Thursday, 18 June 2015
Meridian Foyer/Summit (The Commons Hotel)
One of the most robust predictions of greenhouse warming simulations is that the equator-to-pole temperature difference will decrease at the lower levels of the troposphere, while it will increase at the higher levels. The equator-to-pole temperature difference strongly affects the eddy transport and storm track intensity and location. Linear theory of baroclinic instability (i.e., Eady model) predicts how the growth rate of instabilities is related to a meridional temperature gradient that is constant with height, but it does not predict how the growth rate is affected by a meridional temperature gradient that varies with height. We developed an iterative method to control the final mean temperature distribution of an idealized general circulation simulation driven by a Newtonian cooling scheme. This allows investigating the effect of the meridional temperature gradient, at different levels, on eddy characteristics. We find that an increased upper tropospheric temperature gradient causes a significant increase in eddy kinetic energy, eddy fluxes and as a consequence also surface winds. Furthermore, an increased upper tropospheric temperature gradient tends to excite eddies with faster wave speeds. In contrast, an increased lower tropospheric temperature gradient has a significantly smaller affect on eddy kinetic energy, and in some cases even decreases the eddy kinetic energy. Such a decrease might be related to the mid-winter minimum in eddy kinetic energy observed over the Pacific ocean. Using an idealized GCM we demonstrate that the temperature profile above the Northern Pacific ocean in January is accompanied with a decrease in eddy kinetic energy compared to April or November. Furthermore, our results regarding wave speeds imply that a possible cause of the increased phase speeds recently seen in reanalysis data is the increased high tropospheric temperature gradient.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner