Compensating for penetration in radar altimetry cryosphere elevation estimates using deep learning
Radar altimetry is commonly used for monitoring changes within the cryosphere and its contribution to sea-level rise. Recent advances in Swath altimetry processing, using the interferometric mode of CryoSat-2, have enabled fine 500m spatial resolution surface elevation models from each satellite pas...
Main Author: | |
---|---|
Other Authors: | |
Format: | Master Thesis |
Language: | English |
Published: |
The University of Edinburgh
2018
|
Subjects: | |
Online Access: | http://hdl.handle.net/1842/35458 |
id |
ftunivedinburgh:oai:era.ed.ac.uk:1842/35458 |
---|---|
record_format |
openpolar |
spelling |
ftunivedinburgh:oai:era.ed.ac.uk:1842/35458 2023-07-30T04:03:54+02:00 Compensating for penetration in radar altimetry cryosphere elevation estimates using deep learning Ewart, Martin Gourmelen, Noel 29/11/2018 application/pdf http://hdl.handle.net/1842/35458 en eng The University of Edinburgh http://hdl.handle.net/1842/35458 Radar altimetry CryoSat-2 Greenland Swath processing Interferometry Cryosphere Ice-sheet Glaciers DEM Surface elevation change Climate change Sea level change Operation IceBridge IceSat-2 Neural network Machine learning Deep learning MSc Geographical Information Science GIS Thesis or Dissertation Masters MSc Master of Science 2018 ftunivedinburgh 2023-07-09T20:31:34Z Radar altimetry is commonly used for monitoring changes within the cryosphere and its contribution to sea-level rise. Recent advances in Swath altimetry processing, using the interferometric mode of CryoSat-2, have enabled fine 500m spatial resolution surface elevation models from each satellite pass. However, there is variability in radar elevation estimates, often due to the penetration of radar waves into snow and firn, yielding differences compared to local airborne LiDAR altimeters which are less impacted by penetration. While neural networks are increasingly being used in a wide variety of domains to enhance predictions and detect changes, they have not previously been applied to radar altimetry to correct for elevation biases within the cryosphere. In this study, we present a novel approach to adjusting for elevation bias by creating a neural network that is trained to predict the elevation provided by local airborne LiDAR from CryoSat-2 radar elevation data where both data are available and then apply that model where only radar data are present. We investigate the challenges with building such models and review the variety of configurations and considerations. Finally, we present two proof of concepts that show good spatial and temporal transfer ability and compensate for 70-90% of the mean penetration, while reducing the root mean squared error by 10-17%. Master Thesis Greenland Ice Sheet Edinburgh Research Archive (ERA - University of Edinburgh) Greenland |
institution |
Open Polar |
collection |
Edinburgh Research Archive (ERA - University of Edinburgh) |
op_collection_id |
ftunivedinburgh |
language |
English |
topic |
Radar altimetry CryoSat-2 Greenland Swath processing Interferometry Cryosphere Ice-sheet Glaciers DEM Surface elevation change Climate change Sea level change Operation IceBridge IceSat-2 Neural network Machine learning Deep learning MSc Geographical Information Science GIS |
spellingShingle |
Radar altimetry CryoSat-2 Greenland Swath processing Interferometry Cryosphere Ice-sheet Glaciers DEM Surface elevation change Climate change Sea level change Operation IceBridge IceSat-2 Neural network Machine learning Deep learning MSc Geographical Information Science GIS Ewart, Martin Compensating for penetration in radar altimetry cryosphere elevation estimates using deep learning |
topic_facet |
Radar altimetry CryoSat-2 Greenland Swath processing Interferometry Cryosphere Ice-sheet Glaciers DEM Surface elevation change Climate change Sea level change Operation IceBridge IceSat-2 Neural network Machine learning Deep learning MSc Geographical Information Science GIS |
description |
Radar altimetry is commonly used for monitoring changes within the cryosphere and its contribution to sea-level rise. Recent advances in Swath altimetry processing, using the interferometric mode of CryoSat-2, have enabled fine 500m spatial resolution surface elevation models from each satellite pass. However, there is variability in radar elevation estimates, often due to the penetration of radar waves into snow and firn, yielding differences compared to local airborne LiDAR altimeters which are less impacted by penetration. While neural networks are increasingly being used in a wide variety of domains to enhance predictions and detect changes, they have not previously been applied to radar altimetry to correct for elevation biases within the cryosphere. In this study, we present a novel approach to adjusting for elevation bias by creating a neural network that is trained to predict the elevation provided by local airborne LiDAR from CryoSat-2 radar elevation data where both data are available and then apply that model where only radar data are present. We investigate the challenges with building such models and review the variety of configurations and considerations. Finally, we present two proof of concepts that show good spatial and temporal transfer ability and compensate for 70-90% of the mean penetration, while reducing the root mean squared error by 10-17%. |
author2 |
Gourmelen, Noel |
format |
Master Thesis |
author |
Ewart, Martin |
author_facet |
Ewart, Martin |
author_sort |
Ewart, Martin |
title |
Compensating for penetration in radar altimetry cryosphere elevation estimates using deep learning |
title_short |
Compensating for penetration in radar altimetry cryosphere elevation estimates using deep learning |
title_full |
Compensating for penetration in radar altimetry cryosphere elevation estimates using deep learning |
title_fullStr |
Compensating for penetration in radar altimetry cryosphere elevation estimates using deep learning |
title_full_unstemmed |
Compensating for penetration in radar altimetry cryosphere elevation estimates using deep learning |
title_sort |
compensating for penetration in radar altimetry cryosphere elevation estimates using deep learning |
publisher |
The University of Edinburgh |
publishDate |
2018 |
url |
http://hdl.handle.net/1842/35458 |
geographic |
Greenland |
geographic_facet |
Greenland |
genre |
Greenland Ice Sheet |
genre_facet |
Greenland Ice Sheet |
op_relation |
http://hdl.handle.net/1842/35458 |
_version_ |
1772815036934258688 |