Under south-slope conditions, as shown in figure 6, the positive slope coefficient 

 (>1.00) denotes greater differences between forest and recovery plots in winters with 

 heavy snowfall. 



Was there a trend on the south-slope recovery plot? The measurement taken in 1974 

 lies within the 90 percent confidence band that was calculated based upon the variance 

 about the individual value. Apparent lack of trend, if indeed this was the case, was 

 not unexpected since the reestablishment of the new stand was progressing slowly. There 

 were less than 100 trees per acre (<40 trees/ha) on the recovery plot. 



SPRING SNOWMELT RATE 



North Slope 



Spring snowmelt rates were determined for the recovery plot by dividing the accumu- 

 lated snowmelt--up to the time snow disappeared on the forest plot--by the number of 

 days elapsed. For the forest plot, total snowmelt equaled the value shown for the peak 

 W.E. Average rates by year are listed in table 2. 



The regression curves (fig. 7) show much faster snowmelt on the recovery plot 

 from 1942 to 1952 than from 1967 to 1974. The suggested downward trend was supported 

 by results of the time-trend analyses (table 31 . We expected the spring temperature 

 coefficient to be positive and highly correlated with snowmelt. Again, the time-trend 

 variable under forest conditions was unimportant. In the 34-year period, changes in 

 melt rate were apparently associated only with deviations in spring temperature. 



For the recovery plot, the analysis revealed a highly significant negative time- 

 trend coefficient (P<=0. 01). The occurrence of slight gradual decreases in snowmelt 

 rate in recent years was entirely independent of temperature rise and fall. The thinned 

 sapling stand- -growing taller and enlarging its crown area each year--was apparently 

 acting as a barrier to wind movement and was casting an ever-expanding shadow to shade 

 the snowpack, thus effecting a gradual reduction in average melt rate. 



Of special interest in figure 7, the 1967 to 1974 regression curve is plotted 

 slightly beneath the 100-percent recovery line. After 34 years, the managed, sapling 

 stand had recovered with melt rates equal to or slightly reduced from those associated 

 with the residual mature timber. Crown-cover density in 1974 was estimated to be 0.22, 

 compared with the mature forest figure of 0.80. Similar relations between crown-cover 

 density and snowmelt on the north slope were reported by Packer (1971) . The rate of 

 snowmelt at his site decreased as crown-cover density increased from zero to 0.25, but 

 then increased from 0.25 to 1.00. Experimentation conducted in the central Sierra 

 Nevada with radiation instruments supports Packer's empirical results (U.S. Army 

 Engineers 1956). 



South Slope 



Snowmelt rates were consistently greater in the recovery plot for the six measure- 

 ment periods (table 2). The lack of sampling data precludes clear evidence regarding 

 trend. Because reductions in peak W.E. did not appear, we might not expect measurable 

 change in melt rate. The lack of trend (not statistically tested) appears borne out by 

 the 1974 data point, plotted within the 90 percent confidence band (fig. 7). 



12 



