Essington et al.: Shifts in the estuarine demersal fish community after a fishery closure in Puget Sound, Washington 
209 
Table t 
Summary of total catches for species or species groups that were commonly collect- 
ed during bottom trawl surveys conducted from 1991 to 2012 in Port Madison, Puget 
Sound, Washington. A full listing of all species collected is presented in the appendix 
table. 
Catch 
Species (number of fish) 
Percentage 
of total 
Blackbelly Eelpout ( Lycodes pacificus) 
3398 
9.3 
Dover Sole (Microstomus pacificus ) 
366 
1.0 
English Sole (Parophrys vetulus ) 
10,427 
28.6 
Flathead Sole ( Hippoglossoides elassodon ) 
1665 
4.6 
Pacific Hake (Merluccius productus ) 
958 
2.6 
Pacific Herring ( Clupea pcillasii) 
478 
1.3 
Pacific Tomcod ( Microgadus proximus) 
2677 
7.3 
Plainfin Midshipman ( Porichthys notatus ) 
497 
1.4 
Rock soles ( Lepidopsetta bilineata and L. polyxystra) 
773 
2.1 
Sanddabs ( Citharichthys sordidus and C. stigmaeus ) 
1075 
3.0 
Sand Sole ( Psettichthys melanostictus) 
680 
1.9 
Shiner Perch ( Cymatogaster aggregata) 
1139 
3.1 
Slender Sole ( Lyopsetta exilis) 
2242 
6.2 
Spotted Ratfish ( Hydrolagus colliei) 
4068 
11.2 
Sole ( Parophrys vetulus), the most common species, 
available aging data (senior author, unpubl. data) in- 
dicate that this procedure effectively separates age-1 
individuals from those individuals aged 3 years and 
older. 
Results 
During the 20-year survey 65 fish species were sam- 
pled, and the 14 species that were sampled frequently 
enough (>200 individuals) to evaluate time trends ac- 
counted for more than 85% of the total catch (Table 1). 
Notably, 7 of these species were flatfishes (Pleuronec- 
tidae and Paralichthyidae). English Sole and Spotted 
Ratfish were by far the most common species, collec- 
tively, contributing more than 40% of all individuals 
sampled. Other common species included Blackbelly 
Eelpout ( Lycodes pad ficus ) and Pacific Tomcod (Micro- 
gadus proximus). 
Time series of catch anomalies were nonstationary 
for most species (Fig. 2). Spotted Ratfish was the only 
species for which no trend or apparent change in abun- 
dance over the sampling period was observed (Table 2; 
Fig. 2). As for changes in abundance values that were 
seen, a change point was identified for 9 species and a 
continuous linear trend was found for only 3 species. 
In all cases where the change-point model provided 
the best fit to the data, the relationship indicated a 
reduction in the mean catch rate in the later portion 
of the time series. These cases included the one for 
the most abundant species, English Sole, for which the 
mean catch anomaly shifted from +27/tow before 1998 
to —7/tow afterward. In contrast, increases in the mean 
catch anomaly were observed for all 3 species for which 
abundance trends were best described by the linear 
model. Trends in total catch (unstandardized) summed 
across all species mirrored the trends of English Sole 
(Fig. 2). 
Many species exhibited changes in catch rates at 
similar time periods. Of the 9 species whose dynamics 
were best described by a change-point, 5 species had es- 
timated change points between 1997 and 1999 (catches 
were not sampled in 1998). Three species had change 
points between 1999 or 2000 and 2000 or 2001. There- 
fore, there was evidence of a change in catch rates be- 
tween 1997 and 2001 reflected by several species. 
Analysis of catch-anomaly trends among different 
size classes of flatfishes did not support the hypoth- 
esis that trends were driven by changes in recruitment 
(Fig. 3). For most species, anomalies for small- and 
large-size fishes were synchronous with no apparent 
lag. On the basis of the length-frequency distribution 
of each species or species group, fishes were catego- 
rized as small if they fell in the 33 rd percentile or 
lower and as large if they were assigned to the 67 th 
percentile or higher. For instance, the catch anomalies 
for English Sole were nearly identical between small 
(age 1) and large (age 3+) size classes. Catch-anomaly 
trends for the rock sole species group were more con- 
sistent with recruitment shifts because catch anom- 
alies of small rock soles declined steeply after 1997 
but catch anomalies for large rock soles had a 
less sudden and delayed decline. For Dover Sole (Mi- 
crostomus pacificus ) and sanddabs, nearly all the 
variation in catch anomaly was attributed to large-size 
individuals; catches of small individuals changed 
little. 
