Kemberling and Darnell: Distribution, abundance, and reproductive output of spawning female Callinectes sapidus 349 
crab that were female for that year was used. Sex was 
not determined for blue crab in 2004; therefore, the over- 
all mean proportion that was female (0.930) was used. 
Catch per unit of effort (CPUE) of adult female blue crab 
was calculated by dividing the adult female catch (mea- 
sured as number of individuals) by the area swept by 
the trawl (calculated by using the recorded speed and 
duration of a tow and by assuming a 12.2-m net spread). 
Values of CPUE are given as number of individuals per 
hectare. 
To examine overall temporal and spatial patterns in 
CPUE of blue crab in the SEAMAP survey area, adult 
female catch was modeled by using a 2D generalized 
additive model (Potts and Rose, 2018) with a negative 
binomial error distribution and log link function. Mul- 
tiple candidate models were fit and compared by using 
Akaike information criteria (AIC) to determine the most 
appropriate model. Adult female catch was the response 
variable. Year was included in all models as a categori- 
cal variable. Geographic location was included as a 2D 
smooth function with thin-plate spline bases. Season 
(summer or fall) was included as an additional temporal 
covariate. All models also included an offset term for the 
area swept by the trawl to account for differences in sam- 
pling effort among trawl] tows. Following model selection, 
a Wald chi-square test was used to test for significance of 
predictor variables. 
An index of abundance for the spawning stock in the 
northwestern Gulf of Mexico was estimated to allow 
examination of population trends, accounting for spatial 
and temporal variation in sampling effort. Estimated mar- 
ginal means for each year were extracted from the model 
output, reflecting the main effect of year averaged over the 
full study area and across both seasons. Estimated mar- 
ginal means were then standardized relative to the aver- 
age annual estimated marginal mean across the study 
period by using a Z-transformation, providing an index of 
abundance in units of standard deviations above or below 
the mean for the 20-year period for which we analyzed 
survey data. 
Relative importance of offshore and estuarine waters 
To assess the importance of offshore spawning areas rela- 
tive to estuarine spawning areas for the spawning stock of 
blue crab in the Gulf of Mexico, total abundance of adult 
female blue crab was estimated and compared between off- 
shore and estuarine waters of the study area. Abundance 
estimates were limited to the areas west of the Mississippi 
River (SSZ 13-21), because of the lack of consistent and 
spatially widespread fishery-independent trawl sampling 
for blue crab in estuarine waters of the state of Mississippi. 
Temporal coverage of abundance estimates was limited to 
the period 2006—2018 because of limitations on availabil- 
ity of data for sampling in estuarine areas in other years. 
Observed CPUE was used as a proxy for density of blue 
crab. Because gear efficiencies are unknown for the trawl 
surveys examined in our study, we were unable to cor- 
rect for gear efficiency; actual densities of blue crab, and 
therefore abundances, are likely substantially higher than 
estimates calculated in our study. 
Offshore abundance of blue crab was estimated by using 
the annual mean observed CPUE of adult female blue 
crab from the summer SEAMAP groundfish trawl survey 
and the total area of potential offshore spawning habitat, 
defined as all waters in the Gulf of Mexico west of the Mis- 
sissippi River (SSZ 13-21) from the estuarine mouth or 
barrier islands to the 109.7-m (60-fathom) depth contour 
(depth contours were derived from gridded bathymetry 
data from the General Bathymetric Chart of the Oceans 
(GEBCO_2020 Grid, available from website). Annual mean 
CPUE values, meaning numbers of individuals per hect- 
are, were multiplied by the total area of potential offshore 
spawning habitat to estimate total abundance of blue crab 
in offshore waters for each year. These estimates of off- 
shore abundance were compared with similar estimates of 
abundance in estuarine waters of Texas and Louisiana, by 
using CPUE from fishery-independent trawl surveys con- 
ducted in estuarine areas by the Texas Parks and Wildlife 
Department (TPWD) and Louisiana Department of Wild- 
life and Fisheries (LDWF). Only data collected during 
June and July were included in these analyses because 
these summer months represent the time of peak adult 
female abundance in estuarine waters of Louisiana and 
Texas and also correspond with the timing of the summer 
SEAMAP trawl! survey. 
The TPWD conducts fishery-independent sampling by 
using 6.1-m trawl nets with 38.1-mm stretched mesh, 
towed for 10 min at a speed of 4.8 km/h. Sampling is based 
on a stratified random design, with each of the 8 major 
bay systems in Texas (Sabine, Galveston, Matagorda, San 
Antonio, Aransas, Corpus Christi, Upper Laguna Madre, 
and Lower Laguna Madre) divided into grid cells that 
were 1.85 by 1.85 km. Each month, 10-20 grid cells in 
each major bay system are randomly selected for trawl 
sampling, with the number of cells depending on the bay 
system. Statewide annual mean CPUE of adult (=125 mm 
CW) females was calculated for the period June—July 
and then was multiplied by the total available habitat 
area to estimate the total abundance of mature females 
in estuarine waters of Texas for each year. Habitat area 
was defined as the total area of the Texas coastal bay sys- 
tems extracted from geographic information systems data 
obtained from TPWD (available from website). 
The LDWF conducts fishery-independent sampling 
by using 4.9-m flat trawl nets with a 38.1-mm stretched 
mesh body and a 12.7-mm stretched mesh bag, towed for 
10 min. Sampling occurs at fixed stations, semimonthly 
during November—February and weekly during March— 
October. Vessel speed is not recorded but typically ranges 
from 2.6 to 4.8 km/h (Cagle®). For CPUE calculations, an 
average vessel speed of 3.7 km/h was assumed because it 
is the midpoint of the typical range. Because stations were 
added and deleted during the period of interest, we used 
only data from stations that were sampled in 2018 and had 
® Cagle, P. 2020. Personal commun. Louisiana Dep. Wildl. Fish., 
P.O. Box 98000, Baton Rouge, LA 70898. 
