‐2 photons with Sentinel‐2. The results show an improvement in model performance of 18 percentage points, based on F1 score. The procedures and source code are publicly available and will enhance the value of the new ICESat‐2 bathymetry data product, ATL24, which is scheduled for release in Fall 2024. These procedures may also be applicable to data from NASA's upcoming CASALS mission.https://doi.org/10.1029/2024EA003735ICESat‐2cloud computingbathymetrybenthic habitatsdata fusiondeep learning
spellingShingle Forrest Corcoran
Christopher E. Parrish
Lori A. Magruder
J. P. Swinski
A Scalable, Cloud‐Based Workflow for Spectrally‐Attributed ICESat‐2 Bathymetry With Application to Benthic Habitat Mapping Using Deep Learning
Earth and Space Science
ICESat‐2
cloud computing
bathymetry
benthic habitats
data fusion
deep learning
title A Scalable, Cloud‐Based Workflow for Spectrally‐Attributed ICESat‐2 Bathymetry With Application to Benthic Habitat Mapping Using Deep Learning
title_full A Scalable, Cloud‐Based Workflow for Spectrally‐Attributed ICESat‐2 Bathymetry With Application to Benthic Habitat Mapping Using Deep Learning
title_fullStr A Scalable, Cloud‐Based Workflow for Spectrally‐Attributed ICESat‐2 Bathymetry With Application to Benthic Habitat Mapping Using Deep Learning
title_full_unstemmed A Scalable, Cloud‐Based Workflow for Spectrally‐Attributed ICESat‐2 Bathymetry With Application to Benthic Habitat Mapping Using Deep Learning
title_short A Scalable, Cloud‐Based Workflow for Spectrally‐Attributed ICESat‐2 Bathymetry With Application to Benthic Habitat Mapping Using Deep Learning
title_sort scalable cloud based workflow for spectrally attributed icesat 2 bathymetry with application to benthic habitat mapping using deep learning
topic ICESat‐2
cloud computing
bathymetry
benthic habitats
data fusion
deep learning
url https://doi.org/10.1029/2024EA003735
work_keys_str_mv AT forrestcorcoran ascalablecloudbasedworkflowforspectrallyattributedicesat2bathymetrywithapplicationtobenthichabitatmappingusingdeeplearning
AT christophereparrish ascalablecloudbasedworkflowforspectrallyattributedicesat2bathymetrywithapplicationtobenthichabitatmappingusingdeeplearning
AT loriamagruder ascalablecloudbasedworkflowforspectrallyattributedicesat2bathymetrywithapplicationtobenthichabitatmappingusingdeeplearning
AT jpswinski ascalablecloudbasedworkflowforspectrallyattributedicesat2bathymetrywithapplicationtobenthichabitatmappingusingdeeplearning
AT forrestcorcoran scalablecloudbasedworkflowforspectrallyattributedicesat2bathymetrywithapplicationtobenthichabitatmappingusingdeeplearning
AT christophereparrish scalablecloudbasedworkflowforspectrallyattributedicesat2bathymetrywithapplicationtobenthichabitatmappingusingdeeplearning
AT loriamagruder scalablecloudbasedworkflowforspectrallyattributedicesat2bathymetrywithapplicationtobenthichabitatmappingusingdeeplearning
AT jpswinski scalablecloudbasedworkflowforspectrallyattributedicesat2bathymetrywithapplicationtobenthichabitatmappingusingdeeplearning