Squire Aerial monitoring of abundance of pelagic fishes 



357 



From inspection of core-area fluctuations (Fig. 8), it 

 is difficult to draw any conclusions regarding interac- 

 tion among most of the species included in this study. 

 The annual abundance values that appear to be more 

 closely related are those of the northern anchovy and 

 bluefin tuna which were significantly correlated 

 (y=0.76,P<0.1). 



Pilot estimation of tonnage 



Differences between pilots' ability to estimate school 

 or group size are difficult to evaluate using the type of 

 data recorded in flight logs. Computer recording of 

 flight track and species tonnage observed did not in- 

 clude details of what portion of the block was searched, 

 or where the species were sighted. Aerial spotters en- 

 tering the block area (an 8xl0nm area) may have 

 observed fish schools of the same species in two geo- 

 graphically different locations. Only the sum of the 

 school sizes was coded. 



The time of observation is probably the most critical 

 factor in comparisons of pilot sightings because near- 

 surface fish abundance may be highly variable over a 

 short period of time. Therefore, unless pilots are search- 

 ing the same area at the same time, different levels of 

 abundance are likely to be observed. 



Williams (1981) in surveys of pelagic fish resources 

 off southeast Australia indicated that the professional 

 fish spotters were extremely accurate in tonnage esti- 

 mates of individual schools. In tests conducted by 

 NMFS and CDF&G to determine northern anchovy 

 school size using acoustic gear, a purse seiner and aerial 

 spotter were used. The aerial spotter had a more accu- 

 rate estimate of school tonnage than obtained from 

 acoustic gear (P. Mardesich, Flying Fishermen, Inc., 

 pers. commun. 1978). 



MacCall ( 1975), using the NMFS computer database, 

 investigated variation in anchovy abundance estimates 

 (comparison of data collected during the same lOd pe- 

 riod in 1973 and 1975) and found considerable differ- 

 ences among spotters. The differences were consistent 

 for individual pilots for day and night observations. 

 Variability in reported tonnage was large, but it could 

 be reduced by long-term averaging. MacCall ( 1975) fur- 

 ther suggested that corrections for pilots estimating 

 efficiency be incorporated in calculations of the abun- 

 dance index to avoid biasing the result toward the 

 spotter with the highest tonnage. The accuracy of fish- 

 spotter estimates of tonnage was approached statisti- 

 cally by Lo et al. (In press) using a delta-lognormal 

 model. All aerial spotter data were used in this analy- 

 sis. However, until spotter-pilot estimating efficiency 

 is measured in field experiments, correction factors 

 are unknown. It is possible that these differences be- 



tween spotters may be due more to differences in num- 

 bers of schools seen than to differences in size esti- 

 mates. 



Comparison of aerial index to other 

 abundance indices 



The problem of variability among spotter pilots com- 

 plicates comparisons; however, pilots are paid relative 

 to the amount of tonnage caught. Therefore, the aerial 

 spotter needs to have a reasonable estimate of school 

 size to compare with the amount of fish caught by the 

 seiner. The aerial spotter is continually obtaining 

 "ground truth" from the purse-seine vessel relative to 

 species composition and size of the school caught. 



Since the decline of the Pacific sardine resource in 

 1984, considerable effort has been expended on sur- 

 veys, using various techniques to collect apparent abun- 

 dance data that could be modeled to produce an esti- 

 mate of biomass. Acoustic, egg and larva, and 

 aerial-spotter surveys have been conducted. Each sur- 

 vey measures a different component of the population 

 and, therefore, comparisons of abundance estimates 

 from different methods and models must be viewed 

 with caution. 



The aerial-spotter program measures apparent abun- 

 dance of adults and subadults of epi-pelagic schooling 

 fishes, and assumes that most adults in the popula- 

 tion are available to the spotters at some time during 

 the year. The aerial index is a direct calculation (T/ 

 BAF). It is a non-random type survey, but is not time- 

 restricted. 



The operational procedures used by aerial spotters 

 affect the results. Although they do work closely with 

 the fleet in locating and sometimes directing the set- 

 ting of the purse seine, much of their time is spent 

 searching large areas of the ocean for fishable resources. 

 Many times, the aerial spotter will conduct flight op- 

 erations in a general survey mode for the fleet, but 

 does not work directly with the fleet. Even though 

 aerial monitoring is conducted throughout the year, 

 most observation effort is in the summer and early fall 

 (49% of BAF during July, August, and September) when 

 concentrations of commercial species are more com- 

 mon in the nearshore areas. 



Quantitative apparent-abundance data derived from 

 field surveys (egg and larvae, and acoustic) are statis- 

 tically treated under various biological assumptions to 

 ultimately obtain an estimate of biomass or 

 tonnage: spawning biomass and total biomass in the 

 case of egg and larva surveys (Lo 1985), and schooled 

 biomass in the case of acoustic surveys (Mais 1974). 

 Although some biomass estimates have been made for 

 other species, such as the chub mackerel and Pacific 



