Date of Award

May 2015

Degree Type

Thesis

Degree Name

Master of Science

Department

Mathematics

First Advisor

Vincent E. Larson

Committee Members

Clark Evans, Kyle Swanson

Keywords

Climate, Cloud, Parameterization

Abstract

The representation of clouds and turbulence remains one of the foremost challenges in modeling earth's climate system and continues to remain one of the greatest sources of uncertainty in future climate projections. Increased attention has been given to unifying cloud and turbulence parameterizations in order to avoid the artificial categorization of cloud and turbulence regimes. One such unified parameterization is known as the Cloud Layers Unified by Binormals (CLUBB). CLUBB is a single column model of clouds and turbulence that assumes subgrid scale variability can be represented by a joint probability density function (PDF) of temperature, moisture, momentum, and hydrometeors. An advantage of CLUBB's joint-PDF is that it allows for the interaction of microphysics and subgrid variability which may be important in unified parameterizations.

In order to improve any parameterization, like CLUBB, 'key' model errors must first be diagnosed. This is complicated by numerous feedbacks within the model. In order to elucidate 'key' errors in CLUBB's representation of warm-rain processes, a semiprognostic test was performed in which CLUBB's joint-PDF was supplied with 'perfect' moments derived from a cloud resolving model. An idealized case of the transition from shallow to deep convection over land was used. It was shown that CLUBB's assumed correlations between hydrometeors play a major role in CLUBB's microphysical budgets. It was also shown that for highly skewed cases, CLUBB's current joint-PDF closure may inadequately represent the marginals of the subgrid scale atmopheric state. Finally, CLUBB's assumption that the skewness of temperature and moisture are proportional to the skewness of vertical velocity may break down in highly skewed cases such as the one tested here.

Share

COinS