Errors in the Simulated Heat Budget of CGCMs in the Eastern Part of the Tropical Oceans

Thursday, 18 December 2014
Julia Hazel1, Matthew T Masarik1, Carlos R Mechoso1, R. Justin Small2 and Enrique N Curchitser3, (1)UCLA, Los Angeles, CA, United States, (2)National Center for Atmospheric Research, Boulder, CO, United States, (3)Rutgers University New Brunswick, Institute of Marine and Coastal Sciences, New Brunswick, NJ, United States
The simulation of the tropical climate by coupled atmosphere-ocean general circulation models (CGCMs) shows severe warm biases in the sea-surface temperature (SST) field of the southeastern part of the Pacific and the Atlantic (SEP and SEA, respectively). The errors are strongest near the land mass with a broad plume extending west, Also, the equatorial cold tongue is too strong and extends too far to the west. The simulated precipitation field generally shows a persistent double Inter-tropical Convergence Zone (ITCZ). Tremendous effort has been made to improve CGCM performance in general and to address these tropical errors in particular.

The present paper start by comparing Taylor diagrams of the SST errors in the SEP and SEA by CGCMs participating in the Coupled Model Intercomparison Project phases 3 and 5 (CMIP3 and CMIP5, respectively). Some improvement is noted in models that perform poorly in CMIP3, but the overall performance is broadly similar in the two intercomparison projects. We explore the hypothesis that an improved representation of atmosphere-ocean interaction involving stratocumulus cloud decks and oceanic upwelling is essential to reduce errors in the SEP and SEA. To estimate the error contribution by clouds and upwelling, we examine the upper ocean surface heat flux budget. The resolution of the oceanic component of the CGCMs in both CMIP3 and CMIP5 is too coarse for a realistic representation of upwelling. Therefore, we also examine simulations by the Nested Regional Climate Model (nRCM) system, which is a CGCM with a very high-resolution regional model embedded in coastal regions. The nRCM consists of the Community Atmosphere Model (CAM, run at 1°) coupled to the global Parallel Ocean Program Model (POP, run at 1°) to which the Regional Ocean Modeling System (ROMS6, run at 5-10 km) is nested in selected coastal regions.