Sensitivity of High-Resolution Fully-Coupled Approximate Present Day Transient Climate Simulations to Initial Conditions.
Julie McClean1, David Bader2, Mathew E Maltrud3, Detelina Ivanova4, Katherine J Evans5, Qi Tang6, Milena Veneziani3 and Mark Taylor7, (1)Scripps Institution of Oceanography, La Jolla, CA, United States, (2)Lawrence Livermore National Laboratory, Livermore, United States, (3)Los Alamos National Laboratory, Los Alamos, NM, United States, (4)Climformatics, Inc, San Diego, United States, (5)Oak Ridge National Laboratory, Oak Ridge, TN, United States, (6)Lawrence Livermore National Laboratory, Livermore, CA, United States, (7)Sandia National Laboratories, Albuquerque, NM, United States
Abstract:
Fine-resolution climate simulations explicitly resolving or permitting weather-scale/mesoscale features and processes in the ocean and atmosphere are expected to provide improved projections of future climate due to their greater realism relative to standard climate models. However, their computer resource intensiveness dictates the use of alternative initialization approaches to standard climate model protocol. Here, the evolution of upper ocean conditions and ventilation processes in the high-latitude oceans of a suite of fine-resolution present day transient simulations approximating 1970-2015 climate change are compared with each other and observations. They were initialized from either a multi-decadal atmospheric reanalysis-forced ocean/sea-ice simulation or a centennial 1850 fully-coupled pre-industrial control (PICNTRL) simulation. The simulations from which the initial conditions were drawn are used to provide context for the evolution of the transients. The simulations use 0.1º Parallel Ocean Program (POP2) and CICE4 (sea-ice) models while the coupled simulation also uses 1/4º Community Atmosphere Model 5 - spectral element (CAM5-SE) and the Community Land Model 4 (CLM4).
Transients initialized from forced POP/CICE and the PICNTRL display cold and warm sea surface temperature (SST) biases in the high-latitude Northern Hemisphere, respectively, while all transients display warm SST biases in the Southern Ocean. The cold SST bias, which sets up immediately after coupling, is due to unrealistically shallow mixed layers inherited from POP/CICE causing weak convective mixing. The warm bias, also present in the PICNTRL, is primarily due to excessive absorbed solar radiation (ASR) at high latitudes. High-latitude mode water formation, and Atlantic meridional heat transport and overturning circulation all increase in the transients initialized from POP/CICE such that they approach and sometimes surpass the PICNTRL values to converge with values in the PICNTRL-initialized transient.
