Evaluating Convective Parameterization Schemes and Their Scale-awareness Using Simulated Convection in a Hierarchy of Models

University Of California-San Diego Scripps Inst Of Oceanography, La Jolla CA

Investigators

Abstract

This project will yield new insight into factors governing the onset and amount of convection. It will lead to a much better understanding of the strengths and weaknesses of existing representations of atmospheric convection in major climate models in the world. As the global climate model grid spacing decreases to ~10 km or less, representing atmospheric convection in such models is a great challenge. The proposed work to use cloud-resolving model simulation of atmospheric convection to evaluate the accuracy and scale-awareness of widely used algorithms for representing convection is innovative, and will pave the way for further improving global climate models. The representation of convection and associated clouds strongly affects the statistics of extreme events in both current and future climates. It is also a major source of uncertainty in climate change projection. Thus, the proposed research will contribute to the simulation and prediction of the occurrence of natural disasters associated with climate variability at different timescales. The project will also contribute to educating future generation climate scientists through training of a postdoctoral researcher and providing learning opportunities for undergraduate summer interns. The representation of the effects of atmospheric convection is one of the most challenging scientific issues in climate modeling. Over the last few decades, despite tremendous efforts going into improving the treatment of physical processes in climate models, major problems still exist in simulating important climate systems. These deficiencies are largely associated with the lack of accurate representation of atmospheric convection in the models. This project proposes to systematically investigate the criteria used for determining the onset of convection and assumptions that determine the amount of convective activity (known as closure assumptions) in representations of atmospheric convection using a hierarchy of numerical models. The goal is to evaluate the many different ways of representing atmospheric convection in state-of-the-science global climate models, thereby identifying their strengths and weakness and further improving them. To achieve this goal, simulations of atmospheric convection by fine-resolution numerical models that can resolve convection are used to evaluate the onset criteria and closure assumptions for convection. Statistical analysis methods, including skill score calculation, lead-lag correlation, and composite techniques, will be used to analyze the model output data. The simulations of convection will also be used to investigate whether the algorithms of representing convection in current global climate models can still be used when the grid spacing of the models decreases to ~10 km or less, the so-called grey zone scales, from the current spacing of ~100 km or larger. New ideas and formulations for representing convection in global climate models resulting from the analysis will be tested using the Community Atmosphere Model, CAM5.