to estimate the magnitude of the decrease. The location of the gravel re- 

 moval area may provide an explanation for the significant intergravel flow 

 at Washington Creek and Aufeis Creek. At these two sites the scraping occur- 

 red near the downstream end of a sharp meander bend (Figure 39). It appeared 

 that the scraping in this location caused most of the flow to leave the 

 confinement of the channel. The lack of a well defined channel caused the 

 flow to spread over the gravels in the material site and deposit the sedi- 

 ment load that it was carrying. These deposits were quite loose and un- 

 stable, and thus were very conducive to intergravel flow. Other sites having 

 a similar specific location of scraping were slightly different in configur- 

 ation from that shown in Figure 39; either the bend upstream from the 

 scraped area at these sites was not as sharp or the scraping occurred fur- 

 ther downstream on the bend, thus allowing some of the flow and likely 

 much of the bed load to be retained in the original channel. 



Three possible explanations for the continued loss of surface flow 

 at Washington Creek are (I) that the suspended load is not sufficient to 

 fill the openings in the gravel, (2) the presence of aufeis in the site 

 protects the gravels from the significant snowmelt floods, and 13) water 

 freezes in the gravel, expanding and separating the gravels in the process. 



Pit sites, such as Dietrich Ri ver-Upstream and the two Tanana River 

 sites, had a potential to locally increase the mean annual flow as a result 

 of intercepting intergravel flow and allowing it to surface at the pit. 

 However, the percentage increase in the mean annual flow at these sites 

 is probably quite small. 



132 



