WMOD00419

• Title: Convective Rain Rates and their Evolution During Storms in a Semiarid Climate
• Date: 8/8/1984
• Weather Modification - Doc Type: Report

1610

MONTHLY WEATHER REVIEW

VOLUME 112

50

NDCMP, SUMMER 1981 BOWMAN AND PARSHALL DATA

40
~
~

>
o
z
W 30
::l
"
W
a:
LL
W20
>
i=
<(
...J
W
a:
10

.......
.......

o
1

-GROWING PERIOD
- -DECAYING PERIOD
284 CLUSTERS

~. .

9 13 17
AVERAGE RAIN RATE (mm hr-t)

'-

FiG. 7. The relative frequency distributions of R for growing and decaying periods.

accuracy improves as integrations are made over
larger areas and over time periods approaching the
storm lifetime. In the Quadra data, where R was
computed by averaging over a long period of time,
there is no evidence that RERV versus ATI correla-
tions become higher if the averaging time interval
approaches the storm lifetime; however, there is
an increase of the correlation coefficient from 0.91
to 0.94 when the cell duration increases from 4 to
10 h.
A multiple linear regression analysis was performed
using the 1981 Bowman data consisting of 587 clus-
ters. The maximum radar-estimated rain volume
(MRERV) on an individual scan basis, the maximum
area (MAREA), the maximum reflectivity (ZMX),
and the maximum echo height (MEH) versus RER V

(for the lifetime of the duster), shows that the corre-
lation coefficient with MRERV is the highest (0.97),
followed by MAREA (0.95), ZMX (0.82) and MEH
(0.68). These results suggest that the total rain volume
of a storm may be estimated at its maximum stage
of development. The rain volume prediction equation
based on the North Dakota data can be written.

log(RERV) = 0.67(MRERV)
+ 0.47(MAREA25) - O.l(MEH) - 0.27. (4)

We recognize the problem of identifying a storm's
maximum stage of development in real time, and do
not suggest this idea as the answer to short-term
rainfall prediction. Such studies may result in im- ,
proving objective schemes for estimating storm rain

'..,

TABLE 3. RERV versus area-time correlations for different time increments using Bowman 1981 NDCMP and GATE data.

.'

NDCMP radar Quadra" radar
Time increments (h) LTb 1/6" 2 4 2 4 10
Number of cluster
sequences 507 2609 472 131 18 328 146 59 15
Correlation coefficients 0.98 0.94 0.96 0.97 0.94 0.91 0.91 0.91 0.94
Standard error of
estimates .0.14 0.20 0.15 0.16 0.18

" Canadian Coast Guard ship, part of QA TE (Lovejoy and Austin, 1979).
b Lifetime of cluster. '
e Scan-by-scan.