This study examines the effects of a low oxygen environment, in concert with fluctuating temperature and salinity conditions, on the nearshore depth distributions (0-1 m) and flood tide movements of juvenile chinook (Oncorhynchus tshawytscha) and chum (O. keta) salmon. Comparisons are made between an unpolluted and a sewage polluted estuarine intertidal flat in the Fraser River estuary, British Columbia, the polluted area being characterized by the regular occurrence of low dissolved oxygen levels. Results are based on 380 beach seine samples taken between April and June of 1984. In general, chum and chinook salmon of increasing length were captured in increasing depths, though this pattern was modified by seasonal changes in water temperature. Low dissolved oxygen conditions in deeper waters may have been responsible for the presence of larger, and often sluggishly swimming fish in higher oxygenated surface water layers or in shallow waters near the shore. In both areas, the risk of aerial predation was high. On a flood tide, the likelihood of capturing a chinook salmon was reduced as temperatures increased and oxygen levels decreased. A combination of avoidance behaviour and a regularity in the movement patterns of chinook onto the study area in the later stages of the flood tide may account for their rare occurrence in low oxygen concentrations (<6 mg/1) and high temperatures (>20 °C). Fish mortalities were most likely to occur on the ebb tide when fish were forced into waters of low oxygen content by the drainage patterns characteristic of the polluted study area. Though wide ranges in salinity were recorded on both tidal flats, this factor was not strongly correlated to Chinook distributions. However, significantly higher salinity levels in the unpolluted area may account for the greater numbers of chum salmon captured there. An understanding of the influence of estuarine water quality conditions on the distribution of juvenile salmonids may assist in the identification of significant sources of mortality in their early marine life. This knowledge is particularly important in the evaluation of water quality changes as caused by human activity. / Science, Faculty of / Zoology, Department of / Graduate

Visual estimation techniques were used to quantify habitat characteristics, habitat type (pool, riffle) use and longitudinal distribution of steelhead (Oncorhynchus mykiss), cutthroat trout (0. clarki), and coho salmon (0. kisutch) in spring, summer and fall in 8.8 km of Cummins Creek, a basin in the central coast of Oregon. Fish were distributed significantly different than habitat type availability in most samples. Pool habitats contained a disproportionate percent of the salmonid assemblage and 1+ fish in each sample, and the percentage of fish in pools increased as flow decreased. In spring, coho salmon fry were concentrated in side channels and valley floor tributary habitats. Large woody debris formed 57- 68% of pool habitats and was significantly correlated with pool volume, maximum pool depth, slow surface velocity in pools, and pieces of small woody debris. Longitudinal distribution of the salmonid assemblage did not differ from habitat distribution seasonally or between years, even though certain species differed Coho salmon and cutthroat trout were distributed in proportion to longitudinal habitat availability only when fish abundance was relatively high and streamflow was low. In most samples, both 0+ and 1+ steelhead were distributed in proportion to longitudinal habitat availability. Differences in coho salmon abundance between years appeared to influence longitudinal distribution of each species and age class. Certain reaches had consistent numbers of fish between years while the number of fish in other reaches varied widely. In most samples, reaches with highest abundance for steelhead were in the lower basin, cutthroat trout in the upper basin and coho salmon between the two other species. Timing of reduction in number of fish varied among species. Fifty-five percent of 0+ steelhead and 73% of 1+ steelhead lost between August 1988 and April 1989 were lost between August and October during low flow conditions. However, only 18% of the losses, for 0+ coho salmon, occurred between August and October with the remaining losses occurring after October. This study illustrates that habitat availability is not a good index of fish distribution when fish abundance is low, and it highlights the importance of habitat in the lower portions of basins when fish abundance is high. It also demonstrates that the basin wide distribution of salmonids varies among species, age classes, seasons, and years and suggests that our understanding of salmonid distribution and abundance could be greatly enhanced by adopting a basin-wide, community, and seasonal perspective. In addition, the methods used in this study offer one way to assess the seasonal distribution and abundance of salmonids in a relatively quick, inexpensive, and non-destructive manner. / Graduation date: 1994