Stochastically forced mode water variability

• Main Authors: Hazeleger, W., Drijfhout, S.S.
• Format: Article in Journal/Newspaper
• Language: unknown
• Published: 1999
• Subjects: North Atlantic
• Online Access: https://eprints.soton.ac.uk/349197/

Substantial interannual to decadal variability is observed in the properties of subtropical mode water of the North Atlantic. In this study the response of mode water to stochastic atmospheric forcing is investigated in a numerical model. In a series of experiments the response is studied to different components of stochastic atmospheric forcing, such as wind stress, freshwater flux, and heat flux. The numerical model consists of an isopycnal ocean model with explicit mixed layer physics. The stochastic forcing is superimposed on the climatological forcing. The stochastic forcing function has an idealized form, but the amplitude, the spatial, and the temporal variability are based on observations. When a stochastic heat flux is applied, an atmospheric anomaly model is coupled to the ocean model. The geometry of the model is idealized and mimics the subtropical gyre of the North Atlantic. The stochastic wind stress forcing excites an internal mode in the mode water layer of the model. The response is characterized by the propagation of baroclinic waves. The spectrum of the response to stochastic freshwater flux is red. In the coupled model the stochastic heat flux forcing generates variability characterized by a dipole pattern in the mode water. The spectrum of the response is red and dominates the response to the stochastic wind stress and freshwater flux. The response is damped by an atmospheric feedback that consists of anomalous heat fluxes, depending on the SST anomalies generated by the stochastic forcing itself. Only stochastic heat flux forcing can generate mode water variability of the observed amplitude. A preferred timescale in mode water variability should be contained in the forcing itself or it may result from modes that could not be simulated by the present model.