Aller au contenu. | Aller à la navigation

Laboratoire d’Etudes en Géophysique et Océanographie Spatiales

Outils personnels

This is SunRain Plone Theme


Vous êtes ici : Accueil / Projets / Axes Transverses & Processus / TEASAO / Description of the project


Par Webmaster Legos Dernière modification 19/04/2017 17:26



Turbulence Effects on Active Species in Atmosphere and Ocean
submitted to the
Toulouse IDEX program “Attractivity Chairs 2014”


Professor Peter HAYNES
Professor at University of Cambridge (UK)
Department of Applied Mathematics and Theoretical Physics (DAMTP)


Directeur de Recherche CNRS
(coordinator for a consortium of participants)

Université de Toulouse III, Paul Sabatier (UPS)
Observatoire Midi-Pyrénées
Laboratoire d’Aérologie (LA)
Laboratoire d’Etudes en Géophysique
et Océanographie Spatiales (LEGOS)
Centre National de Recherches Météorologique

Keys words: Ocean – Atmosphere –Tracers – Turbulence – Active/reactive species – Radiative effects – Chemistry – Biogeochemistry – Environmental issues – Fundamental to Applied studies.

1. Presentation

1.1 Participants

The TEASAO (“Turbulence Effects on Active Species in Atmosphere and Ocean”) project aims through the IDEX Chaire d’Attractivite programme to build a new long-term partnership within Toulouse between the three laboratories CNRM, LA and LEGOS (Toulouse Universities and scientific Institutes consortium, COMUE-Toulouse), and externally to DAMTP (University of Cambridge).
The participants are:

  • Peter Haynes, DAMTP, Professor at University of Cambridge, candidate for the Toulouse IDEX attractivity chair;
  • Alexandre Paci, CNRM, IPEF-chercheur (Météo-France & CNRS);
  • Philippe Ricaud, CNRM, Directeur de recherche (CNRS) ;
  • Jean-Pierre Chaboureau, LA, Physicien (CNAP,  UPS);
  • Francis Auclair, LA, Maître de conférence (UPS);
  • Véronique Garçon, LEGOS, Directrice de recherche (CNRS);
  • Yves Morel, LEGOS, Directeur de recherche (CNRS), general coordinator of the project.

1.2 Context and general description of the TEASAO project

Toulouse is the main national center for the development of operational meteorology (Météo France) and operational oceanography (Mercator Océan, SHOM operational center, contribution to the Previmer project …) and the general background of the present project is the preparation of the next generation prediction systems. Enabled by the improvement of space or in situ observation systems, computing facilities and numerical models, a common and constant theme for atmospheric and oceanographic sciences is our ability to understand and represent smaller and smaller scale processes. The upscaling effect by which smaller scale processes have a strong influence on larger scales means that the systematically improving representation of small scales potentially offers major improvements for prediction systems.

One very important and broad class of processes, common to both atmosphere and ocean, concerns the interaction at small scales between fluid dynamics and tracers (substances transported and mixed by the flow) such as chemical species (including atmospheric water vapour) and biological species. Many of these tracers are active in the sense that they undergo non-trivial transformation (e.g. reaction with other chemical species, interaction with other biological species or, in the case of atmospheric water vapour, change of phase) with these transformations depending strongly on fluid processes. In some cases there is a significant effect of the tracer evolution on the flow itself (e.g. through heat generated in phase changes, or through radiative effects, or when the tracer contributes significantly to the density field). Understanding and modelling the effect of turbulence and mixing on biologically, chemically or physically active species in density-stratified geophysical flows remains a current major scientific challenge, at the heart of international scientific programmes (Future Earth, SOLAS, IMBER, ICACGP ...), with many implications for society challenges and our ability to model the local environment and the climate system as a whole.

The scientific goal of the present project is therefore to improve our general understanding and the modelling of the processes coupling fluid dynamics to the evolution of active species at local to global scales.

