Weijerman et al. Trends in biomass of coral reef fishes, derived from creel surveys in Guam 
243 
of at least 3, we assumed that the mean biomass was the 
same between 2010 and 2012 and equivalent to that de- 
rived from the stationary-point-count surveys conducted 
in 2011. For each gear type, we derived smoothed esti- 
mates of q in accordance with Equation 3, with an n of 
3, / equal to the gear-specific annual CPUE values for 
2010-2012, and B equal to the summed biomass of func- 
tional groups exploited by the gear type. Information 
on which functional groups to include in the summed 
biomass was derived from the catch composition of total 
landings for 1985-2012. Functional groups that com- 
posed more than 10% of the catch per gear type were 
included: for example, over this time span, 23% of the 
targeted browsers (e.g., unicornfishes [Naso spp.]), 12% 
of the target grazers (most surgeonfishes [Acanthurus 
spp.]), and 37% of the roving piscivores (e.g., jacks) were 
caught by gill-net fishing; therefore, we added the es- 
timates of biomass from the visual surveys for those 3 
groups for the calculation of q for gill-net fishing. 
To reconstruct the annual total biomass for the pe- 
riod 1985-2012, we divided the gear-specific CPUE for 
each year by the estimated gear-specific catchability 
q, using Equation 2 rewritten as B t =I t /q. Within each 
year, we then calculated the reconstructed biomass for 
each functional group by multiplying the annual total 
biomass estimate by the species composition observed 
in 2011. For example, in 1985, the total biomass was 
estimated at 10,025 t, and excavators (large-bodied 
parrotfishes [e.g. Chlorurus spp.]) composed 1.3% of 
the species composition in 2011; given those values, the 
biomass of excavators in 1985 was estimated at 130 
t (10,025x0.013). Therefore, the results of this analy- 
sis indicate trends in relative fish stock size over time, 
with the assumption of constant catchabilities and spe- 
cies composition. 
Results 
Past and recent periods: comparing total catch, effort, and 
catch per unit of effort 
Hook-and-line was the most commonly used gear type, 
accounting for 59% of total estimated annual effort in 
1985-1990 (past) and 71% of total estimated annual 
effort in 2007-2012 (recent) (Table 2). Cast nets were 
the second most often used gear type, accounting for 
-14% of total annual estimated effort in both time pe- 
riods. Effort for all other gear types was reported less 
frequently in recent years than during the earlier time 
period, and declines in effort between the time peri- 
ods ranged from 16% for drag net fishing to 87% for 
hook-and-gaff fishing (Table 2). The decline in effort 
for those other gears was balanced by a 14% increase 
in effort for hook-and-line fishing, resulting in an es- 
timated total fishing effort that was similar in the 2 
time periods: an annual mean of 217,420 gear hours for 
1985-1990 in contrast to an annual mean of 205,850 
gear hours for 2007-2012. 
Reported catch was considerably higher during the 
period 1985-1990, with an estimated mean annual 
total catch of 100 t, compared with an average of 37 
t in the period 2007-2012 (Fig. 1, Table 2). The most 
noteworthy difference between the 2 time periods was 
a 74% decrease in the catch for cast net fishing despite 
little change in effort and a 21% decrease in the catch 
from hook-and-line fishing despite a 14% increase in 
effort. The largest decline in catch was for spearfish- 
ing: reported catch of snorkel spearfishing and scuba 
spearfishing dropped by 87% and 94%, respectively 
(with reported effort declining by 36% and 90%, re- 
spectively). However, the observed decline in estimated 
spearfishing effort and landings is certainly related to 
the refusal of some spear fishermen to participate in 
creel surveys since around 2005 (Lindfield et al., 2014). 
Based on the total landings for the shore-based fish- 
ery and the use of 95 km 2 as the reef area at depths 
of 0-30 m around Guam (determined with the adapted 
habitat maps of hard- and soft-bottom habitats), annu- 
al yields dropped from 1.30 t/km 2 during 1985-1990 to 
0.58 t/km 2 during 2007-2012. For reef-associated fishes 
(i.e. , excluding pelagic fishes, invertebrates, and non- 
reef-associated species), annual yield was 1.59 t/km 2 
and 0.74 t/km 2 , respectively, and 71 km 2 were used as 
the reef area for hard-bottom habitat at depths of 0-30 
m around Guam (determined with one of the adapted 
habitat maps). 
Hook-and-line fishing was assumed to be the most 
reliable data source because hook and line gear was 
the dominant gear type in the catch and effort records. 
On the basis of the fishery data for this gear type only, 
landings showed no clear trend over time, albeit with 
large annual fluctuations (Fig. 2A). However, landings 
of species other than bait fishes (e.g., bigeye scad [Selar 
crumenophthalmus]) showed a downward trend (Fig. 
2A). The mean annual CPUE, based on the unexpanded 
interview data, for hook-and-line fishing varied because 
of the interannual fluctuation in bait fish landings, but 
the overall mean CPUE decreased 49% between the 2 
time periods, from an annual mean of 0.15 kg/gear hour 
(standard error [SE] 0.02) for 1985-1990 to 0.08 kg/ 
gear hour (SE 0.02) for 2007-2012 (Table 2, Fig. 2B). 
CPUE was also significantly lower in the recent time 
period for gill net fishing and snorkel spearfishing, 
compared with levels in the past. The only gear type 
for which CPUE did not significantly decrease (P=0.1) 
was the gear type other methods (i.e., reef gleaning), 
although CPUE for that type trended downward (Table 
2). For the remaining gear types, there was not enough 
information to statistically compare differences be- 
tween the time periods (annual number of interviews 
was less than 3); however, for all gear types, except 
hooks-and-gaffs, mean CPUE was lower in the recent 
time period and the overall trend showed a decline in 
CPUE (Appendix 3 Figure). 
Species composition of landings 
In most years, combined landings of browsers (predomi- 
nantly unicornfishes and rabbitfishes) and invertebrate 
