and standard deviation at the end of the month

gible but evaluate it for the pairs of months of

are given in Table 2. December 1977 and January

largest flow increase and decrease in the period

1980 were the months of maximum ice growth,

of record. Qsub in eq 8 can be evaluated consider-

and February 1976 was the month of maximum

ing each component of the water balance as a

melt in this period of record.

random variable. The measurement error for dis-

Figures 4 and 5 give results for January 1979,

charge at the gages in winter is typically about

the coldest month of the study. The mean, stan-

8%, as reported on the USGS discharge measure-

dard deviation, and corresponding beta distribu-

ment notes. We will use this value as the coeffi-

tion for Qice and Qsub of each main-stem White

cient of variation for the measured monthly aver-

age discharges at the gages. Qin and Qout correla-

River reach are presented in Figure 4. The storage

tions were computed for each pair of gages over

of water as ice is an important term in the water

the period of study and have coefficients that in-

balance of the White River below Oglala, but it is

crease with distance downstream.

of less significance farther upstream and on the

Little White River, where the stream widths are

small and the flows are relatively high. The

CrawfordOglala subbasin had a negative yield

RESULTS

to the river. The distributions for Qice and Qsub

Seven consecutive winters from November 1974

have similar shapes in the two reaches below

to February 1981 provide a representative range