Weijerman et al. Trends in biomass of coral reef fishes, derived from creel surveys in Guam 
247 
Figure 4 
Reconstruction of relative biomass for 3 trophic groups for the period 1985- 
2012 based on a constant gear catchability coefficient per gear type and the re- 
constructed biomass determined with catch-per-unit-of-effort data from shore- 
based creel surveys conducted by the Guam Division of Aquatic and Wildlife 
Resources and with estimates of biomass, derived from the fishery-indepen- 
dent surveys conducted in 2011, for the functional groups targeted in shore- 
based creel surveys. Piscivores include mid-water piscivores te.g., barracudas), 
benthic piscivores (e.g., groupers), and roving piscivores (e.g., jacks); inverte- 
brate-consuming species include target species of this category; and herbivores 
include grazers (e.g., surgeonfishes), browsers (e.g., unicornfishes), scrapers 
(small-bodied parrotfishes), and excavators (large-bodied parrotfishes). 
CPUE and catch began before the 
first period for which we have 
creel data (before 1985). Fishery 
yields of reef fishes around Guam 
also have been low compared with 
previously published global yields 
for coral reef fisheries — annual 
levels that ranged from 0.2 to 44.0 
t/km 2 (Dalzell, 1996). These global 
values are, however, confounded 
by differences in species group- 
ings and effort reported for differ- 
ent locations around the world, as 
well as by the different assump- 
tions about habitats included in 
area calculations. Still, our results 
clearly show that the downward 
trend in catch and CPUE, a pat- 
tern that earlier studies already 
noted (Hensley and Sherwood, 
1993; Myers, 1993), has continued, 
and another study has reported 
that declines in catch started at 
least 50 years ago (Zeller et al., 
2015). 
Declines in catch are not re- 
stricted to Guam. A recent synthe- 
sis of domestic catch reconstruc- 
tion for 25 Pacific island countries 
and territories from 1950 to 2010 
showed that, at 60% of those lo- 
calities, peak catch occurred before 2000 and that in 
only one location did the peak catch occur after 2010 
(Zeller et al., 2015). Also noteworthy is the shift in 
catch composition, where the majority of the landings 
in recent years comprises lower-valued species, such 
as the bigeye scad ( Selar crumenophthalmus). This 
change in catch composition could be a result of the 
large pulses in bait fish populations, but it could also 
indicate a change in composition of the fish community. 
A similar change in catch composition was found in an 
analysis of landings by spear fishermen (Lindfield et 
al., 2014). Scuba spearfishing around Guam was associ- 
ated not only with a decline in the size of parrotfishes 
caught but also with a shift from a dominance of large 
parrotfishes to a mixed assemblage with increasing 
proportions of surgeonfishes. This shift in community 
structure and a possible loss in biodiversity could lead 
to altered ecosystem functioning. Shifts in community 
structure and an altered ecosystem functioning have 
been observed in Kiritimati, where, in comparison with 
an unfished location (Palmyra Atoll), apex predators 
were relatively smaller, had a shorter life span, and 
were less abundant and prey fishes were larger but not 
more abundant (Ruttenberg et al., 2011). In a global 
analysis of reef fish surveys, Mora et al. (2011) found 
a strong link between ecosystem functioning and bio- 
diversity, the latter of which is negatively influenced 
by human populations through overexploitation or the 
loss or degradation of habitats. Effective efforts to con- 
serve biodiversity should, therefore, include the pres- 
ervation of the functional roles that species perform 
(Brown and Mumby, 2014). 
Although we are confident that our results provide a 
good indication of trends in CPUE and in the derived 
relative fish biomass, our study has several limitations. 
Calculating the CPUE for a multispecies, multigear 
reef fishery is not straight forward. Among other issues, 
many factors influence the decisions that fishermen 
make about when and where to fish and which gear to 
use. Consequently, there are undocumented changes in 
selectivity or catchability of specific gear types over dif- 
ferent temporal and spatial scales. Accounting for all of 
these dependent decisions of fishermen was beyond the 
scope of our study; however, there was no quantitative 
or anecdotal information that documented any system- 
atic changes in those decisions over the time period 
for which we have data. Also, too little creel data are 
available for some gear types, such as cast nets, and 
for spearfishing, and that lack of data clearly limits the 
scope by which the WPacFIN-expanded data can fully 
represent the total catch and effort. 
Additionally, the assumption that the species com- 
position has stayed the same between 1985 and 2012 
is also unlikely to be true because the catch composi- 
tion did change. Fishing affects the species composition 
of piscivores (e.g., sharks and jacks), and large fishes 
are often the first ones to disappear from catch records 
(Friedlander and DeMartini, 2002; Fenner, 2014). We 
