CE46 - Modèles numériques, simulation, applications

High Performance computing for quantifying climate change impacts on boreal areas – HiPerBorea

High Performance computing for quantifying climate change impacts on Boreal Areas

HiPerBorea aims to enable quantitative and predictive modeling of the thermo-hydrological evolution of cold environments covered by permafrost (25% of lands of northern hemisphere) under climate change. Permafrost thaw is associated with major feed-backs on greenhouse gases cycles (e.g.: thawing of previously frozen organic carbon pools).

Permafrost modeling: climate, environmental engineering in cold regions, geotechnics, ...

Permafrost thaw dynamics is a key factor for climatic evolution. The involved physics are highly coupled and non-linear, and simulating them at the scale of long-term monitored watersheds requires the development of spatially distributed and process oriented high performance computing tools. The fields of applications are numerous, from the hydro-bio-geochemistry of cold areas to engineering applications such as infrastructure stability in cold environments or site scale cryo-barriers in polluted industrial areas.

The approach that we aim to apply in order to reach this goal is to develop an OpenFOAM® framework for parallel computing modeling of thermo-hydrodynamics of cold continental surfaces. The team of the project already recently produced a validated and high performance simulation tool for coupled thermo-hydrological transfers in the ground with freeze/thaw of the pore water, the so-called permaFoam solver (Orgogozo et al., 2019), with a good scalability tested on tier-2 and tier-1 supercomputers up to 4000 cores, and enabling to deal with large problems such as a 1.2 billion cells mesh problems (Orgogozo et al., 2015). In this project we would like to further develop the high performance computing capabilities of permaFoam and to apply them to perform highly challenging, watershed scale, centennial simulations of thermo-hydrologic transfers in long term environmental monitoring stations of arctic and sub-arctic regions. The goal is to give a proof of concept of the benefit that can be expected from the use of modern high performance computing techniques for cold regions sciences and engineering.

To date the first results of HiPerBorea are the following:
- a new version of the numerical tool permaFoam (Orgogozo et al., 2019) with the help of ESI-Group - OpenCFD, the industrial maintainer of OpenFOAM (openfoam.com). This new version has been submitted for publication in the Computer Physics Communications code library (Orgogozo, soumis).
- first 3D simulations of permafrost at the watershed scale in the Kulingdakan catchment (Orgogozo et al., 2020, 2021).
- the gathering of a complete data set to be used for simulating the permafrost nearby the Abisko Interact station in Sweden (field mission in late september 2020 (https://hiperborea.omp.eu/field-mission-in-abisko-northern-sweden/).

First, the post-doc that have just been hired at GET will start the simulations of impacts of climate change on Kulingdakan watershed, then he will focus on the Abisko region. He will mainly use computationnal ressources iof the CINES (OCCIGEN) and of the TGCC (IRENE-ROME). The PhD at IMFT is focusing on the peatlands of Khanymey (Western SIberia). He mainly computes on the tier-2 center CALMIP (Olympe). Besides a master trainee has started the modelling effort on Syrdakh watershed (Eastern SIberia), where the LSCE partners are currently in field work. He mainly computes with the computing ressources of the LSCE.

1. 2020 L. Orgogozo, C. Grenier, E. Mouche, M. Marcoux, M. Quintard. Numerical modeling of coupled heat and water transport for the study of permafrost dynamics: High Performance Computing simulations for watershed scale analysis. 12th International Conference on Porous Media, Août-Septembre 2020, Tsingtao, Chine (conférence virtuelle). hal.archives-ouvertes.fr/hal-02931744
2. 2021 L. Orgogozo, O. Pokrovsky, C. Grenier, E. Mouche, M. Marcoux, M. Quintard. Modeling of water fluxes in boreal peatlands : perspectives offered by High-Performance Computing techniques. Accepté pour le VIth International Field Symposium West Siberian Peatlands and Carbon Cycle: Past and Present, 28 Juin – 7 juillet 2021, Khanty-Mansiysk, Russie
3. 2021 S. J. Cazaurang, M. Marcoux, S. Loiko, A. Lim, S. Audry, L. Shirokova, O. Pokrovsky, L. Orgogozo. Numerical Estimation of Effective Hydraulic Properties of Sphagnum Moss, Lichen and Peat From Western Siberian Lowlands. Accepté pour le VIth International Field Symposium West Siberian Peatlands and Carbon Cycle: Past and Present, 28 Juin – 7 juillet 2021, Khanty-Mansiysk, Russie
4. 2021 C. Grenier, A. Séjourné, E. Pohl, P. Konstantinov, A. Fedorov, A. Saintenoy, I. Khristoforov, L. Orgogozo, F. Costard. Study of Thermo-hydrological river-ground transfer processes in the context of permafrost environments – the Syrdakh study site in Central Yakutia (Siberia, Russia). Accepted for the international confrerence Advancing critical zone science. Ozcar-Tereno meeting, 5-7 Oct. 2021, Strasbourg, France
5. (submitted) L. Orgogozo. RichardsFoam3: a new version of RichardsFoam for continental surfaces hydrogeology modeling.
6. (submitted) L. Orgogozo. permaFoam: an extension of RichardsFoam for cryohydrogeological modeling.