The many different physical, chemical and biological processes that are involved in different applications means that many problems in this class are addressed independently, in Toulouse and elsewhere, within separate atmospheric and oceanic communities. There are however strong scientific connections between them and our communities would benefit from a framework allowing exchange of knowledge and ideas and development of common approaches and investigation tools. The first strategic goal of the present project is to therefore to build a long term efficient partnership of laboratories across the COMUE-Toulouse community to improve our understanding of the effect of small scale turbulence and mixing on the evolution of active species in both the atmosphere and the ocean.

The link of the TEASAO project with applications and prediction systems is twofold:

  • The numerical models used in the project are developed by the participants and are used for operational systems. They will be evaluated and improved in their capacity to represent small scale turbulent processes and their effects on active species. This will prepare the future operational models for applications at local scale (on horizontal domains covering typically 100kmx100km).
  • It will however be at least another decade (and probably longer) before the direct numerical simulations of the small scale turbulent processes can be directly represented in global or even regional scale prediction systems. The processes and their effects on tracers have to be parameterized and the studies proposed here will also address this issue. We will indeed study the effect of these small scale processes at macro-scales and the way to take them into account in models whose spatial resolution is too coarse to allow their direct simulation.


The second strategic goal of the TEASOA project is thus to prepare the next generation of prediction systems for both the atmosphere and the ocean at local to global scales.

Professor Peter Haynes is an international expert on the dynamics of atmosphere and ocean and on the transport and mixing of tracers, including reacting chemical and biological species. He has in his previous research addressed these scientific issues with both theoretical and applied studies. Much of his work has been focused on atmospheric problems but he has successfully applied his knowledge and results to the ocean and indeed has previously collaborated with oceanographers from LEGOS. Over the proposed project period he will be able to spend extended periods in Toulouse and additionally he will provide links to researchers in Cambridge both in fluid dynamics and in several aspects of climate science. He is thus the ideal candidate to help us gather our communities and the present project has been built around his availability for the attractivity chair.

2. Detailed research program and fallouts for Toulouse University

2.1 General strategy

Whilst turbulence in geophysical fluids has some aspects in common with turbulence in other contexts, e.g. industrial flows, many aspects are very specific, with the stabilizing effect of vertical density stratification and the planetary Coriolis effect leading to many different types of flow dynamics at different scales and, in particular, to large discrepancies between vertical and horizontal scales . The rationale for the project is thus that the effect of small scale turbulence and mixing on the evolution of active species is an important topic, with relevance to key scientific problems in both atmosphere and ocean, and that the atmospheric and oceanic science communities in Toulouse would benefit from a cross-community framework allowing exchange of knowledge and ideas and development of common approaches and investigation tools. Bearing in mind that there are common fundamental mechanisms and specific applications, we propose to study the effect of small scale turbulent processes on active species through four subprojects: a general theoretical study and three specific studies concerning either the ocean or the atmosphere. This will constitute a new approach in Toulouse and, with Peter Haynes’ help, will federate local scientific expertise in this subject area. The sub-projects have been chosen with the dual aim of matching the overall theme of the project and promoting broad participation across the Toulouse communities.

Each sub-project will be a collaboration between Toulouse scientists from different laboratories, together with Peter Haynes, with a specified Toulouse scientist acting as lead and with the responsibility of supervising the postdoc or PhD student. Each sub-project will also incorporate scientific visits of the postdoc or student to Cambridge, where they will interact not only with Peter Haynes, but also with other fluid dynamicists and climate scientists.

Computing limitations and the complex chaotic character of geophysical fluid motions, plus the large domain sizes required by operational prediction systems, means that direct numerical simulation of small-scale turbulence is beyond the reach of the present computing facilities, but specific modelling in idealized configurations -or realistic configurations but in a very limited area- is accessible. We will thus address the effect of small scale turbulence and mixing on the evolution of atmospheric and oceanic active species in idealized configurations but nonetheless using state-of-the-art numerical models, used for realistic modelling of the atmosphere and ocean:

  • CROCO-NH (ocean), developed at LA;
  • AROME/MESO-NH (atmosphere), developed at CNRM and LA.


The participants of the TEASOA projects actively participate to the development of the latter numerical models in configurations ranging from academic to realistic.


2.2 (SUB-PROJECT 1) Theoretical study: Fundamental aspects of turbulence, mixing and tracers evolution in stratified geophysical flows (3-year postdoc).

