Hawaiian Rainfall Estimates — Stidd 
269 
Figure 4^ shows a correlation field which 
associates 5 -day mean 700-millibar heights 
with 5-day rainfall totals at Hilo. It is different 
from the monthly chart of Figure 1 in three 
noteworthy respects: 
1. The maximum correlation coefficient is 
much lower. 
2. A negative correlation is found in the 
Hawaiian area. 
3. The positive maximum is displaced well 
to the east of its monthly position. 
This chart points up the need for considering 
the short-period phenomena which go to 
make up the mean monthly pressure maps 
and indicates that the reasoning, previously 
presented to explain the monthly pattern, is 
not completely adequate. Further evidence for 
this is provided by Figure Ab, which shows 
the correlation field relating 24-hour rainfall 
amounts at Hilo to mid-period 700-millibar 
heights. This chart substantiates Figure Aa in 
showing a negative correlation near Hawaii, 
but the positive maximum is displaced back 
to the approximate position which it occupies 
on the monthly chart. In considering these 
two shorter-period charts, it should be borne 
in mind that they include only 5 or 7 years of 
data and that, in addition, the 24-hour- 
amounts chart is only for the months of 
January and February, so that the apparent 
differences may be due in part to the pecu- 
liarities of the particular time periods which 
the charts represent. The author suspects that 
this consideration might account for the zonal 
shift in the positive center but is inadequate 
to explain the negative correlation in the 
island area. 
The smaller correlation coefficients dis- 
played by the shorter periods are due in part 
to the larger degree of randomness or "noise” 
in the basic data. This is largely suppressed 
in the process of taking 30-day means. In the 
previous study (Stidd, 1954), short-period 
data for the Tennessee Valley gave patterns 
similar to the monthly data, although the 
strengths of the relationships were reduced 
in about the same degree as that shown here. 
Fig. 4. Correlation field patterns of 700-millibar 
height versus Hilo area rainfall, a. Five-day means, 
November-April, 255 cases; b, twenty-four-hour rain- 
fall amounts and mid-period synoptic 700-millibar 
heights, data taken every fifth day from the months of 
January and February, 66 cases. 
The Hawaiian short-period patterns, however, 
are distinctly "different” from the monthly 
patterns. 
It is evident that patterns associated with 
daily rainfall amounts need not be the same 
as those for monthly amounts. Suppose that 
a strong positive relationship exists between 
pressure at a given point and concurrent rain- 
fall at some other point. During periods when 
no rain is falling, the pressure at the given 
point will be low, and in a month’s time it 
might happen that nearly all the days are free 
of rainfall and marked by low pressure at the 