HiPerBorea aims to enable quantitative and predictive modeling of the thermo-hydrological evolution of cold environments covered by permafrost (25% of lands of northern hemisphere) under climate change. Permafrost thaw is associated with major feed-backs on greenhouse gases cycles (e.g.: thawing of previously frozen organic carbon pools). Its dynamics is a key factor for climatic evolution. The involved physics are highly coupled and non-linear, and simulating them at the scale of long-term monitored watersheds requires the development of spatially distributed and process oriented high performance computing tools. The fields of applications are numerous, from the hydro-bio-geochemistry of cold areas to engineering applications such as infrastructure stability in cold environments or site scale cryo-barriers in polluted industrial areas.

The approach that we aim to apply in order to reach this goal is to develop an OpenFOAM® framework for parallel computing modeling of thermo-hydrodynamics of cold continental surfaces. The team of the project already recently produced a validated and high performance simulation tool for coupled thermo-hydrological transfers in the ground with freeze/thaw of the pore water, the so-called permaFoam solver (Orgogozo et al., 2019), with a good scalability tested on tier-2 and tier-1 supercomputers up to 4000 cores, and enabling to deal with large problems such as a 1.2 billion cells mesh problems (Orgogozo et al., 2015). In this project we would like to further develop the high performance computing capabilities of permaFoam and to apply them to perform highly challenging, watershed scale, centennial simulations of thermo-hydrologic transfers in long term environmental monitoring stations of arctic and sub-arctic regions. The goal is to give a proof of concept of the benefit that can be expected from the use of modern high performance computing techniques for cold regions sciences and engineering.

In order to relevantly use this computational power, the permaFoam solver will be further developed to cope with all the determinant processes for water catchment hydrology and hydrogeology in boreal areas. As such it will:
1) integrates in a numerically efficient way the main external processes that control permafrost hydrology, such as bryophytic layer (i.e. moss and lichen cover) dynamics, solar radiation penetration and snow cover thermal insulation;
2) be validated for watershed scale modeling according to the current international standards, and parameterized based on field data sets from sites of long term monitoring of permafrost (eu-interact.org, IRN CAR WET SIB);
3) be used to perform 3D, watershed scale modeling of arctic and sub-arctic permafrost dynamics in response to climate change, thanks to its good parallel performances, which will be optimised for current supercomputers along the project (from the Tier-2 regional meso-centre to the Tier-0 PRACE european computing infrastructures).

HiPerBorea will proceed in three main steps:
1) developing, testing and validating the needed numerical tool,
2) applying this tool to the already numerically studied Kulingdakan experimental watershed (Orgogozo et al., 2019) in order to establish an efficient, consolidated methodology for 3D watershed scale, centennial cryohydrologic modeling and finally
3) simulating the responses of four reference long term monitoring boreal catchments to various IPCC scenarios of climate change (CMIP5 projections).

Technically speaking the goal is to test the limit size of the applications that can be dealt with the present performances of permaFoam (ability to deal with billion cells mesh, with up to 4000 cores on Tier-1 supercomputers), and to go beyond these limits along the project – towards 10 billions cells mesh, and tens of thousands of used cores on Tier-0 supercomputers.

Project coordinator

Monsieur Laurent Orgogozo (Centre National de la Recherche Scientifique/Géosciences Environnement Toulouse)

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.

Partner

CNRS/GET Centre National de la Recherche Scientifique/Géosciences Environnement Toulouse
LSCE CEA SACLAY - DRF - LSCE
CESBIO Centre d'études spatiales de la biosphère
IMFT INSTITUT DE MECANIQUE DES FLUIDES DE TOULOUSE

Help of the ANR 627,367 euros
Beginning and duration of the scientific project: December 2019 - 48 Months

Useful links

Explorez notre base de projets financés

 

 

ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter