Sensitivity of Atlantic meridional overturning circulation variability to parameterized Nordic Sea overflows in CCSM4

• Published in: Journal of Climate
• Other Authors: Yeager, Stephen (author), Danabasoglu, Gokhan (author)
• Format: Article in Journal/Newspaper
• Language: English
• Published: American Meteorological Society 2012
• Subjects: Abyssal circulation; Decadal variability; Oceanic variability; Parameterization; Meridional overturning circulation; Labrador Sea; Nordic Sea; North Atlantic
• Online Access: http://nldr.library.ucar.edu/repository/collections/OSGC-000-000-010-626; https://doi.org/10.1175/JCLI-D-11-00149.1

The inclusion of parameterized Nordic Sea overflows in the ocean component of the Community Climate System Model version 4 (CCSM4) results in a much improved representation of the North Atlantic tracer and velocity distributions compared to a control CCSM4 simulation without this parameterization. As a consequence, the variability of the Atlantic meridional overturning circulation (AMOC) on decadal and longer time scales is generally lower, but the reduction is not uniform in latitude, depth, or frequency-space. While there is dramatically less variance in the overall AMOC maximum (at about 35°N), the reduction in AMOC variance at higher latitudes is more modest. Also, it is somewhat enhanced in the deep ocean and at low latitudes (south of about 30°N). The complexity of overturning response to overflows is related to the fact that, in both simulations, the AMOC spectrum varies substantially with latitude and depth, reflecting a variety of driving mechanisms that are impacted in different ways by the overflows. The usefulness of reducing AMOC to a single index is thus called into question. This study identifies two main improvements in the ocean mean state associated with the overflow parameterization that tend to damp AMOC variability: enhanced stratification in the Labrador Sea due to the injection of dense overflow waters and a deepening of the deep western boundary current. Direct driving of deep AMOC variance by overflow transport variations is found to be a second-order effect.