Content uploaded by April E. Hall (née Boaden)
Author content
All content in this area was uploaded by April E. Hall (née Boaden) on Apr 30, 2021
Content may be subject to copyright.
ORIGINAL RESEARCH
Partially protected areas as a management tool on inshore
reefs
April E. Hall .Darren S. Cameron .Michael J. Kingsford
Received: 29 June 2020 / Accepted: 3 April 2021
ÓThe Author(s), under exclusive licence to Springer Nature Switzerland AG 2021
Abstract Partially Protected Areas (PPAs) are a
widely-used management tool, yet comparatively little
is known about their effectiveness compared to more
commonly studied No-Take Marine Reserves
(NTMRs). Here, we examine the efficacy of two kinds
of PPAs (with and without spearfishing) within the
Great Barrier Reef Marine Park (GBRMP) that are
subject to a range of fishing limitations, and assess
their utility as a marine park zoning and fisheries
management tool. Fish abundance, size, and habitat
composition were compared inside PPAs and NTMRs
on inshore reefs of the central GBR. Fish abundances
were lower inside PPAs relative to adjacent NTMRs
for primary fishing targets, with no detectable effects
for secondary targets and non-targets, or for species
richness. Fish assemblages differed amongst zones,
but these variations were minor compared to regional
variations in species composition. Partially Protected
Areas supported 46%–69% of the relative abundance
of total primary targets compared to adjacent NTMRs,
with no evident increase in abundance in zones where
spearfishing was prohibited. There were no reductions
in the size of two key target species: coral trout
(Plectropomus spp.) and stripey snapper (Lutjanus
carponotatus) inside PPAs, and only stripey snapper
had significant reductions in abundance inside PPAs
compared to NTMRS. Habitat and biophysical char-
acteristics (especially topographic complexity) were
strong drivers of fish abundance, but the relative
influence of zone was greater for target species
compared to non-targets. This study provides novel
data on PPAs and highlights their utility as a spatial
management tool in contributing to conservation and
fisheries management goals.
Keywords Conservation Fish assemblages
Fisheries Great Barrier Reef Marine Protected
Area Zoning
Introduction
Marine Protected Areas (MPAs) are a ubiquitous
management tool for the conservation of marine
ecosystems (Edgar et al. 2014; Gaines et al. 2010).
Globally, most MPA networks consist of a range of
zones which vary in their level of protection, including
No-Take Marine Reserves (NTMRs), Partially
Supplementary Information The online version contains
supplementary material available at https://doi.org/10.1007/
s11160-021-09654-y.
A. E. Hall (&)M. J. Kingsford
College of Science and Engineering, ARC Centre of
Excellence for Coral Reef Studies, James Cook
University, Townsville, QLD, Australia
e-mail: april.hall@jcu.edu.au
D. S. Cameron
Great Barrier Reef Marine Park Authority, Townsville,
QLD, Australia
123
Rev Fish Biol Fisheries
https://doi.org/10.1007/s11160-021-09654-y(0123456789().,-volV)(0123456789().,-volV)
Protected Areas (PPAs), and open (less regulated)
fishing zones (Ban et al 2014; Denny and Babcock
2004; Zupan et al. 2018). There is great variation in the
management of fishing activities within zones, partic-
ularly for PPAs. Implementation of NTMRs, where
fishing is prohibited, can be polarising in multi-
stakeholder systems, making their widespread imple-
mentation politically fraught (Lester and Halpern
2008; Zupan et al. 2018). Partially Protected Areas,
which allow for limited fishing, are often considered
desirable as an alternative or complementary zone
type, representing a ‘‘compromise’’ solution that
allows access to marine resources with the anticipation
of a concomitant conservation benefit (Fox et al. 2012;
Sciberras et al. 2015). However, it is critical that their
performance is quantified and reviewed, to assess
whether partial protection measures are meeting
intended conservation and fishery management goals.
The majority of research on MPAs is focused on
understanding the contribution of NTMRs to conser-
vation and fisheries management. Effectively enforced
NTMRs have been clearly demonstrated to provide a
range of ecological benefits. They provide a sanctuary
for target species and their associated habitats, and this
can have flow on ecological benefits that influence
entire ecosystems (Boaden and Kingsford 2015; Edgar
et al. 2014; Hughes et al. 2007; McCook et al. 2010).
