15 
NOAA 
National Marine 
Fisheries Service 
Abstract— Underwater video sam- 
pling has become a common ap- 
proach to index fish abundance and 
diversity, but little has been pub- 
lished on determining how much 
video to read. We used video data 
collected over a period of 6 years in 
the Gulf of Mexico to examine how 
the number of video frames read af- 
fects accuracy and precision of fish 
counts and estimates of species rich- 
ness. To examine fish counts, we fo- 
cused on case studies of red snapper 
( Lutjanus campechanus), vermilion 
snapper ( Rhomboplites aurorubens), 
and scamp (Mycteroperca phenax). 
Using a bootstrap framework, we 
found that fish counts were unbi- 
ased when at least 5 of 1201 video 
frames within a 20-min video were 
read. The relative patterns of coeffi- 
cients of variation (CVs) were nearly 
identical among species and declined 
as an inverse power function. Initial 
decreases in CVs were rapid as the 
number of frames read increased 
from 1 to 50. However, subsequent 
declines were modest, decreasing 
only by ~50% when the number of 
frames read increased by 300%. Es- 
timated species richness increased 
asymptotically as the number of 
frames read increased from 25 to 
200 frames, and reading 50 frames 
documented 86% of the species ob- 
served across all 1201 frames. Last- 
ly, we used a generalized additive 
model to show that the most likely 
species to be missed were fast-swim- 
ming fishes that are solitary or form 
relatively small schools. Our results 
indicate that the most efficient use 
of resources (i.e. , maximum informa- 
tion gained at the lowest cost) would 
be to read -50 frames from each 
video. 
Manuscript submitted 30 July 2013. 
Manuscript accepted 13 November 2014. 
Fish. Bull. 113:15-26 (2015). 
doi: 10.7755/FB.113.1.2 
The views and opinions expressed or 
implied in this article are those of the 
author (or authors) and do not necessarily 
reflect the position of the National 
Marine Fisheries Service, NOAA. 
Fishery Bulletin 
fv established 1881 •<?. 
Spencer F. Baird 
First U S. Commissioner 
of Fisheries and founder 
of Fishery Bulletin 
Estimating relative abundance and species 
richness from wide© surveys ©f reef fishes 
Nathan M. Bacheler (contact author) 
Kyle W. Shertzer 
Email address for contact author: nate.bacheler@noaa.gov 
Beaufort Laboratory 
Southeast Fisheries Science Center 
National Marine Fisheries Service, NOAA 
101 Pivers Island Road 
Beaufort, North Carolina 28516 
Underwater video sampling has be- 
come a ubiquitous approach around 
the world to index the abundance 
of marine fish and invertebrate spe- 
cies and to quantify marine biodiver- 
sity (see reviews by Somerton and 
Gledhill, 2005; Murphy and Jenkins, 
2010). Although numerous underwa- 
ter video approaches and techniques 
have been used to index abundance, 
many researchers now employ some 
version of a stationary point-count 
method with baited remote underwa- 
ter video stations (BRUVS) (Willis et 
ah, 2000; Cappo et al., 2004). BRU- 
VS sampling has many advantages; 
1) it is nonextractive and, therefore, 
preferred in no-take areas, 2) is less 
size- or species-selective than other 
baited gears, 3) can sample deeper 
waters more easily than diver sur- 
veys and do so at lower costs than 
can autonomous underwater vehi- 
cles, 4) provides a permanent record 
available to be reviewed for accuracy 
by multiple readers, and 5) can cap- 
ture habitat characteristics of a sur- 
vey site and behavioral interactions 
among species (Silveira et ah, 2003; 
Wells et al., 2008; Langlois et ah, 
2010; Bacheler et al., 2013). 
Nearly all BRUVS studies now use 
an approach called MinCount ( or MaxN 
or MaxNo) to index the number of indi- 
viduals of various species present at a 
site (Ellis and DeMartini, 1995; Mur- 
phy and Jenkins, 2010). MinCount is 
defined as the maximum number of 
individuals (of each species) present 
in a single frame during a viewing 
interval, and this approach is popu- 
lar because it provides a conservative 
estimate of the number of individuals 
of each species present at a site (Ellis 
and DeMartini, 1995; Willis and Bab- 
cock, 2000; Merritt et al, 2011). How- 
ever, MinCount may be nonlinearly 
related to actual abundance because 
it measures a smaller and smaller 
proportion of individuals present at 
a site as abundance increases (Conn, 
2011; Schobernd et al., 2014). Instead, 
Conn (2011) proposed an alternative 
approach, MeanCount, which is calcu- 
lated as the mean number of individu- 
als observed in a series of frames over 
a viewing interval. Schobernd et al. 
(2014) found that MeanCount tracked 
true abundance linearly with levels of 
precision similar to that of MinCount. 
A linear relationship is highly desir- 
able for developing indices of abun- 
dance in stock assessment models 
(Kimura and Somerton, 2006). 
A logical next step in the devel- 
opment of the MeanCount approach 
for indexing fish abundance, as well 
as in estimating species richness, is 
to determine the optimal number 
of frames to be read over a given 
time interval. Previous studies have 
shown a strong relationship between 
