Head in a Cloud

troposphere and stratosphere meet blogosphere

Head in a Cloud random header image

A Summary of “Effect of Dynamics on Mixed-Phase Cloud: Theoretical Considerations”- Korolev & Field

February 11th, 2008 by BDA · 2 Comments

Editor’s Note: The following is the first in a series of posts by undergraduate students in the course “Physical Meteorology: Atmospheric Radiation and Cloud Physics” that I am teaching at Metro State College of Denver. I hope you enjoy! -Sean Davis
The following is a brief overview of the paper “The Effects of Dynamics on Mixed-Phase Clouds: Theoretical Considerations” written by Alexei Korolev and Paul R. Field which appears in the January 2008 issue (Volume 65 pg. 66-86) of Journal of the Atmospheric Sciences. This paper furthers existing research of mixed-phase clouds, specifically the activation of liquid water within non-equilibrium state ice clouds. The liquid-activation process within ice clouds is complex process is dependent on many variables. An accurate parameterization will increase model validity especially affecting precipitation prediction and aviation safety.

In a mixed-phase environment basic thermodynamics tells us that because the saturation vapor pressure over ice is less than that over water, liquid water will deposit itself onto ice crystals until there is no more liquid water. However, direct measurements have detected stable liquid layers within ice clouds and this paper studies the necessary requirements for stable liquid layers to exist under three specific conditions which are uniform vertical ascent, harmonic vertical oscillation, and turbulent motion.

In general, there are two requirements for activating a liquid later. First, an area in the ice cloud must be at a sufficient vertical velocity. This vertical motion can be caused by dynamic processes such as updrafts, convection and gravity waves. The amount of vertical motion required to activate a liquid layer is a positive linear relationship between the velocity of the parcel and threshold velocity. Typical values range from particle concentrations of 10^-1 nm/cm^3 to 10^2 nm/cm^3 which then require velocities from 10^-3 m/s to 10^1 m/s in order to maintain a steady state. The threshold vertical velocity is a function of concentration of ice particles, the mean size of those particles, temperature, and pressure. Areas in an ice cloud at threshold velocity will not necessarily form a liquid layer. Vertical motion must also take the parcel to a sufficient altitude to obtain water saturation.

In addition to obtaining threshold velocity, a parcel must meet a second condition, threshold altitude, in order to achieve liquid saturation. From modeling done in the experiment, it was shown that even with a vertical velocity that exceeded the initial threshold velocity, the parcel may not reach the liquid adiabat due to the fact that ice particles will grow as the parcel rises. Since the velocity threshold equation is dependent upon the number of ice particles, if the velocity does not increase with height, the parcel will rise only as far as the liquid adiabat and no liquid layer will form. If the velocity is sufficiently large to reach and extend past the liquid adiabat, liquid water will be activated.

For the case of uniform vertical motion, the liquid layer will then continue to grow as the parcel continues to rise in the new mixed-phase environment. Eventually, (after about 250 m in the model), ice growth will increase to a point such that the relative humidity will drop below liquid saturation. The liquid layer will begin to evaporate and deposit onto the ice crystals at approximately the same rate as liquid water was activated. Modeling has shown that a higher initial temperature requires both a lower threshold velocity and altitude. For example, and parcel at -5C requires a velocity of approximately 0.22 m/s and a 200 m change in height to begin to activate a liquid layer, while a parcel at -20C requires a velocity of about .75 m/s and a change in height of nearly 600 m. Therefore, liquid layer are more likely from areas with higher initial temperatures.

For the special case of vertical harmonic oscillation, the general principles apply with a few twists. A parcel must have a threshold tangential velocity and must still rise to an altitude that exceeds the moist adiabat in order to activate the liquid layer. Modeling the harmonic rise and fall of a parcel can be thought of like a slinky. The “slinky” must be moving fast enough to get the parcel back above the wet adiabat before phase changes in the parcel due to the slinky going up and down prevent a mixed layer from forming. If the time it takes a parcel to rise and fall and rate of phase relaxation are close to each other, the parcel will be in quasi-equilibrium and thus no activation of liquid water will occur and the ice water mixing ratio will be close to the wet adiabat, not good for getting liquid water. If the tangential velocity is much larger, the slinky is moving up and down quickly, a mixed layer will form whenever the harmonic motion takes the parcel above the wet adiabatic just like in the simple rising motion example.

Since, these two examples are simple compared to actual atmospheric motions, an attempt at modeling turbulent motion was also done. To model turbulence, typical and random updraft velocities were used. After the cycle was run, it was shown that harmonic motion was more produced more liquid water than the turbulent model.

In summary, in order for a mixed layer of liquid water and ice to form in an ice cloud, an area of that cloud must have a sufficient upward velocity which takes it to a sufficient altitude in order to form a mixed layer. These threshold requirements change based on the properties of the parcel. Areas of uniform vertical motion will create a three stratified lawyers (ice, mixed, ice) as you gain altitude. Harmonic and turbulent motions will form layers with the water layer at the top of the cloud. Layers formed from turbulent flow will not be have a clean boundary between the ice and liquid layer due to the variability of the vertical velocities can created pockets of liquid water. In terms of parameterization for modeling, a mix between a turbulent and harmonic solution best represents actually motions in the atmosphere.

For my own two cents, I found this paper very interesting, if a little over my undergraduate head at times. However the authors did a great job of explain the ‘jist’ of what was going on without my needing to spend too many brain-byes on the equations. Here’s a couple of questions I had.

1. They say in the paper that liquid regions with horizontal extents as short as 100m can be formed. However, they never give the other end to the scale and I never got a good idea of how much liquid water can actually be produced by these processes.

2. Going of off #1, dependant on the size of the liquid layer, what actual impact would these layers have on a weather system and what scale would they affect? How big would they have to be to have an effect?


Dan “BDA” Kulp

Tags: modeling · MTR3440

2 responses so far ↓

  • 1 Sean Davis // Feb 12, 2008 at 8:31 am


    I emailed Alexei about this article, and he sent me a couple of replies to your questions, which I’ve put below:

    RE #1 Liquid adiabate limits max condensed cloud water.

    RE #2 e.g. radiation transfer will definitely be affected due large amount of small droplets, enhanced growth of ice due to riming (not considered in the paper)

  • 2 Kramer // Feb 12, 2008 at 5:16 pm


    Just wanted to give this entire paper one big SO TO SPEAK, expecially this line

    For the special case of vertical harmonic oscillation, the general principles apply with a few twists. A parcel must have a threshold tangential velocity and must still rise to an altitude that exceeds the moist adiabat in order to activate the liquid layer. Modeling the harmonic rise and fall of a parcel can be thought of like a slinky. – SO TO SPEAK

    - Kramer

Leave a Comment