Deep Learning-Aided High-Resolution Temporal Gravity Field Simulations: Monthly global mass grids and Spherical Harmonics from 1994 to 2021
Starting from April 2002, Gravity Recovery and Climate Experiment (GRACE) and, its successor mission GRACE-FO (FollowOn) have provided irreplaceable data for monitoring mass variations within the hydrosphere, cryosphere, and oceans, and with unprecedented accuracy and resolution for over two decades...
Main Authors: | , , , , , |
---|---|
Format: | Conference Object |
Language: | English |
Published: |
2023
|
Subjects: | |
Online Access: | https://avesis.yildiz.edu.tr/publication/details/f558041f-991a-4c31-853f-4e9e3f3ac534/oai |
id |
ftyildiztuniv:f558041f-991a-4c31-853f-4e9e3f3ac534 |
---|---|
record_format |
openpolar |
spelling |
ftyildiztuniv:f558041f-991a-4c31-853f-4e9e3f3ac534 2023-09-26T15:18:09+02:00 Deep Learning-Aided High-Resolution Temporal Gravity Field Simulations: Monthly global mass grids and Spherical Harmonics from 1994 to 2021 Atman, Kazım Gökhan Olgun, Sevda Güneş, Özge Uz, Metehan Akyılmaz, Orhan Shum, C.K. 2023-07-15T00:00:00Z https://avesis.yildiz.edu.tr/publication/details/f558041f-991a-4c31-853f-4e9e3f3ac534/oai eng eng f558041f-991a-4c31-853f-4e9e3f3ac534 https://avesis.yildiz.edu.tr/publication/details/f558041f-991a-4c31-853f-4e9e3f3ac534/oai info:eu-repo/semantics/closedAccess info:eu-repo/semantics/conferenceObject 2023 ftyildiztuniv 2023-08-27T21:20:01Z Starting from April 2002, Gravity Recovery and Climate Experiment (GRACE) and, its successor mission GRACE-FO (FollowOn) have provided irreplaceable data for monitoring mass variations within the hydrosphere, cryosphere, and oceans, and with unprecedented accuracy and resolution for over two decades. However, the long-term products of mass variations prior to GRACE-era may allow for a better understanding of spatiotemporal changes in climate-induced geophysical phenomena, including terrestrial water cycle, ice sheet and glacier mass balance, sea level change and ocean bottom pressure. In this study, total water storage anomalies (TWSA) are simulated/reconstructed globally at 1.0°x1.0° spatial and monthly temporal resolutions from January 1994 to December 2020 with an in-house developed hybrid Deep Learning (DL) architecture using GRACE/-FO mascon and SLR gravimetry, ECMWF Reanalysis-5 (ERA5) data. We validated our simulated mass change data products both over land and ocean, not only through mathematical performance metrics (internal validations) such as RMSE or NSE along with comparisons to previous studies, but also external validations with non-GRACE datasets such as El-Nino and La-Nina patterns, barystatic global mean sea level change, degree (d) 2 order (o) 1 spherical harmonic coefficients (C21, S21) retrieved from Earth orientation parameters, Greenland Ice sheet mass balance and in-situ Ocean Bottom Pressure measurements were carried out. The overall validations show that the proposed DL paradigm can efficiently simulate high-resolution monthly global gravity field both in GRACE/GRACE-FO and pre-GRACE era. The resulting simulated data products are available as monthly mass change grids as well as spherical harmonic models up to d/o 180. Conference Object glacier Greenland Ice Sheet Yıldız Technical University Research Information System Greenland |
institution |
Open Polar |
collection |
Yıldız Technical University Research Information System |
op_collection_id |
ftyildiztuniv |
language |
English |
description |
Starting from April 2002, Gravity Recovery and Climate Experiment (GRACE) and, its successor mission GRACE-FO (FollowOn) have provided irreplaceable data for monitoring mass variations within the hydrosphere, cryosphere, and oceans, and with unprecedented accuracy and resolution for over two decades. However, the long-term products of mass variations prior to GRACE-era may allow for a better understanding of spatiotemporal changes in climate-induced geophysical phenomena, including terrestrial water cycle, ice sheet and glacier mass balance, sea level change and ocean bottom pressure. In this study, total water storage anomalies (TWSA) are simulated/reconstructed globally at 1.0°x1.0° spatial and monthly temporal resolutions from January 1994 to December 2020 with an in-house developed hybrid Deep Learning (DL) architecture using GRACE/-FO mascon and SLR gravimetry, ECMWF Reanalysis-5 (ERA5) data. We validated our simulated mass change data products both over land and ocean, not only through mathematical performance metrics (internal validations) such as RMSE or NSE along with comparisons to previous studies, but also external validations with non-GRACE datasets such as El-Nino and La-Nina patterns, barystatic global mean sea level change, degree (d) 2 order (o) 1 spherical harmonic coefficients (C21, S21) retrieved from Earth orientation parameters, Greenland Ice sheet mass balance and in-situ Ocean Bottom Pressure measurements were carried out. The overall validations show that the proposed DL paradigm can efficiently simulate high-resolution monthly global gravity field both in GRACE/GRACE-FO and pre-GRACE era. The resulting simulated data products are available as monthly mass change grids as well as spherical harmonic models up to d/o 180. |
format |
Conference Object |
author |
Atman, Kazım Gökhan Olgun, Sevda Güneş, Özge Uz, Metehan Akyılmaz, Orhan Shum, C.K. |
spellingShingle |
Atman, Kazım Gökhan Olgun, Sevda Güneş, Özge Uz, Metehan Akyılmaz, Orhan Shum, C.K. Deep Learning-Aided High-Resolution Temporal Gravity Field Simulations: Monthly global mass grids and Spherical Harmonics from 1994 to 2021 |
author_facet |
Atman, Kazım Gökhan Olgun, Sevda Güneş, Özge Uz, Metehan Akyılmaz, Orhan Shum, C.K. |
author_sort |
Atman, Kazım Gökhan |
title |
Deep Learning-Aided High-Resolution Temporal Gravity Field Simulations: Monthly global mass grids and Spherical Harmonics from 1994 to 2021 |
title_short |
Deep Learning-Aided High-Resolution Temporal Gravity Field Simulations: Monthly global mass grids and Spherical Harmonics from 1994 to 2021 |
title_full |
Deep Learning-Aided High-Resolution Temporal Gravity Field Simulations: Monthly global mass grids and Spherical Harmonics from 1994 to 2021 |
title_fullStr |
Deep Learning-Aided High-Resolution Temporal Gravity Field Simulations: Monthly global mass grids and Spherical Harmonics from 1994 to 2021 |
title_full_unstemmed |
Deep Learning-Aided High-Resolution Temporal Gravity Field Simulations: Monthly global mass grids and Spherical Harmonics from 1994 to 2021 |
title_sort |
deep learning-aided high-resolution temporal gravity field simulations: monthly global mass grids and spherical harmonics from 1994 to 2021 |
publishDate |
2023 |
url |
https://avesis.yildiz.edu.tr/publication/details/f558041f-991a-4c31-853f-4e9e3f3ac534/oai |
geographic |
Greenland |
geographic_facet |
Greenland |
genre |
glacier Greenland Ice Sheet |
genre_facet |
glacier Greenland Ice Sheet |
op_relation |
f558041f-991a-4c31-853f-4e9e3f3ac534 https://avesis.yildiz.edu.tr/publication/details/f558041f-991a-4c31-853f-4e9e3f3ac534/oai |
op_rights |
info:eu-repo/semantics/closedAccess |
_version_ |
1778140146658443264 |