Main partners: Yves Morel, LEGOS, leader of the sub-project ; Peter Haynes, DAMTP;  Alexandre Paci, CNRM;  Francis Auclair, LA;  Jean-Pierre Chaboureau, LA.

The combination of density stratification and turbulence is fundamental to many important flows in both atmosphere and ocean. In some cases (e.g. the oceanic mixed layer) turbulence is persistent and extends over a layer of significant depth within which density stratification is weak. In other cases (the tropical tropopause layer, the oceanic thermocline) there may be short-lived turbulence patches within a strongly stratified environment. We propose to study the fundamental physics of coupling between dynamics and reactive tracers in such flows. Idealised configurations will allow investigation of small scale turbulence and mixing and its effect on the distribution of tracers in several specific instability problems. It is planned to represent and study the following turbulent processes:

  • shear instabilities and the development of Kelvin-Helmholtz instabilities in geophysical fluids;
  • the breaking of internal waves of large amplitude;
  • the development of convection.


We will use the CROCO-NH model in simplified configurations to represent the latter processes with different initial conditions (variable stratification, instability characteristics and initial tracer distribution). We will particularly look at the spatial distribution and evolution of the local concentration of tracers during the instability and mixing phases. The timescales of the different phases will be examined: during the development of instabilities, tracers are stirred or concentrated as a function of the local properties of the velocity fields. The turbulent cascade eventually leads to mixing and homogenization of tracers at small scales with a final distribution that then remains fixed along density surfaces, and which prescribes the further localisation and transport of the tracer.
At the end of the mixing phase, the tracers homogenizes over vertical scales which are usually much smaller than the thickness of the initial instability characteristics (Kelvin Helmholtz billows for instance). Therefore, the final profiles of density and tracers depend on the local volumetric ratio of tracers in small regions of the flow. In this perspective, any preconditioning process acting on the local distribution of tracers and density prior to the turbulent phase may have an impact on the distribution of tracers. In addition, before the mixing phase, if the development of the instability is slow enough in comparison with the timescales of biological or chemical reaction of species, the stirring/concentration events during the instability growth phase will also play an important role in the reaction and in determining the final distribution of species. It is thus expected that the ratio of active tracers greatly varies within the Kelvin-Helmholtz billows leading to a hyper-sensitivity to details of the turbulence preconditioning again.
The main processes and parameters to be examined are:

  • the effect of preconditioning on the tracers/species concentration during the instability development. Previous studies have already identified several processes of interest (for instance for Kelvin-Helmholtz instability: billow pairings, convective spanwise rolls …);
  • the vertical extent of the mixing and homogenization in comparison with the initial instability scale;
  • the timescale of the biological or chemical reaction in comparison with the growth rate of the instability development;
  • the distribution of active species at larger scales (called upscaling effect);
  • the evolution of species spectrum in both space and time.


This will be used to propose new methodologies and new approaches to better represent the upscaling effects of the turbulent processes.  For instance, we will evaluate stochastic parameterizations for larger scale models used for numerical forecasts of the atmosphere and ocean.
The problem will be addressed with increasing complexity: from a single passive tracers to multiple active tracer, from the influence of an external forcing (solar radiative effects) to nonlinear transformation laws for multiple reactions.

2.3 (SUB-PROJECT 2) Specific application to the ocean: Turbulence and control of planktonic development and export production in the ocean (3-year PhD student)

Main partners: Véronique Garçon, LEGOS, leader of the sub-project; Peter Haynes, DAMTP; Francis Auclair, LA; Yves Morel, LEGOS.

