Modeling stable boundary layers in Antarctica : calibration and sensitivity analysis of the parameterization of turbulence in the ARPEGE-Climate model

• Main Author: Audouin, Olivier
• Other Authors: Groupe de Météorologie Expérimentale et Instrumentale (GMEI), Centre national de recherches météorologiques (CNRM), Institut national des sciences de l'Univers (INSU - CNRS)-Observatoire Midi-Pyrénées (OMP), Institut de Recherche pour le Développement (IRD)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université de Toulouse (UT)-Institut national des sciences de l'Univers (INSU - CNRS)-Centre National d'Études Spatiales Toulouse (CNES)-Centre National de la Recherche Scientifique (CNRS)-Météo-France-Institut de Recherche pour le Développement (IRD)-Université Toulouse III - Paul Sabatier (UT3), Université de Toulouse (UT)-Université de Toulouse (UT)-Institut national des sciences de l'Univers (INSU - CNRS)-Centre National d'Études Spatiales Toulouse (CNES)-Centre National de la Recherche Scientifique (CNRS)-Météo-France-Centre National de la Recherche Scientifique (CNRS)-Institut national des sciences de l'Univers (INSU - CNRS)-Observatoire Midi-Pyrénées (OMP), Université de Toulouse (UT)-Université de Toulouse (UT)-Institut national des sciences de l'Univers (INSU - CNRS)-Centre National d'Études Spatiales Toulouse (CNES)-Centre National de la Recherche Scientifique (CNRS)-Météo-France-Centre National de la Recherche Scientifique (CNRS), Institut National Polytechnique de Toulouse - INPT, Fleur Couvreux, Romain Roehrig
• Format: Doctoral or Postdoctoral Thesis
• Language: French
• Published: HAL CCSD 2021
• Subjects: Physical parameterization; Boundary layer; Turbulence; Climate modelling; Statistical calibration; Antarctica; Paramétrisation physique; Couche limite; Modèlisation du climat; Calibration statistique; Antarctique; [SDU.STU]Sciences of the Universe [physics]/Earth Sciences; Antarctic; The Antarctic; Antarc*; Antarctique*
• Online Access: https://theses.hal.science/tel-04170839; https://theses.hal.science/tel-04170839/document; https://theses.hal.science/tel-04170839/file/AUDOUIN__-_Olivier.pdf

Climate modeling is a key tool for understanding the climate system and for making projections of its future evolution. Yet numerical climate models have multiple sources of uncertainty. Among these, the representation of stable boundary layer processes remains one of the main points on which numerical models must make progress. The most extreme stable boundary layers are observed on the Antarctic plateau. A good modeling of Antarctica in numerical models is based on a good representation of energy exchanges within the boundary layer and in particular of turbulent fluxes. However, in a climate model, turbulent processes, as well as any other small-scale processes (not resolved by the dynamical part of the model), or processes not related to fluid mechanics (e.g. radiation), are based on a set of sub-models called physical parameterizations. These parameterizations introduce a certain number of parameters whose values are more or less well documented and which can be considered as the adjustment levers of the model. The model tuning step is the choice of the values of these different parameters and is considered as a crucial step in the development of the models. The classical approach is to look for an optimal model setting based on a set of metrics. This is a long and tedious work, in which one or two parameters are varied one at a time and which does not allow the exploration of all the possibilities of model adjustment. Moreover, the sources of uncertainty are not always taken into account and the whole procedure is not very reproducible. An approach inspired by History Matching has recently been proposed to calibrate the physics of atmospheric models. This approach proposes not to look for a potential optimal setting of the parameters, but to identify a region of the parameter space in which the model « performs well ». Associated with the use of statistical emulators mimicking the climate model behaviour, it allows to answer part of the problems posed by a more classical approach (exhaustive exploration of ...