Gorfine et al.: Two methods for estimating abundance of Haliotis rubra 
443 
Figure 2 
Comparison of mean abundance estimates of H. rubra at Gabo Island 
among surveys using radial transects. Error bars are 95% confidence 
limits. 
(effect size=9%) than for timed searches (ef- 
fect size=44%); thus power in the analysis 
was likely to be low. 
Trial stock surveys 
with radial transects 
There were significant differences in den- 
sity between all locations for the prerecruit 
and postrecruit size classes, but locations 
were not significantly different for juvenile 
abalone (Table 1). Locations and sites 
within locations accounted for most of the 
variation in density of pre- and postrecruit 
abalone. Significant differences between 
divers occurred for postrecruits, but the 
magnitude of the variance effect for divers 
was only about one third of the variance 
effect for sites. Sites varied significantly 
within locations for each size class; for ju- 
veniles and post recruits this was at the 10% 
confidence level, and for prerecruits at the 1% confi- 
dence level. Only one interaction effect, diver x site 
(location) for prerecruit density, was significant. 
Application of methods to stock monitoring 
Variation in abalone abundance estimates from both 
methods showed significant diver effects in two out 
of three instances (Tables 2 and 3). However, the only 
significant year effect detected was for estimates from 
radial transect surveys in the central zone of the fish- 
ery. The lack of a significant diver effect in this in- 
stance may have increased the power to detect a year 
effect. Consistent with the high spatial variability 
that characterizes abalone populations, site effects 
were significant in all instances. Year x site interac- 
tions were significant in two instances for each 
method, indicating that interannual changes in abun- 
dance varied among sites within the affected zones. 
Only one year x diver interaction was significant and 
because some divers were not represented in each 
year, the interpretation of this effect in this instance 
is problematic. 
For each method only a relatively small amount of 
the variation in abundance could be attributed to the 
effect of year. Sites generally accounted for most of 
Diver 1 
Diver 2 
Diver 3 
Diver 4 
Figure 3 
Comparison of mean abundance estimates of H. rubra 
at Gabo Island among divers using radial transects. 
Error bars are 95% confidence limits. 
70 
60 
50 
40 
30 
20 
10 
Radial transect 
Timed search 
o -I . 1 , i 
Diver 1 Diver 2 Diver 1 Diver 2 
Figure 4 
Diver effects on mean abundance estimates of H. rubra for each 
survey method. Error bars are 95% confidence limits; base sam- 
pling units were 30 m 2 for radial transects and 10 min for timed 
searches. 