Ocean biogeochemical models are used to quantify fluxes of nutrients and carbon between different parts of the ocean and through the planktonic food web. It is important to represent these fluxes on basin and global spatial scales and on time scales of years to decades (and longer for some applications), but because of computational limitations, models incorporating these large space and long time scales cannot represent the full spatial complexity of the ocean at small scales and typically have grid cells with horizontal scale O(100 km). There is a growing realization that physical-biological interactions play a crucial role in the ecology of organisms and biogeochemical tracers distributions at a surprising range of scales, from micrometers to kilometres, in the ocean. Understanding both the small-scale physical-biological-biogeochemical interactions and their implications for larger-scale processes requires an intimate understanding of the coupling between dynamics, biogeochemistry and biology at these small scales. The strong spatial variability in plankton, oxygen and nutrient concentrations over scales (O(1–100km) horizontally and O(1-100m) vertically, see Fig. 1) indicates similar variability in key fluxes.  Different questions arise: What is the impact of small-scale turbulent mixing on biogeochemical tracers and primary production? How does it control the distribution at larger scales? How should biogeochemistry be “upscaled” (parameterized) to the grid cells of coarse-resolution ocean models? In particular are they also linked to lateral mixing associated with ‘submesoscale’ processes on scales of O(1-10km)?
We will here focus on this last question and try to characterize the effects of lateral mixing at O(1-10 km) scales. Presently this is not accounted for either in the present coupled physical-biogeochemical models or in the climate models leading to potential errors in predicting distribution of temperature, salt and biogeochemical tracers.

We propose to run several simulations using a three-dimensional non hydrostatic model CROCO-NH, coupled with a simple NPZDO2 biogeochemical model, with 0.5 km (or less)  resolution in the horizontal to resolve the submesoscale. The underlying question will be to investigate whether lateral mixing at these scales is due to a balanced or unbalanced downscale cascade from the mesoscale, or due to local vertical mixing by surface forcing and internal waves, and, correspondingly, the implications for biological species. Submesoscale processes are innately vertical, as well as horizontal (see Fig. 1) and they facilitate cross-frontal exchange.  A simple setting will be adopted with a baroclinically unstable front in a periodic channel with and without initial meanders and with and without wind stress. We will examine the cross-frontal flux of tracers and the formation of filaments. We will use available in situ observations from the MOUTON 2007 (LEGOS – SHOM) campaign at sea to validate the simulation results.
We will pay a particular attention to explain observed spatial characteristics of typical chlorophyll a concentration distribution as shown on Fig. 1. The vertical section in the Iberian Peninsula upwelling system during  a summer  upwelling event reveals the existence of very small scale concentrations spots, much smaller than the vertical extent of the mixed layer, showing that very small scale processes –typical of turbulence- must be taken into account to explain this distribution. These planktonic patterns will also lead to specific localized organic matter export events.

Figure1:  Development of upwelling and associated phytoplankton bloom in summer along the Spanish and Portuguese coasts. Left: Chlorophyll a concentration observed using remote sensing satellite (MERIS) observation. Right: Distribution of Chlorophyll a concentration (measured in situ using fluorometry sensor, MOUTON2007 campaign) of a vertical section (North-South section represented by dots on the left panel) along the Portuguese coast during the summer upwelling events.                   
Figure1:  Development of upwelling and associated phytoplankton bloom in summer along the Spanish and Portuguese coasts. Left: Chlorophyll a concentration observed using remote sensing satellite (MERIS) observation. Right: Distribution of Chlorophyll a concentration (measured in situ using fluorometry sensor, MOUTON2007 campaign) of a vertical section (North-South section represented by dots on the left panel) along the Portuguese coast during the summer upwelling events.

2.4 (SUB-PROJECT 3) Specific application to the atmosphere: Combined effect of mixing and microphysical processes in the tropical tropopause layer (3-year PhD student)

Main partners: Jean-Pierre Chaboureau, LA, leader of the sub-project; Peter Haynes, DAMTP; Yves Morel, LEGOS.

