Toward High-Resolution Global Atmospheric Inverse Modeling Using Graphics Accelerators

Frédéric Chevallier; Zoé Lloret; Anne Cozic; Sakina Takache; Marine Remaud

doi:10.1029/2022GL102135

Toward High-Resolution Global Atmospheric Inverse Modeling Using Graphics Accelerators

Frédéric Chevallier
, Zoé Lloret
, Anne Cozic
, Sakina Takache
, Marine Remaud

Systems Ecology

Research output: Contribution to Journal › Article › Academic › peer-review

Abstract

The ability of global transport models to go up in resolution becomes discriminating for greenhouse gas atmospheric inversions. This paper describes the porting on Graphics Processing Units of the global transport model currently used in the European operational Copernicus Atmosphere Monitoring Service (CAMS) for CO2 and N2O inversions. It represents an important milestone to achieve sub-degree resolution. The code includes not only the direct model but also its tangent-linear and its adjoint versions which are needed in variational inversions. Tests were carried out for CO2 at a resolution of 2.50° in longitude, 1.27° in latitude and 79 layers in the vertical, corresponding to 1,626,768 3D cells, 4.5 times more than the current standard resolution of the model used in the CAMS reanalyzes. A month's worth of computation of the tangent-linear and of the adjoint versions now takes 2.5 min, including 50 s for reading meteorological data.

Original language	English
Article number	e2022GL102135
Pages (from-to)	1-9
Number of pages	9
Journal	Geophysical Research Letters
Volume	50
Issue number	5
Early online date	16 Mar 2023
DOIs	https://doi.org/10.1029/2022GL102135
Publication status	Published - Mar 2023

Funding

This work was granted access to the HPC resources of CCRT under the allocation CEA/DRF, and of TGCC under the allocation A0110102201 made by GENCI. It was funded by the Copernicus Atmosphere Monitoring Service, implemented by ECMWF on behalf of the European Commission (Grant: CAMS2 55). The authors would like to thank Yann Meurdesoif (LSCE) for his long-standing support with the model developments and his advice about GPUs in particular. This work was granted access to the HPC resources of CCRT under the allocation CEA/DRF, and of TGCC under the allocation A0110102201 made by GENCI. It was funded by the Copernicus Atmosphere Monitoring Service, implemented by ECMWF on behalf of the European Commission (Grant: CAMS2 55). The authors would like to thank Yann Meurdesoif (LSCE) for his long‐standing support with the model developments and his advice about GPUs in particular.

Funders	Funder number
CCRT
Copernicus Atmosphere Monitoring Service
TGCC
Duchenne Research Fund
European Commission	CAMS2 55
Commissariat à l'Énergie Atomique et aux Énergies Alternatives
Grand Équipement National De Calcul Intensif

Access to Document

10.1029/2022GL102135

Persistent URL (handle)

Persistent link

Cite this

@article{836b405805764acd8dfea7a3160a68f4,

title = "Toward High-Resolution Global Atmospheric Inverse Modeling Using Graphics Accelerators",

abstract = "The ability of global transport models to go up in resolution becomes discriminating for greenhouse gas atmospheric inversions. This paper describes the porting on Graphics Processing Units of the global transport model currently used in the European operational Copernicus Atmosphere Monitoring Service (CAMS) for CO2 and N2O inversions. It represents an important milestone to achieve sub-degree resolution. The code includes not only the direct model but also its tangent-linear and its adjoint versions which are needed in variational inversions. Tests were carried out for CO2 at a resolution of 2.50° in longitude, 1.27° in latitude and 79 layers in the vertical, corresponding to 1,626,768 3D cells, 4.5 times more than the current standard resolution of the model used in the CAMS reanalyzes. A month's worth of computation of the tangent-linear and of the adjoint versions now takes 2.5 min, including 50 s for reading meteorological data.",

author = "Fr{\'e}d{\'e}ric Chevallier and Zo{\'e} Lloret and Anne Cozic and Sakina Takache and Marine Remaud",

year = "2023",

month = mar,

doi = "10.1029/2022GL102135",

language = "English",

volume = "50",

pages = "1--9",

journal = "Geophysical Research Letters",

issn = "0094-8276",

publisher = "John Wiley and Sons Inc.",

number = "5",

}

TY - JOUR

T1 - Toward High-Resolution Global Atmospheric Inverse Modeling Using Graphics Accelerators

AU - Chevallier, Frédéric

AU - Lloret, Zoé

AU - Cozic, Anne

AU - Takache, Sakina

AU - Remaud, Marine

PY - 2023/3

Y1 - 2023/3

N2 - The ability of global transport models to go up in resolution becomes discriminating for greenhouse gas atmospheric inversions. This paper describes the porting on Graphics Processing Units of the global transport model currently used in the European operational Copernicus Atmosphere Monitoring Service (CAMS) for CO2 and N2O inversions. It represents an important milestone to achieve sub-degree resolution. The code includes not only the direct model but also its tangent-linear and its adjoint versions which are needed in variational inversions. Tests were carried out for CO2 at a resolution of 2.50° in longitude, 1.27° in latitude and 79 layers in the vertical, corresponding to 1,626,768 3D cells, 4.5 times more than the current standard resolution of the model used in the CAMS reanalyzes. A month's worth of computation of the tangent-linear and of the adjoint versions now takes 2.5 min, including 50 s for reading meteorological data.

AB - The ability of global transport models to go up in resolution becomes discriminating for greenhouse gas atmospheric inversions. This paper describes the porting on Graphics Processing Units of the global transport model currently used in the European operational Copernicus Atmosphere Monitoring Service (CAMS) for CO2 and N2O inversions. It represents an important milestone to achieve sub-degree resolution. The code includes not only the direct model but also its tangent-linear and its adjoint versions which are needed in variational inversions. Tests were carried out for CO2 at a resolution of 2.50° in longitude, 1.27° in latitude and 79 layers in the vertical, corresponding to 1,626,768 3D cells, 4.5 times more than the current standard resolution of the model used in the CAMS reanalyzes. A month's worth of computation of the tangent-linear and of the adjoint versions now takes 2.5 min, including 50 s for reading meteorological data.

U2 - 10.1029/2022GL102135

DO - 10.1029/2022GL102135

M3 - Article

SN - 0094-8276

VL - 50

SP - 1

EP - 9

JO - Geophysical Research Letters

JF - Geophysical Research Letters

IS - 5

M1 - e2022GL102135

ER -

Toward High-Resolution Global Atmospheric Inverse Modeling Using Graphics Accelerators

Abstract

Funding

Access to Document

Persistent URL (handle)

Fingerprint

Cite this