158 
Fishery Bulletin 113(2) 
fishermen. This reliance is problematic because recall 
bias can greatly reduce the accuracy and precision of 
discard estimates (Ditton et al., 1978; Holland et al., 
1992; Hanson and Sauls, 2011; Bochenek et al., 2012; 
Donaldson et al. 44 ). Uncertainty in estimates of bio- 
logical parameters may increase the risk of fisheries 
exceeding annual quotas or decrease their value for 
assessment models (Patterson et al., 2001a). Onboard 
observers, on the other hand, provide direct estimates 
of catch and discards that can be used to verify and ad- 
just fleet-wide estimates input into stock assessments 
(Kennelly and Broadhurst, 2002; Bochenek et al., 2012; 
Donaldson et al. 4 ; Sauls, 2014). 
The purpose of this study was to employ onboard ob- 
servers to directly estimate effort, catch, and discards 
in the for-hire recreational fishery for reef fishes in the 
northern GOM. A focus of this study was the red snap- 
per ( Lutjanus campechanus), which is the most highly 
targeted reef fish in the region and for which the es- 
timated ratio of total harvest (landed catch plus dead 
discards) to landed catch has averaged 1.5: 1.0 in re- 
cent years (SEDAR 5 ). A simple protocol was developed 
to incorporate observers into normal fishing operations 
onboard charter boats 1) to provide estimates of effort, 
catch, and discard variables on trips that targeted 
reef fishes, 2) to characterize changes in fishing be- 
havior and catch for both open and closed seasons for 
red snapper, and 3) to examine differences in the size 
and age composition of discarded red snapper for both 
open- and closed-season trips. 
Materials and methods 
Observers collected data during fishing operations on 
charter boats during 2012 and 2013, working onboard 
4 vessels that operated out of home ports between 
Destin, Florida, and Orange Beach, Alabama, the his- 
torical center of for-hire effort in the northern GOM 
red snapper fishery (SEDAR 5 ). Boat captains were 
contacted each week to obtain notice of a trip and to 
notify them of observer participation. Captains were 
compensated $250 for each trip because the presence 
of an observer meant the loss of one potential cus- 
tomer. Captains provided the depth, reef type (artifi- 
cial or natural), and approximate GPS coordinates for 
each site fished. Observers recorded the total number 
of fishermen onboard, the number of fishermen that 
fished at each site, the time spent fishing at each 
4 Donaldson, D., G. Bray, B. Sauls, S. Freed, B. Cermak, P. 
Campbell, A. Best, K. Doyle, A. Strelcheck, and K. Bren- 
nan. 2013. For-hire electronic logbook pilot study in the 
Gulf of Mexico: final report, 63 p. Report submitted to 
the Marine Recreational Information Program Operations 
Team. [Available from https://www.st.nmfs.noaa.gov/Assets/ 
recreational/pdf/Charter_Boat_Logbook_Project_report.pd£] 
5 SEDAR (Southeast Data, Assessment, and Review). 2013. 
SEDAR 31 — Gulf of Mexico red snapper stock assessment 
report, 1103 p. [Available from http://www.sefsc.noaa.gov/ 
sedar/.] 
site, and the terminal tackle used to catch each fish. 
General comments were recorded with respect to the 
types of fishing rigs used (e.g., multiple-hook bottom 
rigs versus single-hook bottom rigs suspended in the 
water column), and the manufacturer, model number, 
and dimensions of circle hooks were recorded. A Letter 
of Acknowledgment from the National Marine Fisher- 
ies Service’s Southeast Regional Office, in accordance 
with the definitions and guidance at Federal Register 
(2012), permitted observers to collect red snapper (of 
any size during any season) on the 4 vessels that par- 
ticipated in this study. 
All fish captured during normal bottom fishing op- 
erations (i.e., excluding fish caught when a vessel was 
trolling to and from offshore sites) were identified to 
species (Hoese and Moore, 1998; FishBase, http://www. 
fishbase.org), weighed to the nearest 0.1 kg, and mea- 
sured in fork length (FL) and total length (TL) to the 
nearest millimeter. Observers recorded whether fish 
were retained or discarded. Species that were identi- 
fied but accidentally released without having been 
measured (i.e., fish that escaped when they were 
hoisted into a boat) or could not be brought aboard 
(i.e., large elasmobranchs) were recorded and included 
in that trip’s discard total. A sample of red snapper 
discards was collected at each site during both open 
and closed seasons. These fish were systematically and 
randomly sampled through retention of every nth fish, 
depending on catch rate, such that approximately 20 
discarded red snapper were sampled on each trip. A 
similar systematic sampling routine was employed to 
sample approximately 20 red snapper retained by fish- 
ermen as landed catch during each open-season trip. 
All sampled fish were identified by affixing a number- 
coded plastic cable tie through the opercular and buc- 
cal cavity of each fish. 
The retained catch from each vessel was processed 
by the crew at the dock. Carcasses from those fish, as 
well as whole discards, were transported back to the 
laboratory at the Dauphin Island Sea Lab. Otoliths 
were collected from all fish and processed for aging. 
Otoliths were rinsed of adhering tissue and stored dry 
in plastic cell wells; they were then embedded in epoxy, 
sectioned with a low-speed watering saw, and aged ac- 
cording to standard protocols (Patterson et al., 2001b). 
Opaque zones were counted by 2 independent read- 
ers and average percent error was calculated between 
reader counts (Beamish and Fournier, 1981). 
Models for single-factor analysis of variance (ANO- 
VA) (a=0.05) were computed in R, vers. 3.1.0 (R Core 
Team, 2014; Crawley, 2007; Kabacoff, 2011) to test the 
effect of red snapper fishing seasons, open versus closed, 
on effort (distance from port, depth, number of fisher- 
men, number of sites fished per trip, time fished per site, 
and percentage of natural reefs fished), catch (number 
of species, fishes, and red snapper caught and percent- 
age of the catch that was red snapper), and discard 
(number of discards for all fishes and red snapper only) 
variables. Data were ln-transformed to achieve normal- 
ity and homogeneity of variances when necessary. 
