172 



Fishery Bulletin 91(1). 1993 



Figure 1 



Study area and track maps of hori- 

 zontal movements by chinook 

 salmon Oncorhyrwhus tshawytscha 

 tracked in the Columbia River es- 

 tuary. Sampling during flooding 

 ( ) and ebbing (•) tides. Each 

 circle represents 30min of track- 

 ing time. 'H' indicates extended 

 holding period occurred. 



relatively shallow water (about 5 m) on the south side 

 of Sand Island (except Fish 1 which was captured on 

 the north side of Desdemona Sands), and all fish were 

 released at Buoy 21 (Fig. 1 ). 



A single fish was released each day and followed 

 primarily during daylight hours from the gillnet ves- 

 sel Midnight Gambler. Transmitted signals were re- 

 ceived by a directional hydrophone and tunable re- 

 ceiver/decoder (Vemco Ltd.). During tracking, the boat 

 typically stayed 50-400 m away from the fish, and the 

 following data were collected: (1) boat position every 

 5 min from a loran C receiver; (2) water depth beneath 

 the boat every 5 min from a fathometer; (3) fish depth 

 every lmin from the decoder; (4) approximately every 

 30 min the fish was more closely approached (usually 

 to within 50 m, based on triangulation and signal 

 strength), and secchi disk and CTD casts were made 

 while the boat drifted. CTD casts took about 5 min to 

 perform and measured the conductivity and tempera- 

 ture at intervals of 1 or 2 m, usually to within 4 m of 

 the bottom. In deeper waters, casts were generally 

 limited to 12 m to avoid losing the fish. Except for 

 fish swimming close to the bottom, this range always 

 encompassed the depth at which the fish was 

 swimming and any large changes in temperature or 

 salinity. 



Data analysis 



Boat positions were used to reconstruct each fish's path 

 on a horizontal track map and to determine ground 

 speed. A 15 min sampling interval was chosen to cal- 

 culate ground speeds because shorter intervals may 

 overestimate fish speed due to extraneous boat move- 

 ments, and longer intervals may underestimate fish 

 speed because calculations based on a straight line 



between positions may mask shorter-scale movements. 

 Water and fish depths were used to reconstruct each 

 fish's path on a vertical track map. Conductivity was 

 converted to salinity (Perkin & Walker 1972) for con- 

 struction of temperature and salinity profiles. 



To determine whether salmon showed preferences 

 for ranges of temperature or salinity, the salinity and 

 temperature of the water experienced by each fish were 

 determined indirectly by substituting the appropriate 

 values from the temperature and salinity profile for 

 the depth at which the fish was swimming during each 

 observation. Salinities and temperatures between the 

 measured depth-intervals were determined by linear 

 interpolation. The range of temperatures and salini- 

 ties available to each fish was determined from tem- 

 perature and salinity profiles separated into 1-unit (°C 

 or %o) intervals. The fraction of the water column that 

 each unit of temperature or salinity occupied within 

 the sampled depth was calculated and multiplied by 

 the time-interval of the representative temperature and 

 salinity profile. Each temperature and salinity profile 

 was assumed to represent water conditions over a time- 

 interval midway between consecutive profiles. Fish that 

 swam near the bottom sometimes exceeded the depth 

 of the CTD casts, and these observations were omitted 

 from analysis of salinity or temperature preference. 

 Frequencies of temperature and salinity were summed 

 over all profiles for each track to obtain the salinity 

 and temperature distribution available to each fish. 

 These distributions were tested statistically by good- 

 ness-of-fit analysis to determine if the distributions of 

 available and experienced conditions were similar. Dif- 

 ferences were assumed to indicate fish were display- 

 ing non-random vertical movements, presumably to se- 

 lect for a favorable combination of environmental 

 factors. 



