over, weirs do not provide information on the 

 distribution of spawners in the stream. 



In 1960 .and 1961 several different methods 

 of estimating the nun-iber of spawners were 

 used including periodic counts made on foot 

 surveys to determine distribution. A method 

 of estimating the total nunnber of fish that 

 spawned during the season based on periodic 

 foot survey coxints was evaluated by comparing 

 the results with estinnates made by other 

 methods. 



Description of Methods 



In 1960 and 1961, several methods were 

 used to estimate the number of spawners in 

 the fresh-water and intertidal stream areas. 

 Fresh-water spawners were counted as they 

 passed weirs at the mean high tide line of 

 Olsen Creek (Helle et al., 1964). Frequent 

 floods made maintenance of the weirs too 

 costly. In I960 a "peak count" method was 

 used to estimate the number of intertidal 

 spawners. In this method observers walking 

 along the streambank counted live and dead 

 fish. The peak count estimate was made by 

 adding the maximum count of live fish to the 

 accumulated count of carcasses on any one 

 day. In 1961 the number of intertidal spawners 

 was estimated by a tag ratio method. In this 

 method live pink salmon were tagged in the 

 bay with Petersen disk tags, and the ratio of 

 tagged to untagged carcasses on the spawn- 

 ing grounds was used to estimate the size of 

 the population. 



Because the combination of methods tried 

 was not satisfactory, we needed to develop a 

 single, more efficient method for estimating 

 the populations of both intertidal and fresh- 

 water spawners. McNeil (1962) had been suc- 

 cessful in estimating total spawning popula- 

 tions of pink salmon in southeastern Alaska 

 by periodically walking along a stream and 

 counting the fish on the spawning grounds (so- 

 called foot survey). We had counted pink salmon 

 during periodic foot surveys of Olsen Creek in 

 1960-61 to determine the time of spawning and 

 distribution of fish in fresh-water and inter- 

 tidal areas. I therefore decided to try this 

 method for estimating the populations of in- 

 tertidal and fresh-water spawners. 



Evaluation of the Foot Survey Method 



To evaluate the foot survey method, I used 

 data collected during the studies in 1960-61, 

 when foot surveys were used to estimate the 

 numbers and distribution of spawners in the 

 entire stream every 7 to 10 days. Two biolo- 

 gists counted salmon separately on each sur- 

 vey, and although the counts were similar, 

 those of the more experienced observer were 

 used in all calculations. Sheridan (1962) and 

 Willis (1964) compared spawning ground (foot 

 survey) counts of different observers made 



under ideal field conditions and found no sig- 

 nificant differences. Variability increased, 

 however, when counts of salmon in pools were 

 included and also when nunnbers of fish were 

 large (Sheridan, 1962). 



Methods of estimating the total number of 

 spawners fron-i repeated foot surveys were 

 suggested by Gangmark and Fulton (1952) and 

 Sheridan (1962), and described in detail by 

 McNeil (1962, 1964, 1966a, and 1966b). McNeil 

 (1964) plotted daily counts, drew a curve by 

 inspection, and divided the area under the 

 curve (fish days) by the average redd life (the 

 estimated number of days each fish spent on 

 the spawning ground) to obtain an estimate of 

 the number of pink salmon that spawned. J£sti- 

 mates made by this method were within 6 per- 

 cent of the actual number of fish present 

 (counted through a weir). 6 



The foot survey counts for 1960 and 1961 

 were plotted (figs. 3 and 4), and the area under 

 the curve for each year (fish days) was calcu- 

 lated and divided by the average stream life 

 (number of days each fish spends in the streann) 

 to obtain estimates of the intertidal and fresh- 

 water spawning populations. Stream life was 

 used in the calculations instead of redd life 

 because all of the pink salmon in the stream 

 were counted whether they were spawning or 

 not. In 1961 the average stream life of inter- 

 tidal spawners was shorter for the late run 

 (8.5 days) than the early run (15.3 days)--see 

 Helle et al. (1964). The estimated stream life 

 of fresh-water spawners was the difference in 

 days between the highest daily count of live 

 fish and the highest daily count of carcasses 

 that had drifted downstream and accumulated 

 at the weirs. This difference was 9 days for 

 the East Fork and 16 days for the West Fork 

 in 1960 and 9 days for the East Fork and 13 

 days for the West Fork in 1961. It is not clear 

 why stream life was longer in the West Fork. 

 Perhaps the large pool on the West Fork 

 allowed spawned out fish to remain in the 

 stream longer before they died and were con- 

 veyed downstream to the weir. 



Estimates of the populations of adult pink 

 salmon in the three areas made by the foot 

 survey method are compared with estimates 

 by other methods in table 3. The greatest dif- 

 ference between estimates (6 and 22 percent) 

 was for the fresh-water population in the East 

 Fork. The number of fresh-water spawners 

 estimated by the foot survey method was larger 

 than the number counted at the weir for all 

 areas except the West Fork in 1961. The weir 

 counts were known to be low because sections 

 of the weirs were removed several times during 



6 WllUam J. McNeil, William 0. McLamey, Clark T. 

 Fontaine, Robert Coates, David Brlckell, and Ralph Wells. 

 1964. Pink salmon studies at Little Port Walter, summer 

 and autumn, 1963. Manuscript on file. Bur. Commer. Fish. 

 Biol. Lab., Auke Bay, Alaska. 



