oxygen levels throughout the winter. The lower Tanana R i ver-Upstream pit did 

 not contain a great average depth, 1.7 m, but the lack of littoral vege- 

 tation reduced the probability of oxygen depletion. The water in the pit was 

 turbid during the summer, limiting production of aquatic vegetation. The 

 lower pit maintained higher dissolved oxygen than the upper pit during the 

 winter (Table 24). The Dietrich R i ver-Upstream and Tanana River-Downstream 

 pits both contained deep, clear water regions and did not have well-devel- 

 oped littoral vegetation. Oxygen levels probably remained high through- 

 out the year. The depth and lack of productivity combined to make these two 

 pits excellent overwintering areas; the same features limited their value as 

 rearing areas. 



There are other possible effects of grave! removal on overwintering 

 areas, but they are difficult to assess because of the absence of data on 

 the study sites before gravel removal. A primary effect is the loss of 

 overwintering areas due to diversion of flow from an original channel, as 

 occurred at four sites (Penny River, Dietrich River-Downstream, Middle Fork 

 Koyukuk R i ver-Upstream, and Middle Fork Koyukuk River-Downstream). In these 

 cases, complete or partial diversion of flow could lead to loss or reduction 

 of overwintering habitat. Another effect is the loss of overwintering hab- 

 itat due to increased braiding and the associated changes — loss of pool- 

 riffle sequence and reductions in depth and velocity which promote rapid 

 freezing. In some areas, gravel removal created or aggravated the formation 

 of aufeis fields, thus leading to a reduction in water available for over- 

 wintering downstream (Washington Creek, Oregon Creek, McManus Creek, pos- 

 sibly some of the North Slope sites). 



EFFECTS OF HABITAT ALTERATION ON AQUATIC MACRO I NVERTEBRATES 



Observed Effects on Density and Species Assemblage 



Habitat alterations expected to affect assemblages of riffle macro- 

 invertebrates would be changes in velocity, substrate, depth, and water 

 quality. During the present study, habitat alterations resulting in a change 

 of each of these parameters were accompanied by changes in the riffle com- 

 mun i ty (Tab I e 25 ) . 



198 



