na near the 
Nds after 
' four main 
s of the 
5, two days 
istal area 
a, counter 
| a small bay 
e. 
ta 
MSS band 4 
tal line is 
adiant area 
adiant pat- 
river water 
| 9115. 
at MSS image- 
a along the 
Indicated in 
direction 
‘ig.l) sug- 
are describ- 
| salinity 
'Okachi River 
50-to 120 
lume of the 
pwater:'a- 
= August (See 
ater is con- 
of August 25 
"Of around 
tember 30 
line.  There- 
in accordance 
re an ocean 
ront. 
+ off both 
5). On the 
e that the 
lain is from 
current is 
    
  
thought. to exist as is shown by a thick arrow in Fig.5, and it is 
observed in MSS band 4 imagery on August 25 that the current from 
southwest along the coast mixes with that from northeast near 
Kombumori (See Photo 1). 
II.3 Bay Current 
  
Clockwise rotation of muddy water in Akkeshi Bay is recognized 
in MSS imageries on both August 25 and September 30. Especially, 
extremely muddy water rotation is observed in the former imagery. 
Multi-temporal comparison of Landsat MSS data shows the fea- 
ture of water flows change according to tide or season(Photo 3). 
In winter, water in the bay is clearer, and the flowing pattern of 
ice there is obviously different from that of muddy water in summer. 
Bay current is considered to be affected by the ocean current off 
Akkeshi Bay. 
II.4 Clear Water Flowing into Lake 
According to MSS band 4 and 5 imageries on both August 25 and 
September 30, relatively clear water flowing into Lake Akkeshi (In- 
dicated in Fig.1), while the lake is not clear owing to high plank- 
ton density (See Photo 3 (a) and (b)). Water of the river is not 
muddy even immediately after the typhoon though that of other rivers 
contain plenty of mud. 
III. Detection of Flooded Area 
  
An attempt was made to detect the flooded area in the Tokachi 
Plain from Landsat imageries taken two days after Typhoon 8115 which 
passed over Hokkaido on August 23 and caused flood in the area. 
Unfortunately the area in the MSS imageries taken before the 
flood were entirely covered with cloud, so the scenes which were 
applicable to the analysis were only those of August 25 and Septem- 
ber 30, 1981 after the flood was over. 
It was not easy to detect the flooded area due to the follow- 
ing reasons; 
a) The flood caused by Typhoon 8115 was not so great. 
b) The surface condition changed between August 25 and Septem- 
ber 30, which makes multi-temporal analysis difficult. For 
example, at the beginning of September a spring wheat farm, 
where wheat was reapt at the end of July, was digged, and 
potatoes were cropped for crop rotation almost at the begin- 
ning of September. 
C) A part of the area concerned was covered with thin cloud 
which makes digital analysis difficult. 
To clear the difficulties above, color areal photos (Photo 6 
and 7) were greatly effective. The photos were taken not after 
Typhoon 8115 but on around August 7, four days after previously 
passed Typhoon 8112 which caused great and widespread flood in 
Hokkaido. There is a good reason to assume that flooded area by 
heavy rainfall of the previous typhoon may be flooded again by the 
rainfall of the present typhoon, Typhoon 8115. Thus it is reason- 
able to make use of the areal photos taken on around August 7, 1981. 
In fact, correspondence of traces of submergence between areal 
photos and Landsat MSS false color images (band 4, 5 and 7) was rec- 
ognized (See Photo 5 and 7). 
The analized areas are shown in Fig.l and Photo 4 and 5. 
Using Landsat digital image data analysis system, the spectral 
characteristics (See Fig.6) of the sites interpreted as flooded area 
(See Photo 5) in Landsat image on August 25 were computed and the 
flooded area was classified.