DeepBedMap: a deep neural network for resolving the bed topography of Antarctica

<p>To resolve the bed elevation of Antarctica, we present DeepBedMap – a novel machine learning method that can produce Antarctic bed topography with adequate surface roughness from multiple remote sensing data inputs. The super-resolution deep convolutional neural network model is trained on...

Full description

Bibliographic Details
Main Authors: W. J. Leong, H. J. Horgan
Format: Article
Language:English
Published: Copernicus Publications 2020-11-01
Series:The Cryosphere
Online Access:https://tc.copernicus.org/articles/14/3687/2020/tc-14-3687-2020.pdf
_version_ 1819086087935492096
author W. J. Leong
H. J. Horgan
author_facet W. J. Leong
H. J. Horgan
author_sort W. J. Leong
collection DOAJ
description <p>To resolve the bed elevation of Antarctica, we present DeepBedMap – a novel machine learning method that can produce Antarctic bed topography with adequate surface roughness from multiple remote sensing data inputs. The super-resolution deep convolutional neural network model is trained on scattered regions in Antarctica where high-resolution (250&thinsp;<span class="inline-formula">m</span>) ground-truth bed elevation grids are available. This model is then used to generate high-resolution bed topography in less surveyed areas. DeepBedMap improves on previous interpolation methods by not restricting itself to a low-spatial-resolution (1000&thinsp;<span class="inline-formula">m</span>) BEDMAP2 raster image as its prior image. It takes in additional high-spatial-resolution datasets, such as ice surface elevation, velocity and snow accumulation, to better inform the bed topography even in the absence of ice thickness data from direct ice-penetrating-radar surveys. The DeepBedMap model is based on an adapted architecture of the Enhanced Super-Resolution Generative Adversarial Network, chosen to minimize per-pixel elevation errors while producing realistic topography. The final product is a four-times-upsampled (250&thinsp;<span class="inline-formula">m</span>) bed elevation model of Antarctica that can be used by glaciologists interested in the subglacial terrain and by ice sheet modellers wanting to run catchment- or continent-scale ice sheet model simulations. We show that DeepBedMap offers a rougher topographic profile compared to the standard bicubically interpolated BEDMAP2 and BedMachine Antarctica and envision it being used where a high-resolution bed elevation model is required.</p>
first_indexed 2024-12-21T21:14:41Z
format Article
id doaj.art-3c64fbbe2f9a4386942da242a12c5b37
institution Directory Open Access Journal
issn 1994-0416
1994-0424
language English
last_indexed 2024-12-21T21:14:41Z
publishDate 2020-11-01
publisher Copernicus Publications
record_format Article
series The Cryosphere
spelling doaj.art-3c64fbbe2f9a4386942da242a12c5b372022-12-21T18:50:03ZengCopernicus PublicationsThe Cryosphere1994-04161994-04242020-11-01143687370510.5194/tc-14-3687-2020DeepBedMap: a deep neural network for resolving the bed topography of AntarcticaW. J. LeongH. J. Horgan<p>To resolve the bed elevation of Antarctica, we present DeepBedMap – a novel machine learning method that can produce Antarctic bed topography with adequate surface roughness from multiple remote sensing data inputs. The super-resolution deep convolutional neural network model is trained on scattered regions in Antarctica where high-resolution (250&thinsp;<span class="inline-formula">m</span>) ground-truth bed elevation grids are available. This model is then used to generate high-resolution bed topography in less surveyed areas. DeepBedMap improves on previous interpolation methods by not restricting itself to a low-spatial-resolution (1000&thinsp;<span class="inline-formula">m</span>) BEDMAP2 raster image as its prior image. It takes in additional high-spatial-resolution datasets, such as ice surface elevation, velocity and snow accumulation, to better inform the bed topography even in the absence of ice thickness data from direct ice-penetrating-radar surveys. The DeepBedMap model is based on an adapted architecture of the Enhanced Super-Resolution Generative Adversarial Network, chosen to minimize per-pixel elevation errors while producing realistic topography. The final product is a four-times-upsampled (250&thinsp;<span class="inline-formula">m</span>) bed elevation model of Antarctica that can be used by glaciologists interested in the subglacial terrain and by ice sheet modellers wanting to run catchment- or continent-scale ice sheet model simulations. We show that DeepBedMap offers a rougher topographic profile compared to the standard bicubically interpolated BEDMAP2 and BedMachine Antarctica and envision it being used where a high-resolution bed elevation model is required.</p>https://tc.copernicus.org/articles/14/3687/2020/tc-14-3687-2020.pdf
spellingShingle W. J. Leong
H. J. Horgan
DeepBedMap: a deep neural network for resolving the bed topography of Antarctica
The Cryosphere
title DeepBedMap: a deep neural network for resolving the bed topography of Antarctica
title_full DeepBedMap: a deep neural network for resolving the bed topography of Antarctica
title_fullStr DeepBedMap: a deep neural network for resolving the bed topography of Antarctica
title_full_unstemmed DeepBedMap: a deep neural network for resolving the bed topography of Antarctica
title_short DeepBedMap: a deep neural network for resolving the bed topography of Antarctica
title_sort deepbedmap a deep neural network for resolving the bed topography of antarctica
url https://tc.copernicus.org/articles/14/3687/2020/tc-14-3687-2020.pdf
work_keys_str_mv AT wjleong deepbedmapadeepneuralnetworkforresolvingthebedtopographyofantarctica
AT hjhorgan deepbedmapadeepneuralnetworkforresolvingthebedtopographyofantarctica