Characterizing Cumulus Cloud Cover with Transilient Matrices

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

Model simulations of shallow cumulus clouds in the lower atmosphere demonstrate the sensitivity of predictions of future states to the representation of these clouds in the initial state of the atmosphere. Numerical schemes to handle these clouds in weather and climate models have been in existence for many years, yet new techniques continue to build upon and improve these schemes. This research will use a mathematical tool, called Transilient Matrices, to better characterize these clouds in the model's representation of the lower atmosphere and in turn improve forecasts of future weather and climate states based on these clouds. These improvements will in turn benefit society through an improved understanding of the sensitivity of weather and climate models to the characterization of shallow cumulus clouds in the lower atmosphere. The proposed research seeks to test hypotheses for the amount of cloud cover associated with shallow cumulus. One such hypothesis relates the amount of cloud cover to the rate of cloud detrainment by a timescale that depends on cloud size and eddy diffusivity; this scale differs from the typical scale used in climate models by two orders of magnitude. The expression for timescale will be tested and refined using a well established mathematical tool known as the Transilient Matrix (TM) to output from large-eddy simulation (LES) and observations from cloud radar. Since differences in shallow cumulus cloud cover among various Global Climate Models (GCMs) have been identified as major contributions to uncertainty in climate projections, this research has the potential of improving forecasts of future climate.