The tropical tropopause layer (TTL) is a transition layer sharing upper tropospheric and lower stratospheric characteristics in which the motion is a complicated combination of large-scale circulations, waves on many different scales and turbulence. Vertical transport of air across the TTL is accepted as a key process regulating the composition of the stratosphere. The detailed mechanisms are however the subject of ongoing debate not least because they depend on several complex processes in which dynamics and physics are closely linked across a large range of scales. The details are particularly important with respect to water vapour, since the low concentrations of water vapour in the stratosphere are set by ‘freeze drying’ as air moves through the very cold tropical tropopause region. But they also have significant implications for other chemical species, such as both naturally occurring and anthropogenic halogen species that will become important controls on stratospheric ozone as concentrations of very long-lived anthropogenic species such as CFCs decrease. Two contrasting mechanisms for vertical transport are slow radiative ascent on the large scale and rapid very deep convection that overshoots the tropopause on the small scale. Both clearly contribute in some part to the overall composition of the stratosphere. The emphasis in this sub-project will be on using high-resolution numerical simulation to examine in more detail the small-scale processes. Ice formation processes at low temperatures affect the efficacy of freeze drying as air passes through the tropopause. Depending on the environmental conditions, the ice crystals can either slowly precipitate and sublimate or mix with the stratospheric air through Kelvin-Helmhotz instabilities and other waves. As an example, Fig. 2 shows a convective plume of ice particles with wavy clouds observed by lidar above the tropical tropopause. This type of complicated interplay between turbulence, mixing and microphysics will be investigated using the the Meso-NH model, taking advantage of its capability to be run over large grids on massively parallel supercomputers. A bin microphysics scheme will be implemented in order to describe the evolution of crystals following their mass diameter as function of ice nucleation, deposition/sublimation, particle collision accounting for turbulence, and sedimentation. Several case studies under idealized conditions and extreme environment of low temperature will be run across a large range of scales to investigate different scenarios including evolution of cirrus decks arising from convective injection and from large-scale transport through anomalously cold regions.


2.5 (SUB-PROJECT 4) Specific application to the atmosphere: Impact of convection on the water vapour budget in the tropical tropopause layer. (3-year PhD student)

Main partners: Philippe Ricaud, CNRM, leader of the sub-project; Peter Haynes, DAMTP; Jean-Pierre Chaboureau, LA.

The importance of the TTL in global distributions of water vapour and other chemical species and the multi-scale nature of TTL processes has already been outlined above with respect to sub-project 3. In order to assess the roles of different scales it is important to use new measurements and new modelling approaches in combination to test scientific understanding. The diurnal cycle of water vapour, temperature and other quantities in the TTL in particular, provides an opportunity for such a test. This diurnal cycle is almost certainly associated with a complicated mix of 'waves' (e.g. 'tidal' signals) and 'turbulence' (i.e. convective injection).
At global scales, very recent studies based on the space-borne Microwave Limb Sounder  measurements of water vapour (H2O) and ice crystals (Fig. 3) have shown that 1) the tropical deep overshooting convection is the most intense above continental areas such as South America, Africa, and the maritime continent, and 2) the diurnal cycle of the convection impacts differently the H2O budget in the TTL depending on the height (upper troposphere, cold point, lower stratosphere), the surface (land/ocean), the latitude (North/South), and the nature of the surface (forest/field).  


Figure 3: Mean relative difference between the daytime and night-time Microwave Limb Sounder H2O measurements for 8 years (2005–2012) in the upper troposphere for December-February (left, southern latitude convective period) and June-August (right, northern latitude convective period) [from Carminati et al., 2014].

Based on our experience on this subject, the present project intends to study in more detail and in very localized areas the diurnal evolution of the convective activity impacting the H2O budget in the TTL based on space-borne observations and mesoscale modelling studies. Firstly, we will base our study on the MLS measurements of H2O, ice and also carbon monoxide (a dynamical tracer) in the TTL considering more than 10 years of measurements. We will focus on particular geographical domains where we have already detected some signals in the diurnal variation of H2O, e.g. South America (Amazonian forest and Brazilian fields), Maritime continent. Within these boxes, we will then study the diurnal evolution of the convective activity through the evolution of H2O, ice and CO along the vertical in the TTL, amplitudes and phases. Secondly, once a domain and a period of interest are selected, we will perform a case study using the mesoscale MESO-NH model that will be able to follow precisely the diurnal cycle of convection and its impact on H2O, ice, and CO. Finally we will compare these modelling results with the observations to assess how the different processes studied in the overall projects (waves, multiscale turbulence and mixing) impact on the diurnal cycle of H2O in the TTL and assess the broader scientific implications for the relative importance of different processes on longer time scales.

Actions sur le document