Furthermore, NTMRs can directly benefit fisheries, by
exporting of larvae, or movement of adult fishes from
NTMRs to adjacent fished zones (Gell and Roberts
2003; Halpern et al. 2009; Harrison et al. 2012; Russ
et al. 2004). There are, however, comparatively few
studies that have asked these questions of PPAs,
despite their prevalence as part of MPA systems, and
their potential as effective tools contributing to
fisheries management (Ban et al. 2014). There is also
disagreement in the outcomes of those studies that
have considered PPAs. Two of the three reviews on the
effectiveness of PPAs, have concluded that PPAs can
provide significant benefits over less regulated/open
fishing areas (Sciberras et al. 2015; Zupan et al. 2018),
whereas the third (Lester and Halpern 2008) reported
no overall conservation benefit of PPAs. Likewise,
individual studies have had inconsistent outcomes,
with some reporting a positive effect of PPAs com-
pared to open fishing zones (e.g. Boaden and Kings-
ford 2015; Bobiles and Nakamura 2019), and others
indicating no significant benefit from the implemented
partial protection measures (e.g. Coleman et al. 2013;
Malcolm et al. 2018). This disparity in outcomes has
occurred even when comparing the same species
amongst MPAs (e.g. for snapper, Chrysophyrys aura-
tus, Denny and Babcock 2004; Harasti et al. 2018).
These inconsistencies may be due to variation in the
overall effective fishing effort amongst PPAs, and
highlights the need for region-specific data to inform
management of PPAs within individual MPAs.
Despite the Great Barrier Reef Marine Park
(GBRMP) being one of the world’s largest and most
comprehensively studied MPA networks, little is
known about the contribution of PPAs to its conser-
vation and management. The GBRMP is a multiple
use marine park, comprising a range of zone types
including NTMRs, PPAs, and less protected open
fishing zones, where most ecologically sustainable
fishing methods are allowed. Fishing is prohibited
inside NTMRs, whereas PPAs aim to limit effective
fishing effort through gear prohibitions and restric-
tions, whilst allowing for limited line fishing, trolling,
and bait netting for fish species. Spearfishing is also
allowed inside regular PPAs, called Conservation Park
Zones (CPZs), but is prohibited inside designated
Public Appreciation Special Management Areas,
referred to as Special Management Areas (SMAs)
from hereon (Day et al. 2019; GBRMPA 2003), and
which overlay CPZs. The majority of fishing occurring
within CPZs and SMAs is by recreational fishers,
although some CPZs also support regionally important
commercial trolling, line-fishing and bait-netting
activities. These zones are often situated close to the
mainland shore, recognising ease of access by small
fishing vessels (GBRMPA 2003). Inshore reefs in such
areas are often highly turbid environments, and in
some regions may be habitat for estuarine crocodiles,
making monitoring using underwater surveys on
SCUBA problematic (Bradley et al. 2017). There is
comparatively little known about the effects of zoning
on fish assemblages in these coastal and nearshore
areas of the GBRMP, despite the fishing that occurs in
them (Department of Agriculture 2020; Webley et al.
2015).
The success of NTMRs inside the GBRMP has
been well documented, with clear benefits for targeted
species and associated ecosystem processes. No– take
Marine Reserves can positively benefit ecosystems by
supporting an increased abundance and size of target
species (Boaden and Kingsford 2015; Emslie et al.
2015; McCook et al. 2010; Williamson et al.
123
Rev Fish Biol Fisheries
2019,2004), however, it is notable that the majority of
prior studies have either excluded PPAs from com-
parisons, or have pooled PPAs together with less
restricted open fishing zones (but see Boaden and
Kingsford 2015; Frisch et al. 2012 as exceptions).
Since these less restricted zones allow a greater range
of fishing activities, pooling them with PPAs may
mask effects of partial protection. There is, therefore, a
need for empirical studies that quantify the effects of
PPAs within the GBRMP; and evaluate their utility as
a zoning and fisheries management tool.
