288 
N 
< 
2 
7) 
f= 
oO 
£ 
QO 
-0.5 
Dimension 1 
Figure 6 
Correspondence analysis for dimensions 1 and 2 for the data set from deple- 
tion experiments conducted during 1997-2011 to examine efficiency of 
hydraulic dredges for capturing clam species and to estimate stock density 
for populations of ocean quahogs (Arctica islandica) and Atlantic surfclams 
(Spisula solidissima) off the mid-Atlantic coast of the United States. Error 
terms Err (R1), Err2 (R2), Err3 (R3), and Err4 (R4) are supplementary vari- 
ables. Estimates of the following characteristics are entered as quartiles: 
dredge efficiency (E); clam density (D); the k parameter (K), which is the 
negative binomial dispersion parameter; coefficient of variation (CV) of effi- 
ciency (C); CV of density (N); CV of the & parameter (P); effective area swept 
(S); overlap score (T); latitude (L); and depth (Z); only quartiles 1 and 4 are 
shown in the plot. Other variables that describe the experiments include 
species, the ocean quahog (O) and Atlantic surfclam (S); region, Long Island 
(LI) in New York, New Jersey (NJ), and the Delmarva Peninsula (DMV) of 
Delaware, Maryland, and Virginia; and dredge width. Dredge widths are 2.54, 
3.05, 3.30, and 3.81 m. Error estimates are entered as 1 (below the 80th per- 
centile) or 2 (at or above the 80th percentile). The inner box demarcates the 
area with loading factors from —0.5 to 0.5 on both axes. 
Fishery Bulletin 119(4) 
dredge width (3.81 m [12.5 ft]); experi- 
ments with dredges of this size clearly 
had superior performance. 
High OS in a depletion experiment 
does not always reduce uncertainty in 
Patch model estimates. An explanation 
for this may come from the pragmatic 
efforts of a field experiment. Depletion 
experiments are costly in vessel time and 
crew effort, often requiring more than 
8 h of nearly continual dredging. Cost at 
sea was sufficient in that adaptive time 
management during the experiment was 
directed at limiting tow number, albeit 
with limited empirical guidance to deter- 
mine the stopping point for the depletion 
experiment. 
One consequence of adaptive time 
management during the depletion exper- 
iment was a decision to add tows if the 
experiment appeared not to be generat- 
ing a clear and consistent reduction in 
catch per tow. Results of correspondence 
analysis indicate the danger of the use 
of adaptive decisions during depletion 
experiments without rigorous empirical 
determination criteria designed to opti- 
mize the cost and benefit of increasing 
tow number. The danger of terminating a 
depletion experiment early on the basis of 
a potentially misleading depletion curve 
was present as well. The OS did not fall 
out cleanly in any of the dimensions on 
the plots of correspondence analysis, the 
opposite of what was expected given the 
clear improvement afforded by higher 
tow numbers, and more tow overlap, in 
the simulation study of Poussard et al. 
(2021). However, the absence of OS did 
not diminish its effect on estimates of 
gear efficiency in depletion experiments. 
Notably, OS and the distribution of clams 
on the bottom (e.g., NP, P, T, or HP), which 
in some fashion are measures of dredge 
average, and their inclusion in the depletion data set may 
bias the overall efficiency estimates used to inform stock 
assessments. 
In correspondence analysis, Err2, Err3, and Err4 also 
fall on the same dimensional axis as a low EAS value. Low 
EAS and low efficiency generally occur together because 
the efficiency value is a variable in the equation determin- 
ing EAS (Equation 3). The relationship is well-documented 
by Poussard et al. (2021). This expectation is confirmed by 
Pearson’s correlation coefficients calculated by using the 
field depletion experiment data set (Figs. 4 and 5). The EAS 
is also positively correlated with year for experiments that 
targeted ocean quahogs and with dredge width for exper- 
iments with both ocean quahogs and Atlantic surfclams 
(Figs. 4 and 5). The relationship is driven by the largest 
overlap with tow paths or with clams in the area, are 
both relatively unbiased parameters. That is, they are not 
associated with any depth, dredge size, species, or other 
characteristic of the experiment. The distribution of clams 
relative to the distribution of tows is a critical constraint 
on efficiency estimation. 
Results of correspondence analysis clearly reveal the 
relationships earlier identified by using the Wilcoxon 
rank sum tests and by Pearson’s correlation coefficients. 
The error terms Err2, Err3 and Err4, which the results of 
Wilcoxon analyses indicate were highly significant, fall 
on the positive side of dimension 1 along with the param- 
eters and experiment characteristics significantly influ- 
enced by them. The error term Err1, which did not have 
significant differences in the Wilcoxon rank sum tests, 
