Search Results

You are looking at 1 - 10 of 119 items for

  • Author or Editor: David Randall x
  • Refine by Access: All Content x
Clear All Modify Search
David Randall
Full access
David A. Randall
Full access
Harshvardhan and David A. Randall

Abstract

No abstract available.

Full access
David A. Randall

Abstract

It is shown that the radiative cooling of a cloud layer strongly influences the turbulent flux profiles and the entrainment rate, and that the radiative cooling should be modeled as acting inside the turbulent layer. Numerical experiments demonstrate that a cloud-topped mixed-layer model, similar to that of Lilly (1968), is quite sensitive to δpR, the depth of the radiatively cooled layer near cloud top. As δpR increases, the model’s sensitivity to the entrainment assumption is markedly heightened. More specifically, for large δpR the cloud top and cloud base rise dramatically as the entrainment parameter k is increased, while for small δpR an increase in k has almost no effect. The model is most sensitive to ΔpR precisely for the cold-water, strong-divergence regime of greatest interest.

Full access
David A. Randall

Abstract

Entrainment closure theories for mixed-layer models entail assumptions about how the net rate of buoyant production of turbulence kinetic energy is partitioned into gross production and consumption. Three alternative partitioning theories are examined in this paper: Eulerian partitioning, process partitioning and Lagrangian partitioning. Lagrangian partitioning provides a definition of the gross production rate, but is difficult to implement directly. Eulerian and process partitioning are attempts to implement Lagrangian partitioning indirectly.

For the buoyancy fluxes due to a single family of plumes, Eulerian and Lagrangian partitioning are shown to be equivalent. Recent observations reported by Wilczak and Businger rule out such a model. However, it serves as a useful conceptual link between Eulerian and Lagrangian partitioning.

Process partitioning can be formulated in a variety of ways. Examples show that mixed-layer model results are very sensitive to the way in which radiative cooling is assumed to influence the production and consumption rates. A quantitative relationship between process partitioning and Lagrangian partitioning has yet to be established. The observations of Wilczak and Businger show that consumption and entrainment are not as closely linked as current versions of process partitioning suggest.

Full access
David A. Randall

Abstract

Numerical simulation of geostrophic adjustment in shallow water is discussed for the case of an unstaggered grid for vorticity, divergence, and mass. The dispersion equation is shown to be very well behaved and superior to that obtained with the Arakawa grids A–E.

Full access
David A. Randall

Abstract

A simple convective mass flux model is used to derive expressions for the fluxes of liquid water and buoyancy in partly cloudy turbulent layers. The results differ radically from those suggested in some previous studies. Physical interpretation is given, and examples are presented. Implications for the dynamics of partly cloudy boundary layers are discussed, and the aftermath of cloud-top entrainment instability is analyzed.

Full access
David A. Randall

Global atmospheric models are proliferating, in part because of the widespread availability of powerful computers. There are about two dozen global modeling groups at work in the United States today. These groups are put into four categories, considering both laboratories and universities and development and applications. Community models are a special subgroup and in principle are both developed and applied by the community. Most U.S. global modeling groups are focusing on applications rather than on development. This is especially true in the university community, although over the years university groups have made important contributions in the model-development arena. A key role of university groups is to train new model developers at a rate matched to the community's demand for such scientists. A simple but functional conceptual organization of the U.S. global modeling community is suggested.

Full access
David A. Randall

Abstract

It is shown that entrainment leads to the generation of turbulence kinetic energy in a stratocumulus layer when the virtual temperature jump at the cloud top is weaker than a critical value. The critical value increases as the relative humidity of the air above cloud top decreases. This result is interpreted as a criterion for the instability of the layer cloud to penetrative downdrafts. The role of the instability in determining the subtropical and tropical distributions of boundary-layer cloudiness is assessed.

Full access
Marat Khairoutdinov and David Randall

Abstract

Results are presented from a high-resolution three-dimensional simulation of shallow-to-deep convection transition based on idealization of observations made during the Large-Scale Biosphere–Atmosphere (LBA) experiment in Amazonia, Brazil, during the Tropical Rainfall Measuring Mission (TRMM)-LBA mission on 23 February. The doubly periodic grid has 1536 × 1536 × 256 grid cells with horizontal grid spacing of 100 m, thus covering an area of 154 × 154 km2. The vertical resolution varies from 50 m in the boundary layer to 100 m in the free troposphere and gradually coarsens to 250 m near the domain top at 25.4 km. The length of the simulation is 6 h, starting from an early morning sounding corresponding to 0730 local time. Convection is forced by prescribed surface latent and sensible heat fluxes and prescribed horizontally uniform radiative heating

Despite a considerable amount of convective available potential energy (CAPE) in the range of 1600–2400 J kg−1, and despite virtually no convective inhibition (CIN) in the mean sounding throughout the simulation, the cumulus convection starts as shallow, gradually developing into congestus, and becomes deep only toward the end of simulation. Analysis shows that the reason is that the shallow clouds generated by the boundary layer turbulence are too small to penetrate deep into the troposphere, as they are quickly diluted by mixing with the environment. Precipitation and the associated cold pools are needed to generate thermals big enough to support the growth of deep clouds. This positive feedback involving precipitation is supported by a sensitivity experiment in which the cold pools are effectively eliminated by artificially switching off the evaporation of precipitation; in the experiment, the convection remains shallow throughout the entire simulation, with a few congestus but no deep clouds.

The probability distribution function (PDF) of cloud size during the shallow, congestus, and deep phases is analyzed using a new method. During each of the three phases, the shallow clouds dominate the mode of the PDFs at about 1-km diameter. During the deep phase, the PDFs show cloud bases as wide as 4 km. Analysis of the joint PDFs of cloud size and in-cloud variables demonstrates that, as expected, the bigger clouds are far less diluted above their bases than their smaller counterparts. Also, thermodynamic properties at cloud bases are found to be nearly identical for all cloud sizes, with the moist static energy exceeding the mean value by as much as 4 kJ kg−1. The width of the moist static energy distribution in the boundary layer is mostly due to variability of water vapor; therefore, clouds appear to grow from the air with the highest water vapor content available.

No undiluted cloudy parcels are found near the level of neutral buoyancy. It appears that a simple entraining-plume model explains the entrainment rates rather well. The least diluted plumes in the simulation correspond to an entrainment parameter of about 0.1 km−1.

Full access