WMOD00319 ‎(2)‎

• Title: Comparison of Cloud Model Predictions: A Case Study Analysis of One- and Two-Dimensional Models
• Date: 6/11/1975
• Weather Modification - Doc Type: Report

,~

..

~

Table 1. MODEL WORKSHOP INITIALIZATION DATA

August 10, 1973

August 17, 1973

Cloud properties
Updraft radius (km)
Base height (km)
CCN (C, K)**
Ice Nuclei (No./l at -160 C)

Environmental properties
Surface maximum temperature (0 C)
Corresponding dew point (0 C)

Average sub-cloud convergence
(sec-l)

1.0, 2.0*
1.2
4493.7, 0.29
0.1-0.15

1.0, 2.0*
1.4
1597.8, 0.37
0.1-0.15

30.3
19.7

29.4
21.4

-1. 0 x 10-4

-5 x 10-5

* Use these as standards of comparisons, other radii optional
**N(cm-3) = CSK, where S is supersaturation

The two-dimensional, time-dependent,
axi-symmetric models predicted tops within the
range of the observations and, in general, had
good predictions of cloud base. Their maximum
vertical velocities were larger, ranging from
20 to 29 sec-1 on August 10 and from 8 to
13 m sec-l on the 17th. In general, their clouds
were larger than the slab model clouds. Except for
one model, their maximum rain rates were in the
proper range and where they were run on both cases
they caught the proper tendency from the 10th to
the 17th. Also, two of the two-dimensional, time-
dependent, axi-symmetric models predicted hail on
the 10th and not on the 17th.

It should be noted that the two-
dimensional models predict a considerable amount of
additional information that is worthy of comparison
but could not be used because of the limited
observations.

In general then, the models predicted
the characteristics of convection on these 2 days
and gave quantitative information about rainfall
rates and quantities of rain from individual cells
in most cases and from the whole field of
convective clouds in two cases. The factor of
three difference in rainfall rate in the observa-
tions was predicted in some of the models. The
vertical velocity of the observations came from the
rise of echo heights, which is probably an under-
estimate of the vertical velocity.

The time for integration is indicated
in the last row of tables 1 and 2. It is seen that
the time range from a few seconds to 2 hours for
the one-dimensional models and up to 4 hours for
the two-dimensional, time-dependent models. The
axi-symmetrical models treat a single cloud,
whereas the slab models treat a field of clouds,
and their integration can then continue for longer
times. One of the axi-symmetric models has quite
detailed microphysics, including the ice stage,
and its time required was not greater than the
field-of-motion model.

Dr. Ackerman stressed the fact that
convection on both of these days was characterized
by a series of pulses, as is convection on most
days. However, the data requested from the

modelers did not include time-height histories,
although some of the presentations indicated a
spectrum of clouds. One field-of-motion model
predicted the evolution of the cloud-size scales
quite accurately. The other field-of-motion model
gave radar echoes, and those data (dashed lines)
have been plotted on figure 3, which also shows
the observed rise of echoes (solid lines) on
August 10 and August 17. The echo data from the
model were analyzed after the fact, being initial-
ized from the time of observed first ech". Again,
the general characteristics of the convection are
captured by the numerical simulations.

3.

INITIALIZATION AND VERIFICATION DATA
REQUIREMENTS

Specific observations required to
initialize and verify a model will vary according
to the complexity of the model. The following
observations, listed in general order of priority,
were considered important for initializing models:

a. "Representative" rawinsondes close in
time and space to the appearanCE! of the
first cloud, and serial releases every
2 hours thereafter.

b. Mesoscale Beta (50-150 km) divergence
measurements every 2 km up to 20 km.

c. Cloud condensation nuclei activity spectra
at cloud base and 5 km.

d. Cloud base temperature and size and magni-
tude of updraft velocity at cloud base
during developing stage.

e. Surface observations of temperature,
humidity, and winds from mesoscale Alpha
(10-50 km) network.

f. Ice nuclei activation spectra at cloud base
as a function of supersaturation and
temperature.

g. Giant nuclei (5-20 1Jm) number density at
cloud base.

h. Cloud cover.

Tl