The objective of our study was to examine the
effectiveness of PPAs as a management tool in two
inshore regions of the central Great Barrier Reef
(GBR): namely the Hinchinbrook and Dunk Island
regions. We hypothesised that fishing inside PPAs
would affect the abundance and/ or size of targeted
fishes, but these effects would be lesser inside SMAs
(compared to CPZs), due to prohibition of spearfish-
ing. Given this, we predicted that targeted species
would be most strongly influenced by zoning, whereas
for non-target species, variations in habitat and
biophysical characteristics would have a stronger
effect. To address this, we assessed the effects of
PPAs on targeted and non-target fish species and
groups, and evaluated the relative importance of
zoning as well as habitat and biophysical
characteristics.
The specific aims of the study were to:
1. Determine the relative efficacy of PPAs compared
to NTMRs by comparing the abundance and size
of fisheries targets amongst management zones,
2. Examine the differential response of targeted and
non-target fish species and groups to zoning in the
region,
3. Determine whether prohibition of spearfishing
inside SMAs provides improved conservation
benefits,
4. Examine the influence of zoning on the species
composition of fish communities,
5. Determine the relative influence of zoning, habi-
tat, and biophysical characteristics in driving
abundances of fish species and groups.
Methods
Great Barrier Reef Marine Park
The GBRMP is a multiple use marine park, with
different zone types allowing varying levels and types
of fishing activities. Approximately one third of the
GBRMP is zoned as NTMRs; Marine National Park or
‘‘green’’ zones, where all forms of fishing are prohib-
ited (GBRMPA 2003). Less restrictive open fishing
zones comprise Habitat Protection Zones (trawling
prohibited) or ‘‘dark blue’’ zones, or General Use
(trawling allowed) or ‘‘light blue’’ zones. Partially
protected areas including CPZs or ‘‘yellow’’ zones,
have intermediate protection between NTMRs and
open fishing zones, and allow for limited fishing, with
gear and effort restrictions applying to both recre-
ational and commercial fishers. Inside CPZs limited
line fishing is allowed, with one hook/liner per fisher
(except for trolling), but large-mesh netting, trawling,
and dive-based harvest and collection fishing is
prohibited (GBRMPA 2003). In all fished zones of
the GBRMP, there are bag and size limits that apply,
and these are tailored to the management of individual
species or groups of species (Day et al. 2019). In this
study, we compare the two kinds of PPAs that occur
within the Marine Park: Conservation Park Zones
(CPZs) and Special Management Areas (SMAs). Both
zone types have the limitations on fishing described
for CPZs above, however CPZs allow for spearfishing
whereas spearfishing is prohibited inside SMAs
(GBRMPA 2003). Both zones are coloured yellow
on zoning maps, and the SMAs are indicated by a pink
hashed line surrounding the designated SMA region
(Fig. 1).
Study area and sampling design
The study area comprised two regions in the central
GBR: the Hinchinbrook and Dunk Island regions,
which are important recreational fishing areas with a
large coverage of PPAs (Fig. 1). Both regions include
a cluster of inshore islands, with a range of shallow
habitats including rocky reefs, fringing coral reefs,
seagrass beds, and macroalgae beds. Coral reefs and
associated biota are explicit conservation values
recognized in the Hinchinbrook Plan of Management
which encompasses the study area (Day et al. 2019).
Importantly, the study area was ultimately chosen
123
Rev Fish Biol Fisheries
!(
!(
!(
!(
!(
!(
!(
!(
!(
Dunk Island
Bedarra Island
Thorpe Island
Purtaboi Island
Coombe Island
Smith
Island
146°12'0"E
146°12'0"E
146°10'0"E
146°10'0"E
146°8'0"E
146°8'0"E
17°56'0"S
17°56'0"S
17°58'0"S
17°58'0"S
18°0'0"S
18°0'0"S
18°2'0"S
18°2'0"S
!(
!(!(
!(
!(
!(
!(
!(
Hinchinbrook Island
Goold Island
Brook
Islands
Agnes
Island
Garden Island
Eva Island
Queensland
146°20'0"E
146°20'0"E
146°10'0"E
146°10'0"E
146°0'0"E
146°0'0"E
18°10'0"S
18°10'0"S
18°20'0"S
18°20'0"S
LEGEND
BRUVS survey sites
!(
Dunk Island Region
!(
Hinchinbrook Island Region
Zoning in map extent
General Use Zone
Conservation Park Zone
No–Take Marine Reserve
Special Management Area
Great Barrier Reef
Marine Park boundary
Indicative Reef boundary
Mainland and Islands
Queensland
MAPPED
AREA
a
b
02
Kms
SDC200 408
This map is indicative only
Names are no t necessarily aut horitative
05
Kms
123
Rev Fish Biol Fisheries
because both regions contain a number of NTMRs and
CPZs, however SMAs only occur in the Dunk Island
region (Fig. 1). The location of these SMAs overlaps
with existing state fisheries legislation, such that
spearfishing was already prohibited in parts of the
current SMAs prior to the implementation of the
current (2004) zoning Plan. Similar to many coastal
regions of the GBRMP, neither region contains
comparable habitats in open fishing zones, so we
could not survey these zone types for comparison.
