natural variation of the median diameter at a site. Degradation was also ob- 

 served at a few sites although at other sites only causative evidence was 

 available to indicate that this process can occur. Sediment transport 

 changes were suggested at several sites where there were observations of 

 bedforms in or downstream from the gravel removal area, observations of 

 changes in the bed material size, computations of changes in shear stress, 

 or observations of sediment sources which remained from the gravel removal 

 operation. The effects of gravel removal activity on these sedimentation 

 characteristics were evaluated and given quantitative ratings for com- 

 parative purposes (Table 8). 



Sediment Size Distribution. The most common significant change in 

 sediment size distribution resulting from gravel removal was a decrease 

 in the size caused by fine material deposition in the material site. This 

 change was reflected in the surface material at six sites and the subsurface 

 material at six sites. Oregon Creek, Penny River, and Ugnuravik River had 

 significant changes in both surface and subsurface material sizes. At Sinuk 

 River, fine and medium sized gravels were nearly missing from the subsurface 

 samples in the material site, causing an increase in the median size. The 

 explanation for this is unknown. At Washington Creek, the subsurface ma- 

 terial size was larger in the material site even though fine material depo- 

 sition in the site reduced the median size of the armor layer. 



A pattern of correlation was not evident between increases or decreases 

 in armor layer median diameter resulting from gravel removal and physical 

 site or gravel removal area characteristics. One reason for this lack of 

 correlation is that armor layer development is a complex function of several 

 interrelated factors including degree of development of undisturbed armor 

 layer, flooding history since gravel was removed, and flow characteristics 

 in the gravel removal area. If the undisturbed size distribution of the 

 armor layer was not significantly different from that of the material under- 

 lying it, the relative change due to gravel removal would have been less and 

 the time required for recovery to the undisturbed condition would also be 

 less. The time for recovery is also a function of the floods during the 

 recovery period; one large recurrence interval flood may be sufficient to 



118 



