WMOD00419

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

1608

MONTHLY WEATHER REVIEW

VOLUME 112

that was the only way to check the rain volume
estimates with an independent set of data. The mean
value of the residual differences between the two rain
volume estimates was 0.965, indicating that the 1980
formula underestimated 1981 cluster rain volume by
an average of 3.5%. This might be attributed to the
much drier weather conditions of 1980. A similar
test was made using the 1972 North Dakota Pilot
Project (NDPP) rain volume-A TI relationship (Do-
neaud et al., 1981). The reason for the double checking
was the fact that the 1972 data were not "floating
target" radar echo clusters, but fixed radar screen
reflectivity data. Moreover, the data were obtained
with a 10 cm radar system, no reflectivity threshold
was employed, and the A TI were calculated on an
hourly (instead of scan-by-scan) basis. In view of
these differences and the resulting larger area-time
integrals, the agreement was remarkably good (the
mean value of the residuals was 0.902, underestimat-
ing cluster rain volumes by - 10%). The scatter and
the linear fit of both the 1980 NDCMP and the 1972
NDPP data are shown in Fig. 6 in a log-log plot.
The correlation coefficients are 0.98 and 0.97 for the
1980 and the 1972 data, respectively. The 95% con-
fidence limits of the correlation coefficient for the

106

5

2

~ 105
E
E 5
)(
N
E
= 2

I
W
~ 104
...J
o
> !!
z
<[
0: 2
o
W
~ 1()3
::!:
I- 5
lJ)
W

~ 2
o
C2 102

SLOPE =1.0 INTERC = 3.63

.
.
.

5

2

10'
100

1980 population are 0.96-0.99. Slight differences
exist in the slope and intercept, but good alignment
of the data for both samples emphasizes the potential
predictive power of the A TI technique for different
storm sizes.

6. The evolution of R during storms

As previously described, the ATI-RERV relation-
ship, implying a slightly increasing average rain rate
with A TI (or, in a first approximation, an average
rain rate independent of A TI), is useful for estimating
the amount of rain from a storm following its dissi-
pation. It would be ~f interest to know whether this
or a similar relationship could be used at intermediate
stages of a storm's lifetime to provide estimates of
the rain volume in real time. This information could
also be used to evaluate the variation in rain produc-
tion with time in a storm's entire history. If the rain
rate averaged over shorter periods of time (e.g., 10
min, I h) or at some intermediate stage of the cluster
lifetime (e.g., during the growing period) could be
considered independent of the A TI, a potential fore-
cast tool might emerge.
Each cluster lifetime was divided into several time
intervals as follows: 1) the time between volume

j

INTERC = 2.32 (1972)

i'

101 102. 103 104
MAXIMUM AREA-TIME INTEGRAL -\km2 ,x,hr1.

FIG. 6. Scatter and linear fits for the 1980 NDCMPaDcl tlie '1972 NDPP data sets.

105