[e-mail]

1
On the Relationships of Large to Small Scales in
a Convecting Atmosphere in the Real World and a
CRM
L. Davies, C. Jakob, V. Kumar, P. May and A.
Protat
School of Mathematical Sciences, Monash
University, Australia.
laura.davies_at_monash.edu
Poster available at users.monash.edu.au/ladavies
/current.html
1. Motivation What relationships exist between
the large-scale atmosphere and rainfall?
The role of convection parameterization is to
represent sub-grid convective processes in
relation to the large-scale state of the
atmosphere. In previous work, we have
investigated the large-scale to small-scale
relationships using a three-year data set at
Darwin derived using a variational analysis
approach (large scales) and C-band radar data
(convective scales). Here, the C-band radar data
is further used to identify congestus clouds.
Congestus clouds are defined as convective cells
with cloud tops below 7km whereas convective
cells with cloud tops above 7km are defined as
deep. A cloud top is defined as the maximum
height of the 5dbz echo above a radar pixel
identified to contain convective rainfall.
2. How much congestus cloud is there and how does
it relate to total rainfall area?
Cloud type Stratiform Convective Congestus
Frequency of occurrence 0.88 0.86 0.85
Mean area fraction when present 0.1186 0.0105 0.0016
Figure 1 Relative area fractions for stratiform,
convective and congestus cloud before and after
heavy rainfall. The relative area is defined as a
percentage of the mean over the 36hr period for
each cloud type.
Table 1 Relative Frequency and area fraction
when present for stratiform, convective and
congestus cloud types.
While all cloud types occur frequently in time,
their area fractions span two orders of
magnitude, with congestus clouds covering a very
small area when present. During heavy rainfall
events (the top tercile of the rainfall
distribution) congestus clouds lead deep
convection, which in turn leads stratiform
clouds. Congestus and deep convection cease to
exist quickly after the rain event while
stratiform clouds persist longer.
3. The relationship of cloud type occurrence and
mid-tropospheric humidity
Figure 3 Area fractions for stratiform,
convective and congestus for terciles of 600mb
moisture (a-c). Congestus area also shown as a
fraction on total convective area d.
All three cloud types increase in their area
fraction as mid-tropospheric humidity increases
(a-c). The largest relative increase is
associated with stratiform clouds (a). This could
be the result of evaporating rain from stratiform
clouds, but could also indicate that stratiform
clouds require a moist environment to exist.
Congestus clouds cover an order of magnitude less
area than deep convective clouds and their
relative contribution to total convective area
decreases with increasing mid-tropospheric
moisture (d). Given this behaviour, there is no
clear evidence for congestus playing a crucial
role in moistening the mid-troposphere.
4. Can a CRM capture these relationships?
Observed domain-mean rainfall is strongly related
to moisture convergence (4a), a relationship that
is largely realized through an increase in rain
area with increasing moisture convergence (4b).
While the CRM shows qualitatively similar
relationships (4c and d), there some notable
differences. The model misses some of the heavier
rainfall events and there are fewer rainfall
events under divergent large-scale conditions.
Rather than a model error, this could be the
result of the simplified model setup in the
topographically complex Darwin area. Overall the
model does provide a useful framework for
studying the underlying physics for the observed
relationships.
Figure 4 Relationships between area-average
large-scale moisture convergence and
domain-average rainfall (a and c) and rain area
fraction (b and d) from observations (a and b)
and the SAM CRM (c and d).