interval events the user would follow

the 20% chance line to come up with

minimum recommended releases.

Similar information for the primary

ice event months of December

Open water

7,100

7,100

7,400

7,400

6,900

6,900

through February is also provided in

Table 12.

50%

7,200

7,500

9,800

8,200

7,100

7,000

This approach could also be

20%

7,400

9,000

12,800

10,400

8,800

7,900

modified to address releases during

a specific year. It would still be based

10%

8,200

11,300

14,300

11,400

9,700

8,900

on long-term weather statistics, but

5%

10,500

13,500

15,600

12,400

10,300

9,600

could allow consideration of years

that are relatively wetter or drier than

normal. If information was available

on expected tributary inflow rates, the monthly mini-

further assuming that it is 3 January, we could esti-

mum release flow distributions, such as those pre-

mate that there is a 10% risk of discharge deficit of

6,900 ft3/s or greater occurring. Since the discharge

sented for long-term averages in Table 11, could be

in reach 2 would be 10,600 ft3/s, and the minimum

calculated for specific time periods. Combining this

required reach flow is 8,000 ft3/s, the discharge from

contained in Figure 15 would provide recom-

Gavins Point would have to be increased by 4,300

ft3/s to avoid difficulty in that reach. Further, since

mended release information similar to that in Figure

24, but tailored more to the inflow conditions ex-

the water travel time from Gavins Point to reach 7 for

pected for a given year.

January 1970 flow rates is on the order of 8 days and

the wave travel time about 6.5 days, the release

would have to begin before the normal 3- to 5-day

This approach would be applicable for near-term

weather forecast period would provide reliable fore-

modification of the planned winter release schedule

casts. Should a cold weather system be anticipated,

in anticipation of an approaching cold weather pe-

the response would need to be based on known

riod. As in the long-term approach just discussed, the

water discharge distributions and a risk-based esti-

estimation of required releases from Gavins Point is

mate of cold snap severity.

again based on an acceptable level of risk. Since the

Alternately, if the accumulated freezing-degree-

travel time of releases from Gavins Point Dam to

days (AFDD) have been tracked as described earlier,

a discharge deficit estimate can be made that ac-

many of the water intakes is significantly longer than

counts for the severity of the current winter. In this

the time period of reliable weather forecasts, winter

case, the user would check the current AFDD tabu-

releases must be based on probabilistic, conservative

lated since 1 December, enter Figure 19 within that

estimates of required flows.

AFDD range and read the discharge deficit magni-

To use this approach it is first necessary to esti-

tude corresponding to the preselected level of risk. If

mate the potential severity of the event. The probable

a cold snap is anticipated that would increase the

level of discharge deficit can be determined from

AFDD to the next range in Figure 19, then values in

either Figure 15 or Figure 19. The most straight-

each range should also be considered in determining

forward approach is to use Figure 15, which provides

an appropriate response.

the probable level of discharge deficit corresponding

For example, if the January 1970 discharge sce-

to a selected level of risk based on calendar date.

nario is assumed along with 180 AFDD, a 10% level

Thus, a 10% chance of exceedence on, say, 23 Decem-

of risk would correspond to a discharge deficit of

ber would give an estimated discharge deficit of

3/s, whereas on 23 February it would be only

6,800 ft3/s. However, if it is 3 January we could

4,200 ft

2,000 ft3/s. This estimated deficit could then be

estimate (using Fig. 8a) that there is a 50% risk of a 5-

coupled with current known (or projected) levels

day cold snap exceeding 80 FDD and a 10-day cold

and flows to determine whether such a deficit might

snap exceeding 140 FDD. In that case we would

create a problem and if so, how much the Gavins

expect that in one out of two years we could go from

Point discharge should be increased.

180 to 260 AFDD within the next 5 days and to as

For example, using the long-term monthly mean

much as 320 AFFD within 10 days. Thus, the ex-

incremental discharge distributions in Table 9, an

pected discharge deficit might actually decrease

3/s, and

from 6,800 ft3/s to 6,400 and then 5,300 as the cold

assumed Gavins Point discharge of 10,000 ft

23