The process-oriented evaluation of Southern Ocean cloud and radiation errors in weather and climate models using cloud regimes

• Main Author: Mason, Shannon
• Format: Thesis
• Language: unknown
• Published: Monash University 2017
• Subjects: Uncategorized; Southern Ocean
• Online Access: https://dx.doi.org/10.4225/03/58b3c1c71bf31; https://bridges.monash.edu/articles/thesis/The_process-oriented_evaluation_of_Southern_Ocean_cloud_and_radiation_errors_in_weather_and_climate_models_using_cloud_regimes/4697503

Cloud and shortwave radiation (SW) biases over the high-latitude Southern Ocean (SO) are persistent in many general circulation models (GCMs), and have been associated with large-scale circulation errors and uncertainties in cloud feedbacks in a warming climate. Physical processes below the grid-scale of GCMs are represented by parameterisations; to prioritise improvements in model physics there is a need to identify which processes are the major contributors to model error. A promising approach to process-oriented model evaluation is to identify cloud regimes from passive satellite observations, the occurrence of which are functions of the large-scale weather state. The aims of this thesis are, first, to expand upon approaches to identifying cloud regimes and their associated weather states, and to more fully characterise the vertical structure and thermodynamic phase of cloud regimes in the Southern Ocean. We then apply this observational knowledge to model evaluation, so as to develop a cloud regime-oriented framework for identifying and quantifying cloud and radiation errors in GCMs, and their associations with processes represented by model physics. In this thesis we present significant and original contributions to the knowledge of SO cloud from observations, and to the use of cloud regimes to evaluate model errors. We identify robust distinctions between key midtopped cloud regimes that have been associated with errors in the SO, and develop a methodology for using active satellite observations to quantify the subgrid-scale variability of cloud structures within cloud regimes. We develop a novel approach to identifying hybrid cloud regimes from both observations and a GCM, and demonstrate applications to quantifying the effects of changes to model parameterisations. We use the hybrid cloud regime approach to compare the development of cloud and radiation errors between timescales by evaluating short-term weather hindcasts and long-term climate simulations of the same GCM. Key methodological contributions to the use of cloud regimes for observations and model evaluation are demonstrated in the context of the high-latitude SO, with promising applications to ongoing efforts to address SO cloud errors in GCMs, and to investigate model errors in other contexts.