irt from 
tructure 
that the 
owever, 
'SSary. 
ations 
uity of 
precipi- 
e calcu- 
voiding 
he time 
| vario- 
Chubu, 
stations 
ich are 
ram at 
ong all 
use the 
oram at 
ost flat: 
nly this 
ype. In 
largest 
creases 
S Is an 
type. A 
though 
assified 
ram at 
show a 
' be not 
than 1 
1 cases 
stations 
Kanto. 
iven as 
(I) 
lue and 
vherical 
Is show 
similar 
> This 
ime is 
district 
T. Shoji, H. Kitaura / Computers & Geosciences 32 (2006) 1007-1024 1017 
03 p 
  
e 
N 
T 
  
Semivariogram / (log (mm x h™1)}2 
© 
T 
  
  
  
  
CO C 
= coacto on. DD ODD os 
/ d "Oum oo VM NO” DD 
dé AMAA AMA b rl A 
RAR ah BR a A 
0 | 1 L 1 L J 
0 24 48 72 
Time Lag/h 
Fig. 10. Temporal variograms of hourly precipitations observed though a year (99) at selected stations in Chubu and Kanto, and their 
fitted spherical models. Locations: Nanbu — solid circle, Hiraiwa — open circle, Tsunan — solid triangle, Hachiman — open triangle, and 
Tsubakuro-dake — plus. 
and the plain Kanto district. The variograms show 
clear nuggets, sills, and ranges. The range values are 
about 8h. Most variograms at individual stations 
show also clear ranges from 6 to 12h. These mean 
that rain intensities have generally no correlation, 
when a time lag is beyond 6-12 h. 
Temporal variograms of speeds, directions and 
velocities of wind suggest daily duration, and wind 
is strong by day and weak by night (Shoji, 2002b, 
2006). In contrast, only one example of temporal 
variograms of precipitations shows a periodical 
oscillation as mentioned above. 
Fig. 12 shows proportion (r,) of the rainy stations 
at Time # [from 01 JST (Japan Standard Time) to 24 
JST] on Day d [from January 1 to December 31], 
which is defined as follows: 
r= 0-3 [3 Wa, Q) 
d d 
where Rj, and W;, represent the numbers of rainy 
stations and worked stations in each district, 
respectively. Fig. 12 indicates two facts: (1) the 
proportions of rainy stations are 8-10% in Chubu, 
and 5-7% in Kanto, and it is concluded that Chubu 
has more rainy hours than Kanto, and (2) the 
number of rainy stations is the largest around 17 
o’clock independently of the districts. The first fact 
is the reason why the distribution curve of annual 
precipitations in Chubu passes through the much 
precipitation region compared with that in Kanto 
(Fig. 5). 
Fig. 13 shows hourly precipitation p, at Time ¢ 
averaged through a year, which is defined by the 
equation: 
A- XY Yn 
d i d 
E Ys Ys. (3) 
d d 
where pj, is the precipitation at Time f£ on Day dat 
Station i, and P,, is precipitation summarized in a