This thesis examines the importance of energy and momentum consistency in subgrid-scale parameterization for climate models. It is divided into two parts according to the two aspects of the problem that are investigated, namely the importance of momentum conservation alone and the consistency between energy and momentum conservation. The first part addresses the importance of momentum conservation alone. Using a zonally-symmetric model, it is shown that violating momentum conservation in the parameterization of gravity wave drag leads to large errors and non-robustness of the response to an imposed radiative perturbation in the middle atmosphere. Using the Canadian Middle Atmosphere Model, a three-dimensional climate model, it is shown that violating momentum conservation, by allowing gravity wave momentum flux to escape through the model lid, leads to large errors in the mean climate when the model lid is placed at 10 hPa. When the model lid is placed at 0.001 hPa the errors due to nonconservation are minimal. When the 10 hPa climate is perturbed by idealized ozone depletion in the southern hemisphere, nonconservation is found to significantly alter the polar temperature and surface responses. Overall, momentum conservation ensures a better agreement between the 10 hPa and the 0.001 hPa climates. The second part addresses the self-consistency of energy and momentum conservation. Using Hamiltonian geophysical fluid dynamics, pseudoenergy and pseudomomentum wave-activity conservation laws are derived for the subgrid-scale dynamics. Noether’s theorem is used to derive a relationship between the wave-activity fluxes, which represents a generalization of the first Eliassen-Palm theorem. Using multiple scale asymptotics a theoretical framework for subgrid-scale parameterization is built which consistently conserves both energy and momentum and respects the second law of thermodynamics. The framework couples a hydrostatic resolved-scale flow to a non-hydrostatic subgrid-scale flow. The transfers of energy and momentum between the two scales are understood using the subgrid-scale wave-activity conservation laws, whose relationships with the resolved-scale dynamics represent generalized non-acceleration theorems. The derived relationship between the wave-activity fluxes — which represents a generalization of the second Eliassen-Palm theorem — is key to ensuring consistency between energy and momentum conservation. The framework includes a consistent formulation of heating and entropy production due to kinetic energy dissipation.

This thesis examines the importance of energy and momentum consistency in subgrid-scale parameterization for climate models. It is divided into two parts according to the two aspects of the problem that are investigated, namely the importance of momentum conservation alone and the consistency between energy and momentum conservation. The first part addresses the importance of momentum conservation alone. Using a zonally-symmetric model, it is shown that violating momentum conservation in the parameterization of gravity wave drag leads to large errors and non-robustness of the response to an imposed radiative perturbation in the middle atmosphere. Using the Canadian Middle Atmosphere Model, a three-dimensional climate model, it is shown that violating momentum conservation, by allowing gravity wave momentum flux to escape through the model lid, leads to large errors in the mean climate when the model lid is placed at 10 hPa. When the model lid is placed at 0.001 hPa the errors due to nonconservation are minimal. When the 10 hPa climate is perturbed by idealized ozone depletion in the southern hemisphere, nonconservation is found to significantly alter the polar temperature and surface responses. Overall, momentum conservation ensures a better agreement between the 10 hPa and the 0.001 hPa climates. The second part addresses the self-consistency of energy and momentum conservation. Using Hamiltonian geophysical fluid dynamics, pseudoenergy and pseudomomentum wave-activity conservation laws are derived for the subgrid-scale dynamics. Noether’s theorem is used to derive a relationship between the wave-activity fluxes, which represents a generalization of the first Eliassen-Palm theorem. Using multiple scale asymptotics a theoretical framework for subgrid-scale parameterization is built which consistently conserves both energy and momentum and respects the second law of thermodynamics. The framework couples a hydrostatic resolved-scale flow to a non-hydrostatic subgrid-scale flow. The transfers of energy and momentum between the two scales are understood using the subgrid-scale wave-activity conservation laws, whose relationships with the resolved-scale dynamics represent generalized non-acceleration theorems. The derived relationship between the wave-activity fluxes — which represents a generalization of the second Eliassen-Palm theorem — is key to ensuring consistency between energy and momentum conservation. The framework includes a consistent formulation of heating and entropy production due to kinetic energy dissipation.

Ce travail a pour objectif de développer et d’appliquer une méthode originale permettant de simuler l’écoulement dans un milieu poreux fracturé. Cette méthode repose sur une approche multicouches double continuum permettant de séparer le comportement des différents aquifères présents sur un site. La résolution des écoulements, basée sur la méthode des Eléments Finis de Crouzeix-Raviart, est associée à une méthode inverse (minimisation de type Quasi-Newton combinée à la méthode de l’état adjoint) et à une paramétrisation multi-échelle.La méthode est appliquée dans un premier temps sur l’aquifère fracturé du site expérimental de Poitiers. Les résultats montrent une bonne restitution du comportement de l’aquifère et aboutissent à des champs de transmissivité plus réguliers par rapport à ceux de l’approche simple continuum. L’application finale est réalisée sur le site de Cadarache (taille plus importante et données d’entrée moins denses). Le calage des deux aquifères présents sur le site est satisfaisant et montre que ceux-ci se comportent globalement de façon indépendante. Ce calage pourra être amélioré localement grâce à données de recharge plus fines. / The aim of this study is to develop and validate a new method for the simulation of flow in fractured porous media. This method is based on a multi-layered and dual continuum approach allowing to discriminate the behavior of different aquifers present on a site. The flow equations are solved using a Crouzeix-Raviart Finite Element method, in association with an inverse method (Quasi-Newton minimization combined with the adjoint state method) and a multi-scale parameterization.The method is first applied and validated on the fractured aquifer of the Hydrogeological Experimental Site of Poitiers. The results closely reproduce the flow behavior of the aquifer and lead to a transmissivity field much more homogeneous than the one obtained with a simple continuum approach. The final application is performed on the site of Cadarache (large scale problem with heterogeneously distributed input data). The model calibration of both aquifers is rather satisfactory and shows that their behavior is globally independent. It could locally be improved if more accurate groundwater recharge data is made available.