294 
Table 1. Number of available and suitable images, compared with theoretically possible number of satellite 
overpasses from April through September, for two irrigation districts, Grande Bonifica Ferrarese and East 
Sesia, from 1980 through 1985 
Year 
Path/row 
Sensor 
Over 
passes 
Available 
images 
Images cloud 
free in 4 
quadrants 
Images cloud 
free in the 
relevant 
quadrants 
Suitable images 
in relevant 
quadrants from 
April to Sept. 
Number of 
required 
images for the 
investigation 
num 
ber 
% of 
total 
num 
ber 
% of 
total 
num 
ber 
% of 
total 
num 
ber 
% of 
total 
Grande 
Bonifica 
Ferrarese 
1980 
206/29 
MSS 
41 
22 
53.7 
3 
7.3 
4 
9.8 
4 
9.8 
3 
207/29 
MSS 
41 
24 
58.5 
5 
12.2 
7 
17.1 
5 
12.2 
3 
1981 
206/29 
MSS 
40 
7 
17.5 
not available 
7 
17.5 
6 
15.0 
3 
207/29 
MSS 
40 
5 
12.5 
not available 
5 
12.5 
5 
12.5 
3 
1982 
206/29 
MSS 
33 
12 
36.4 
not available 
6 
18.2 
4 
12.1 
3 
207/29 
MSS 
33 
12 
36.4 
not available 
4 
12.1 
3 
9.1 
3 
1983 
192/29 
MSS 
26 
20 
76.9 
4 
15.4 
6 
23.1 
1 
3.9 
3 
1984 
192/29 
MSS 
42 
24 
57.1 
2 
4.8 
4 
9.5 
3 
7.1 
3 
TM 
19 
15 
78.9 
1 
5.3 
2 
10.5 
2 
10.5 
3 
1985 
192/29 
MSS 
46 
37 
80.4 
6 
13 
14 
30.4 
10 
21.7 
3 
TM 
23 
19 
82.6 
2 
8.7 
5 
21.7 
4 
17.4 
3 
East Sesia 
1980 
209/28 
MSS 
41 
24 
58.5 
2 
4.9 
4 
9.8 
3 
7.3 
3 
1981 
209/28 
MSS 
40 
13 
32.5 
not available 
6 
15.0 
5 
12.5 
3 
1982 
209/28 
MSS 
33 
13 
39.4 
not available 
4 
12.1 
3 
9.1 
3 
1983 
194/28 
MSS 
26 
18 
69.2 
3 
11.5 
4 
15.4 
1 
3.8 
3 
1984 
194/28 
MSS 
42 
38 
90.5 
2 
4.8 
5 
11 .9 
4 
9.5 
3 
TM 
19 
17 
89.5 
1 
5.3 
3 
21.1 
2 
10.5 
3 
1985 
194/28 
MSS 
46 
42 
91.3 
7 
15.2 
13 
28.3 
6 
13.0 
3 
TM 
23 
21 
91.3 
5 
21.7 
7 
30.4 
3 
13.0 
3 
Figure 1. Attainment of 100% ground cover in the irrigation district Grande Bonifica Ferrarese in Italy, dur 
ing the growing season of 1985 ; A. cumulative distribution of area of occurrence for sugar beet; [t^, t^ ] such 
as to give [ACt^) - A(t^)] = 0.8 A(t^); B. durations [t^, t^] and [t^, 12] for each crop 
overlap between the periods [t-|, 12 ] applying to each 
crop will be rather large, thus complicating crop 
discrimination on the basis of multitemporal reflec 
tance measurements. We can make discrimination eas 
ier by accepting a higher risk of misinterpretation. 
So, on the basis of a graph of A(t) (Fig. 1A) we can 
estimate a time period [t., ¿2] < [t^, 12] such as to 
include 80%, say, of A(t2l. The left and right 10%- 
tails are left out of consideration, thus accepting 
the implied risk of misinterpretation. We can final 
ly obtain a crop calendar (Fig. 1B) giving for each 
crop and stage both the duration [tj, t2] and [t^, 
£2]- Note that these intervals of time are not the 
actual duration at an individual field of the 
phenological stage 'attainment of full ground cover' 
They indicate that period of time within which a