North Pacific upper-ocean changes projected by CMIP5 models

Abstract

IPCC (Intergovernmental Panel on Climate Change) recently have recently released its fifth assessment report (AR5) which provides up-to-date scientific knowledge and socio-economic aspects of climate change, largely based on observational data and CMIP5 (Coupled Model Intercomparison Project Phase 5) global models. In this study, we aim to evaluate performance of global climate models (CMIP5 models) by comparing their historical run simulation with observed climatology, and to analyze their future climate change projection, focusing on North Pacific upper ocean changes that are associated with ecosystem response to global warming. Statistical analyses including Taylor diagram show that CMIP5 models have improved spatial patterns of PDO (Pacific Decadal Oscillation) compared with those of CMIP3 models, mainly through better teleconnection representation between the tropics and mid-latitudes. Furthermore, the improvement is not only by decrease in number of models with poor performance, but also by better simulation of PDO spatial patterns. Sea surface temperature and mixed layer depth in the North Pacific Ocean, however, still appears to have significant biases, thus contributing to simulation biases in ecosystem including estimates of chlorophyll concentration. CMIP5 models project that upper ocean processes including mixed layer depth tend to change considerably in the North Pacific, as projected by CMIP3 models. Poata and CMIP5 (Coupled Model Intercomparison Project Phase 5) global models. In this study, we aim to evaluate performance of global climate models (CMIP5 models) by comparing their historical run simulation with observed climatology, and to analyze their future climate change projection, focusing on North Pacific upper ocean changes that are associated with ecosystem response to global warming. Statistical analyses including Taylor diagram show that CMIP5 models have improved spatial patterns of PDO (Pacific Decadal Oscillation) compared with those of CMIP3 models, mainly through better teleconnection representation between the tropics and mid-latitudes. Furthermore, the improvement is not only by decrease in number of models with poor performance, but also by better simulation of PDO spatial patterns. Sea surface temperature and mixed layer depth in the North Pacific Ocean, however, still appears to have significant biases, thus contributing to simulation biases in ecosystem including estimates of chlorophyll concentration. CMIP5 models project that upper ocean processes including mixed layer depth tend to change considerably in the North Pacific, as projected by CMIP3 models. Po