Untangling the Mistral and seasonal atmospheric forcing driving deep convection in the Gulf of Lion: 2012–2013

Abstract

Abstract. Deep convection in the Gulf of Lion is believed to be primarily driven by the Mistral winds. However, our findings show that the seasonal atmospheric change provides roughly 2/3 of the buoyancy loss required for deep convection to occur, for the 2012 to 2013 year, with the Mistral supplying the final 1/3. Two NEMOMED12 ocean simulations of the Mediterranean Sea were run for the Aug. 1st, 2012 to July 31st, 2013 year, forced with two sets of atmospheric forcing data from a RegIPSL coupled run within the Med-CORDEX framework. One set of atmospheric forcing data was left unmodified, while the other was filtered to remove the signal of the Mistral. The Control simulation featured deep convection, while the Seasonal did not. A simple model was derived, relating the anomaly scale forcing (the difference between the Control and Seasonal runs) and the seasonal scale forcing to the ocean response through the Stratification Index. This simple model revealed that the Mistral's effect on buoyancy loss depends more on its strength rather than its frequency or duration. The simple model also revealed that the seasonal cycle of the Stratification Index is equal to the net surface heat flux over the course of the year, with the stratification maximum and minimum occurring roughly at the fall and spring equinoxes.

Untangling the Mistral and Seasonal Atmospheric Forcing Driving Deep Convection in the Gulf of Lion: 1993-2013

“…For the seasonal response, we further make the assumption that N 2 S negligibly depends on the x, y, and z coordinate directions: (Drobinski et al, 2012). Code used to generate figures and implement the simple model can be found at the following Git repository: https://gitlab.com/dkllrjr/untangling_deep_conv_2012_ 2013_code (Keller, 2022).…”

Untangling the mistral and seasonal atmospheric forcing driving deep convection in the Gulf of Lion: 2012–2013

“…The RegIPSL model and the NEMO model and its configuration used in this paper can be accessed at the following Git repository: https://gitlab.in2p3.fr/ipsl/lmd/intro/regipsl/ regipsl (Drobinski et al, 2012). Code used to generate figures and implement the simple model can be found at the following Git repository: https://gitlab.com/dkllrjr/untangling_deep_conv_2012_ 2013_code (Keller, 2022).…”