Causes of the interannual variability of deep convection
Deep water formation in the Labrador Sea and the Gulf of Lion, for example, results from convection. A cyclonic circulation leads to a doming of the isopycnals at its centre, where stratification is then completely eroded by high surface winter buoyancy loss. This thesis assesses the causes of the i...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2009
|
| Subjects: | |
| Online Access: | https://eprints.soton.ac.uk/72147/ https://eprints.soton.ac.uk/72147/1/Grignon_PhD_2009.pdf |
| Summary: | Deep water formation in the Labrador Sea and the Gulf of Lion, for example, results from convection. A cyclonic circulation leads to a doming of the isopycnals at its centre, where stratification is then completely eroded by high surface winter buoyancy loss. This thesis assesses the causes of the interannual variability of deep convection. We first aim to quantify the relative importance of preconditioning, defined as the temperature and salinity structures and contents of the water column before the onset of convection, and of the buoyancy forcing (averaged over one winter) on the final convective mixed layer depth and on the temperature and salinity of the water mass formed. This study focuses on the Mediterranean and uses datafrom the Medar/Medatlas and Dyfamed data sets. The heat fluxes are studied and characterised. It is shown that the preconditioning is as important as the winter buoyancy fluxes in setting the final depth of convection. At the Dyfamed site (Corsica Strait), the seasonal cycle shows that the stratification frequency reaches a maximum in the intermediate layer in winter. This winter maximum is thought to be of critical importance. The second (and main) part focuses on the effect of the short-term (O(day)) variability of the surface forcing on convection, using an idealised model. The MIT model is integrated over a square box of size 64km x 64 km x 2km initialised with homogeneous salinity and a linear vertical temperature gradient. The configuration of the model is described and validated. A time-periodic cooling is then applied over a disc of radius 20km at the centre of the surface of the box. It is shown that the final mixed layer depth depends little on this short-term time variability because the lateral buoyancy fluxes are very responsive to the surface ones. Our results are compared with traditional parameterisation of the lateral buoyancy fluxes. General characteristics of the patch are also looked at, such as the rim current, the location of the angular momentum surfaces, the ... |
|---|