My Research

Home

About me

My research

Ocean web sites

My research interests:

1- Mesoscale activity in the ocean

2- Submesoscale activity in the ocean
3- Regional numerical modeling of the California Current System.
4- Basin-scale modeling
5- Lagrangian description of oceanic phenomena (advection/dispersion/retention and application to living organisms - eg, diel vertical migration).
6- Regional modeling of the Peru-Chile upwelling System and Alaskan region (in collaboration with F. Colas).

Mesoscale activity in the ocean

From energy spectra, it is quite clear that the ocean dynamics tends to be very active at length scales of 30 to 300 km (typical time scales are of the order of several days to weeks). From this raises the interest of oceanographers for what is known as mesoscale activity. No forcing input can by itself explain the energetic peak corresponding to the mesoscale, which is linked to certain fundamental aspects of a geophysical fluid:

Due to the combined effects of ocean stratification and Earth rotation, geophysical fluids have natural intrinsic length scales called internal deformation radii. The largest of these radii (and also the most important) is typically of the order of tens of kilometers in the ocean and often noted Rd.
In the ocean, Reynolds number are of the order of 10^6. Therefore, the molecular viscosity can not effectively mix momentum, heat and salt. This implies that the ocean dynamics is strongly nonlinear i.e., it is associated with intense inteactions and transfers between all motion time and length scales.As a result, large scale energy inputs associated with atmospheric and solar forcings still tend to cascade toward smaller scales, but through nontrivial ways.
The ocean constitutes a thin layer of fluid in which any motion is nearly bidimensionnal as soon as its horizontal extension reaches a few kilometers. Turbulence in such environments is very different from 3d turbulence. Notably, there exists an "inverse cascade" that drives energy from small scales toward larger scales. In the ocean it has been shown that this inverse cascade becomes uneffective beyond a length scale called Rhines scale, which is at most about 10Rd.

It result from these 3 points that mixing in the ocean does not simply follow a Fickian law and that it produces energy accumulation at scales consistent with Rd.

From the phenomenological viewpoint, an important mechanism associated with remark 2 mentionned above is the destabilization of ocean currents. An example is shown in the right figure. The Gulf Stream flows northeastward on average but an instantaneous snapshot also reveals its meandering nature. When the meandears reach large ampltudes, they can isolate and and shed eddies (Gulf Stream rings) that are visible from space. These eddies are the oceanic equivalent of the well-know (in countries like France !) atmospheric mid-latitude depressions. They have radii of the order of a few Rd (from 3 to 10). Most ocean currents, whether flowing in surface or not. generate eddies.

My Ph.D. work mostly consisted in undestand ing and describing how certain characteristics of processes or flow characteristics favor or prevent the development of baroclinic instability, which is known as an eddy formation mechanism.

Infra red satellite picture from the Gulf Stream off the East Coast of the US. Meandering activity and well defined eddies are visible. The size of the domain is about 600km square.

Submesoscale activity in the ocean

As its name suggests, submesoscale is defined by length and time scales smaller than the mesoscale. Precisely, the submesoscale range encompasses scales between mesoscale and microscale where rotation and stratification are still important (like for mesoscale) but yet departure from geostrophy (ie, balance between Coriolis and pressure forces) can be large. Typically, this class of motions is associated with length scales of the order of kilometers and time scales of hours to a few days. Observing submesoscale phenomena in nature is quite challenging. Although important advances have resulted from several measurement programs (eg, FASINEX, and POMME more recently), a complete description of the submesoscale activity and its role in the ocean (in terms of dynamics and also biology) has yet to emerge. What we already know is that submesoscale activity in the upper ocean is quite energetic and almost ubiquitous on AVHRR and ocean color satellite images (see figures on the right and also the wiggles on the Gulf Stream rings above). Vertical velocities are maybe the most striking signature of the submesoscale, by contrast with the limited vertical motions associated with mesoscale activity. This is particularly true in an upwelling system as revealed by observations and also some of my numerical solutions [Capet et al, submitted]. A presentation I gave at Ocean Science 2006 is available for now.
Refs: Flament, P., L. Armi, and L. Washburn, 1985: The evolving structure of an upwelling filament. J. Geophys.Res., 90, 11765--11778.	SST snapshot from a California Curent System ROMS configuration at 0.750km resolution. Domain size and color coding are consistent with the AVHRR picture on the right.	Enhanced AVHRR image showing submesoscale features in the California Current system related to an upwelling filament (white is cold water). The size of the figure is about 200km by 200km. From Flament et al (1985).

Regional modeling of the California Current System

The California Current System is one of the main four upwelling systems and roughly goes from the southern tip of Baja California (Mexico) to the Oregon/Washington (USA) border - although differences exist between subregions of this area, that spans almost 30 degrees in latitude. Like other upwelling systems, it is a region of high primary productivity because of the cold and nutrient rich water upwelled at the coast and subsequently advected offshore by Ekman (ie wind-induced) currents. Beside environmental and climate issues, the turbulent nature of upwelling systems make them appealing regions, from an aesthetic and fluid dynamics standpoint. In addition, upwelling systems also lends themselves to idealization quite naturally because there are subject to one dominant process (upwelling) as opposed to a myriad of them like in other places. So far my work has consisted in validating model climatologies against observations (satellite altimetry and CalCOFI line 67 off Monterey Bay). Model/data discrepancies have led to sensitivity studies, concerning winds (Capet et al, 2003), heat fluxed, boundary conditions and topography (manuscript in preparation). This research is the continuation of what was done by P. Marchesiello and P. Penven. It relies both on stand-alone and nested grids configurations. The figure on the right shows a SST snapshot from a USWC 5km horizontal solution.
Most importantly, such a numerical approach allows us to investigate the transport in the California System, eg in terms of heat, organisms (Carr et al, submitted); also the role of the mesoscale eddies in the transport (turbulent transport) can be estimated. ROMS online and offline (ROFF) trajectory submodels can be quite useful for that task.

A downscaling approach is also being implemented where the effect of basin-scale climatic signals (eg, El Nino) on an oceanic region like the California Current System are considered. The modeling strategy here is to compute medium resolution (50 to 25km horizontal resolution) Pacific solutions with full interannual variability that provide boundary conditions for regional domains having finer gridscale. So far the coupling is being done offline. Yet another paper is in preparation. A presentation I gave at the EPOC conference in 2004 is available though.

An old animation of sea level that was previously on my research page is still available here.

July SST snapshot from ROMS solution at 5km horizontal solution. The solution is forced by climatological winds derived from QuikSCAT scatterometer measurements. Note the presence of blue filaments carrying cold water away from the nearshore region.

Numerical modeling of the Pacific on decadal time scales

Although there are existing large scale solutions readily available (eg., SODA POP reanalysis), it is quite handy to have configuration you run at will. A. Shchepetkin, F. Colas and I are taking care of a home-made 50 and 25km horizontal resolution Pacific solution. Although none of us can fully dedicate himself to the task of basin scale modeling, multiple solutions have been computed over the period 1970-2000. They look quite reasonable (compared to others, see below) and are continuously improving. Our primary interest is to force regional configurations for the US West Coast and Peru -Chile regions and we are paying particular attention to the well-sampled 1997-1998 El Nino event (see the model/data comparison on the right).
	November 1997 sea level anomaly from a ROMS Pacific 50km horizontal resolution. It is about 20% too weak relative to the data.	November 1997 DUACS sea level anomaly (merged TOPEX ERS altimeter data).