Size Decompositions of Observed and Simulated Shallow Convective Cloud Fields

Principal Investigator(s):
Thijs Heus, Cleveland State University

Parameterizations of boundary-layer clouds remain at the heart of important problems in large-scale models for  weather and climate prediction. A long-standing problem is the unification of the representation of different boundary-layer cloud regimes. A new problem is that processes are becoming partially resolved at present-day resolutions, meaning that parameterizations of boundary layer clouds that assume that all clouds are smaller than the grid size are no longer valid. Ideally, one accommodates for this dynamically through the formulation of scale-dependent parameterizations, which can be achieved for boundary layer clouds through a formulation in terms of size distributions. This cloud size distribution then needs to be constrained by a generally applicable closure. The aim of this proposal is to find a better understanding of how the cloud size distribution is being set, and how it impacts vertical transport of heat and moisture, as well as precipitation. Our specific focus will be on how a variety of cumulus cloud properties scale as a function of cloud size. This includes cloud geometry, precipitation, condensate loading, dynamics, transport and mixing. One of the most important parameters that influences many of those cloud properties is entrainment, the amount of mixing between cloudy air and environmental air. By studying a wide range of different observable properties as a function of cloud size, at multiple sites and covering long periods, we can construct a model that is best capable of reproducing all observed distributions combined. This holistic approach avoids parameterizations that do well in predicting one variable (e.g., cloud height), but less in others (e.g., precipitation).

Many of the current studies on cloud size distributions are based on ultra-fine numerical simulations (LES). While those are often the most convenient tool to retrieve a cloud size distribution, it has to be stressed that the cloud size distribution varies significantly between different LES models. Therefore, there is a critical need to focus on observations if the cloud size distribution is to be understood. Recent single-case LES studies have questioned the common assumption that entrainment is a function of cloud size. While such LES studies of single cases do provide insight, any generally applicable conclusion requires analysis of a large variety of data sets, preferably from observations. Precipitation is also hypothesized to scale with cloud size, but only after a certain threshold size is met. We will link entrainment and precipitation to the cloud size, across a range of different cases, to cover a broad parameter space in terms of atmospheric state, surface fluxes and aerosol concentration.

This project aims to investigate cloud size distributions under a wide range of meteorological conditions. High-frequency, multi-dimensional long-term ARM measurements are combined with semi-idealized LES of the same scenes. The main research goals are:

  1. To gain insight into the mechanisms that determine the shape and time-development of size distributions of shallow convective cloud properties
  2. To understand relationships between cloud size, transport, and microphysics
  3. To thus constrain a new class of parameterizations for models. A new boundary layer scheme based on size distributions will be calibrated against ARM data and tested in the WRF model. New insights resulting from the proposed activities will be directly implemented and tested on their impact on model predictions.