1000

Dec 1

J ec 15

800

Jan 1

Fan 15

Feb 1

600

eb 15

400

200

0

2

5

10 20

100

Return Interval (years)

14 10 3

12

10

Envelope

8

6

4

2

0

100

200

300

400

500

600

700

800

900

1000

0

AFDD

deficits occurring in each AFDD category are shown

In Figure 17 the maximum discharge deficits list-

in Figure 18. As with the previous statistical analysis,

ed in Table 5 are plotted against the AFDD that

the return interval for each AFDD category is found

occurred on the day that the maximum discharge

by assigning each maximum discharge deficit a

deficit occurred. It can be seen that in general there is

an "envelope" that describes the limit of the dis-

charge deficit and that the upper limit of this enve-

of the deficit and *N *is the total number of data points

lope decreases with increasing AFDD. This decrease

available. The return interval is then 1/(1*P*). We can

reflects the fact that as each winter progresses (and

see that, in large part, the data are normally distrib-

uted except for the category 100200 AFDD. This

the number of AFDD increases), the amount of ice in

category is somewhat comparable to the time period

the river is increased and the amount of open water

1631 December. As with that time period, the data

decreased. With the reach fully ice covered, dis-

for this category suggest a mixed population, and

charge deficits should not occur.

separate probability curves have been drawn to indi-

To complete this analysis, we can group the dis-

cate the two populations. At this time it is not pos-

charge deficits into AFDD categories as shown in

sible to explain the mixed population appearance of

Table 7. It can be seen that not every AFDD category

this data.

is reached every year and that ice-impacted periods

do not occur every year during each AFDD category

To further extend the statistical analyses, we

that is reached. The annual maximum discharge

must take note that discharge deficit cannot be ex-

15