104 
Fishery Bulletin 113(2) 
and initiated a calibration study in which the catch- 
ability of the old vessel was compared with the new 
vessel (Miller et al. 6 ). 
Data 
lowing equation, provided the degree of clustering for 
points within a given distance (d): 
I n Eij^ijUi -x) 
d_ Q>ij)£r=i(*i-*) 2 ’ 
CPUE from the NEFSC bottom trawl survey was used 
as an index of relative abundance of spiny dogfish in 
the NES LME and was defined as the number of spiny 
dogfish caught per tow. To enable investigation of fish- 
ery behavior (i.e., effort and catch), CPUE was used to 
index local fish density and was defined as the number 
of spiny dogfish caught per hour fished. Spatial overlap 
analyses between spiny dogfish distribution and com- 
mercial fisheries were based solely on positive catches 
(i.e., CPUE >0) and did not depend upon magnitude. 
No attempts were made to standardize CPUE between 
gear types or compare magnitudes directly in any spa- 
tial analyses. Differences between effort allocation, 
gear configuration, and catchability invalidated direct 
comparison of any trends in relative abundance be- 
tween the survey and fisheries. Spatial locations were 
provided by latitudes and longitudes reported within 
both data sets. 
Sampling during NEFSC bottom trawl surveys typi- 
cally occurred over an 8-week period and proceeded 
from Cape Hatteras, North Carolina, north to the GM 
(Rago, 2005). These surveys generally lasted from Sep- 
tember through November during autumn and from 
March through May during spring. Point data from 
the NEFOP showed that major fisheries, including the 
SGN and OT fisheries, covered both inshore and off- 
shore regions of the NES LME between 1989 and 2009. 
The footprint of the SGN fishery and its coverage by 
observers expanded in that period. To enable temporal 
comparisons of spiny dogfish occurrence in the survey 
and in each fishery, the only fishery-dependent data 
that were used in analyses were collected during these 
time periods. Spatially, spiny dogfish distribution was 
comparable with fishery effort and catch because the 
overall footprint identified from the NEFOP covered 
much of the NES LME. 
Spatial distribution 
Spatial autocorrelation To assess the spatial pattern 
of spiny dogfish within each commercial fishery and 
the NEFSC bottom trawl survey, the spatial depen- 
dency among observations of spiny dogfish CPUE in 
geographic space (i.e., spatial correlation) was assessed 
through the use of Moran’s 7 statistic (Moran, 1948; 
Goodchild, 1986). This statistic, calculated with the fol- 
nental Shelf of the United States. ICES Committee Meeting 
(C.M.) document 2007/Q:20, 25 p. 
Miller, T. J., C. Das, P. J. Politis, A. S. Miller, S. M. Lucey, C. 
M. Legault, R. W. Brown, and P. J. Rago. 2010. Estima- 
tion of Albatross IV to Henry B. Bigelow calibration factors. 
Northeast Fish. Sci. Cent. Ref. Doc. 10-05, 233 p. [Avail- 
able at http://www.nefsc.noaa.gov/publications/crd/crdl005/ 
crdl005.pdf.] 
where n = the number of observations, x; and xj are the 
attribute values (CPUE >0) at points i and 
j; 
x = the mean CPUE; 
it;jj = the weighting function (mjj=l if points are 
within d, otherwise my=0) (Nielsen et al., 
2007); and 
ly = the sum over i and j with i±j. 
Moran’s I tested the null hypothesis of a random spa- 
tial pattern in spiny dogfish CPUE (i.e., Moran’s 7=0) 
with values ranging from -1 (dispersed) to +1 (clus- 
tered). Moran’s 7 was calculated in R software, vers. 
2.14.0 (R Development Core Team, 2011) with the sp- 
dep package and a spatial weights matrix based on the 
5-nearest neighbors (Bivand et al., 2012). 
Center of abundance Annual centers of spiny dogfish 
abundance (Marino et al., 2009) were estimated to 
identify and compare interannual locations of spiny 
dogfish catch for each fishery and the bottom trawl 
survey. This metric was calculated with the following 
equation: 
EtifrjXij 
Zb i ’ 
( 2 ) 
where Xj = the parameter of interest (latitude, longi- 
tude) at station i in year j; and 
b{ - the log e -transformed abundance (log e [CPUE 
>0]+0.05) (Nye et al., 2009). 
Annual centers of spiny dogfish abundance were 
mapped in ArcGIS 7 , vers. 10.1 (ESRI Corp., Redlands, 
CA). Annual centers of spiny dogfish abundance with- 
in each fishery do not reflect true shifts in distribu- 
tion but instead reflect locations of catches (whether 
targeted or occurring as bycatch) on fishing grounds. 
In contrast, centers from the bottom trawl survey are 
unbiased representations of spiny dogfish distribution 
and, therefore, estimate true changes in the population 
distribution in the survey domain. 
Annual centers of spiny dogfish abundance docu- 
mented during the bottom trawl survey and each fish- 
ery were compared to determine whether the spatial 
locations of abundance differed for each season. De- 
spite transformation efforts, non-normality and highly 
correlated dependent variables prevented the use of 
parametric techniques (Quinn and Keough, 2002). In- 
stead, a one-way permutational multivariate analysis 
of variance (Anderson, 2001), for analysis of variance 
with a balanced design based on Bray-Curtis distances, 
was used to test for differences in location through the 
7 Mention of trade names or commercial companies is for iden- 
tification purposes only and does not imply endorsement by 
the National Marine Fisheries Service, NOAA. 
