McKinzie and Szedimayer: Mortality estimates for Balistes capriscus based on acoustic telemetry 179 
of body orientation, and resumption of normal swimming 
motion. Recovered fish were then placed into a weighted 
rectangular cage (46 x 61 x 61 cm) made of plastic-coated 
13-gauge wire and quickly lowered to the seafloor (depths: 
19-31 m) within 10 m of the capture site. Once the cage 
reached the seafloor, a door automatically opened, allow- 
ing the tagged fish to leave on its own initiative (Williams 
et al., 2015). Any tagged fish that did not leave the cage 
after 20 min on the seafloor were not released. 
_ Validation of detection data 
Acoustic receivers can generate false detections that are 
not from valid transmitters in tagged fish (Pincock’). False 
detections can result from incomplete transmission due to 
interference (noise) or collision of signals when 2 or more 
transmitters simultaneously reach a receiver (Pincock’). 
To reduce the potential for data loss due to signal colli- 
sions, no more than 10 transmitters (tagged fish plus a 
control transmitter) were deployed at any individual 
VPS site at any one time (Topping and Szedlmayer, 2011; 
Piraino and Szedlmayer, 2014; McKinzie, 2018). False 
detections that resulted in the production of unknown tag 
IDs were removed from all subsequent analyses. Trans- 
mitter detections of known tags were screened before 
they were accepted as valid. Transmitter detections were 
accepted as valid if there were a minimum of 2 detections 
for a single transmitter ID and if there was at least one 
short interval between detections and more short inter- 
vals than long intervals. The short interval time was set 
at 30 min (30 times the nominal ping interval of 60 s). 
The long interval was set at 12 h (720 times the nominal 
ping interval of 60 s; Pincock”). All false detections of valid 
transmitters were removed from analysis. 
Survival and mortality estimates 
Fish positions (latitude and longitude) were calculated 
by Innovasea Systems from the time differential of signal 
arrival at 3-5 receivers. Status of each fish tagged with 
a transmitter was based on positions and the time inter- 
val between detections after a 3-d post-tagging recovery 
period. Any fish that remained stationary after exiting 
their release cage or that permanently emigrated away 
from the VPS site within the 3-d recovery period were con- 
sidered lost and were removed from analyses. After 3 d, 
tagged gray triggerfish were categorized as active (contin- 
uously swimming around a VPS site), emigrated (progres- 
sively moving farther away from a VPS site until it was 
lost or had moved to a surrounding reef site), dead as a 
result of natural causes (M, tag became stationary or data 
indicates erratic large-scale movements, e.g., shark move- 
ment patterns; Altobelli and Szedlmayer, 2020), or dead 
or removed from stock as a result of fishing (F', suddenly 
disappeared from a VPS site). 
2 Pincock, D. G. 2012. False detections: what they are and how to 
remove them from detection data. Amirix Syst. Doc. DOC-004691, 
vers. 3, 11 p. [Available from website, accessed February 2022.] 
Fish that emigrated were frequently detected with 
receivers at surrounding reef sites, and fish that experi- 
enced a mortality event were not detected at surround- 
ing sites. Fishing mortalities were also confirmed by tags 
returned by fishermen. To increase the probability that 
fishermen would return tags, a high monetary reward was 
offered ($150). Posters describing the study and reward 
offer were posted at marinas, bait shops, and other public 
sites and on the Auburn University fish tagging website. 
It was assumed that tagged gray triggerfish active within 
the VPS sites where they were tagged and released expe- 
rienced similar mortality rates to untagged fish outside of 
these VPS sites (Williams-Grove and Szedlmayer, 2016a; 
Mudrak and Szedlmayer, 2020). 
A known-fate model was applied in the program MARK 
(vers. 8.2, available from website; White and Burnham, 
1999) to estimate probabilities of conditional survival and 
total survival, as well as standard errors and 95% confi- 
dence intervals (CIs) (Schroepfer and Szedlmayer, 2006; 
Topping and Szedlmayer, 2013; Herbig and Szedlmayer, 
2016; Williams-Grove and Szedlmayer, 2016a, 2016b; 
Mudrak and Szedlmayer, 2020). Annual estimates were 
based on weekly time intervals for each year of the study 
(2013-2017). The program MARK was used to calculate 
estimates of survival based on the maximum likelihood 
binomial (Edwards, 1992). The probability of surviving a 
mortality event was determined by calculating the num- 
ber of individuals at risk of dying and the number of indi- 
viduals that survived for that time interval. Fish that 
emigrated or suffered a mortality event not under consid- 
eration were removed (right censored). For example, when 
F was estimated, all data for fish that emigrated or died as 
a result of natural causes were removed (Williams-Grove 
and Szedlmayer, 2016a). 
Instantaneous annual mortality rates were based on 
total survival (S) throughout the entire study period (241 
weeks) adjusted to 52 weeks (annual S=total S®”/*4”) for 
each type of mortality. For example, the following equa- 
tions were used for adjustments: annual F=-InS®?4) 
for fishing mortality, annual M=-InS®”*" for natural 
mortality, and annual Z=-InS®?”*” for total mortal- 
ity (Starr et al., 2005). Confidence limits for instanta- 
neous mortality rates were calculated from the 95% CIs 
estimated from the maximum-likelihood estimator of 
the survival functions at 1 year (52 weeks) (Klein and 
Moeschberger, 2003; Topping and Szedlmayer, 2013; 
Williams-Grove and Szedlmayer, 2016a; Mudrak and 
Szedlmayer, 2020). The reported sample sizes for the 
mortality estimates were the number of tagged gray 
triggerfish at risk during the time interval under analy- 
sis (Klein and Moeschberger, 2003). 
Results 
Mean fish size was 395 mm FL (standard deviation 52) 
and ranged from 276 to 535 mm FL. Most tagged gray 
triggerfish (86%) were above the legal sport and com- 
mercial size limit (356 mm FL). Five fish (9%) left their 
