A Process-Based Assessment Of Global Decadal Climate Variability And Its Model Prediction Skill

The global decadal climate variability can change the global climate background and modulate the interannual climate variability,which has important influence on the medium and long-term development of society and economy.Therefore,the decadal climate prediction is critical for the assessment of climate change risks and the design of mitigation and adaptation strategies,facing with urgent technical problems and scientific questions in the initial stage.This study represents an initial effort in the context of the coupled atmosphere-surface climate feedback-response analysis method?CFRAM?to partition the temporal evolution of the global and the East Asian continent?15°-70°N,70°-160°E?surface temperature into components associated with individual physical and dynamical processes in the NCAR CCSM4's decadal hindcasts.When compared with the observation?ERA-Interim?,it helps to seek the physical origins of prediction skill and the source of errors in the model physical and dynamical processes,thus further identify the problems at the process level in the model and provide effective guidance for model improvement.The global rapid warming occurred from the 1980s to late 1990s with a transient cooling after the Pinatubo eruption in 1991.The global surface temperature has experienced a warming slowdown after 2000s,known as the hiatus.The changes in surface dynamics?mainly ocean heat storage change?dictates the global decadal-scale surface temperature evolution from 1981 to 2015,with changes in CO₂,clouds,surface albedo,and water vapor providing secondary positive contributions at a decreasing order of magnitude.Atmospheric dynamics works against the surface temperature tendency associated with surface dynamics through turbulent and convective heat transport.Changes in the ocean heat storage leads to basin-wide ocean warming over the tropics.The rising CO₂ concentration provides a sustained and global warming contribution.Surface albedo effect is important for the polar regions.Impacts of solar irradiance with the quasi 11-year solar cycle and ozone changes on the surface temperature change are relatively weak during this period.The contribution from an individual physical or dynamical process to the global mean surface temperature change could change signs.Rapid warming over the East Asian continent occurred in late 1980s and early 2000s with a transient pause of warming between the two periods and a rapid cooling in late2000s.The most pronounced feature is the stronger warming over the East Asian continent than the global mean between two decades?2011-15 and 1981-85?,particularly over high-latitudes and East China.Surface dynamics?mainly energy process related to soil heat diffusion,snow/ice melting/freezing,and river runoff?and cloud effect dominates the evolution of surface temperature from 1981 to 2015,while atmospheric dynamics tends to work against surface temperature fluctuations.The changes in water vapor,CO₂,and surface albedo provide secondary positive contribution to the surface temperature evolution at a decreasing order of magnitude.Cloud provides positive contribution to the warming over southern Siberia and East China.Surface albedo effect is important for warming/cooling over high-latitude and high-elevation regions.Impacts of solar irradiance and ozone changes are relatively small.The spatial pattern of the interannual rapid warming/cooling receives major positive contributions from atmospheric dynamics,water vapor,clouds,and surface albedo,suggesting that the rapid warming/cooling discussed here share some common physical origins.Compared to the observation?ERA-Interim?,CCSM4 is able to predict the global mean surface temperature evolution from 1981 to 2005,with some problems in representing various physical and dynamical processes.The model captures fairly well the surface temperature changes due to changes in external forcings and atmospheric dynamics,e.g.,the surface temperature anomalies related to changes in solar irradiance and rising CO₂ concentration.However,the model could not predict the stratospheric ozone restoration in 1990s and the regional sea-ice changes in the polar regions,and tends to overestimate the impact of volcanic eruption on climate system.Most noticeable feature is the difference in the most critical process dictating the global surface temperature evolution:it is the surface dynamics?mainly ocean heat storage change?in the observation but water vapor in the model.In the model,the atmospheric water vapor uniformly increases over the globe,while the change in water vapor exhibits substantial spatial variations in the observation.The magnitude of cloud effect is considerably smaller in the model compared to that in the observation,due to the significantly underestimation of total cloud cover with more warm clouds and less ice clouds.The model misrepresented the processes related to changes in water vapor and cloud,suggesting the major problems in the water cycle of climate system.The net effect of surface dynamics?i.e.,when globally averaging the partial temperature differences?is underestimated by the model,indicating the potential problems in representing long-term changes of ocean dynamical and thermodynamical processes in the oceanic model.The model is able to predict an overall warming trend over the East Asian continent,but not the transient cooling in 1990s and regional cooling.Most of the skill can be attributed to the long-term warming trend.The model has an overly strong water vapor effect that dictates the surface temperature evolution over the East Asian continent.This is in contrast with the observation,where changes in surface dynamics?mainly energy process related to soil heat diffusion,snow/ice melting/freezing,and river runoff?dominate the actual temperature evolution.In the model,surface and atmospheric dynamics act respectively as the main positive and negative contributors to the surface temperature evolution,in a way similar to the observation.However,it is the water vapor effect providing secondary positive contribution to the overall temperature evolution at a decreasing order of magnitude,but clouds effect in the observation.The snow cover in the model prediction is substantially different from that in the observation.The reanalysis datasets are subject to varying degrees of uncertainty?e.g.,cloud fields?.High quality reanalysis data is required to make the accurate attribution of global decadal climate change and provides a good reference for model validation.