Instead, we focused on comparisons of PPAs with
adjacent NTMRs within the study area. To enable
comparisons amongst zones, we used a nested hierar-
chical design, with zones nested within each region,
sites nested within zones, and replicate surveys within
each site (Fig. 1). The arrangement of zones within
each region allowed for a robust sampling design
without spatial segregation by treatment (Hurlbert
1984), and sites were chosen within each zone to
optimally encompass the range of broad habitat types
present in each region. This design allowed us to
assess zoning trends within the study area, in the
context of other key drivers such as habitat and
biophysical characteristics.
Baited Remote Underwater Video Stations
(BRUVS) were used to survey fishes at each site; a
total of 204 replicate BRUVS drops were deployed.
The sampling design differed slightly between
regions, with 3 sites/zone (total = 9 sites for 3
categories of zone) and 4 sites/zone (total = 8 sites
for 2 categories of zone) for the Hinchinbrook and
Dunk Island regions respectively (Fig. 1). Surveys
were carried out in each region in two field trips during
which six replicate BRUVS were deployed at each site
(total = 12 replicates/site). Surveys were conducted in
neap tide periods during the 2018 dry season of
tropical north Queensland (May–October), to ensure
maximal water visibility and minimize tidal effects
(August/October and June/September for the Hinch-
inbrook and Dunk Island Regions respectively). Each
BRUVS had a deployed soak time of 60 min; repli-
cates were placed haphazardly along shallow reef
habitats or island edges, and the depth recorded. To
ensure sampling independence, each of the BRUVS
drops were located 250 m–350 m apart during field
trips, and at sites with limited spatial extent, adequate
replication with this separation was achieved by
alternating the placement of replicates by field trip.
In this manner, each of the twelve BRUVS drops per
site were therefore considered an independent
replicate.
BRUVS apparatus and deployment
The BRUVS apparatus consisted of four simple
BRUVS, and two stereo-BRUVS, all of which were
deployed at each site, giving a mix of stereo and single
view videos for each site. The BRUVS apparatus and
deployment methods used are described in detail by
Stowar et al. (2008). Briefly, the simple BRUVS
consisted of a galvanized steel frame, upon which an
underwater camera housing and bait arm were
attached. The camera housing was secured to the
frame and oriented slightly downward for an optimal
view of the bait bag, which was located one metre
away at the end of a PVC bait arm. The mesh bait bag
contained approximately 1 kg of crushed pilchards
(Sardinops sp.). A rope with attached surface floats
enabled the BRUVS to be deployed and retrieved from
the surface. The stereo-BRUVS apparatus was similar
to that described above, except that it contained two
cameras mounted at either end of a wider steel frame,
to capture stereo-video that enabled lengths of fish to
be measured. Sony handicams were used for simple
BRUVS, and GoPro Hero 4 cameras were used for
stereo-BRUVS. For both setups, camera settings were
adjusted for the maximal focal range and optimal
video resolution.
Video analysis
Video footage was analysed for fish abundance using a
custom interface developed by the Australian Institute
of Marine Science, ‘‘BRUVS 3.0.mbd’’, as described
in Cappo et al. (2011). Surveys included all species
that were large and conspicuous enough to accurately
count and identify (Online resource 1: Table S1). We
excluded small fishes such as wrasses \5 cm, dam-
selfishes and gobies, that could not be accurately
recorded. To ensure accuracy of data, replicates were
excluded from analysis if the visibility was less than
1.5 m, or if the BRUVS were positioned such that the
bFig. 1 Map of Baited Remote Underwater Video Stations
(BRUVS) survey sites in athe Dunk Island Region, and bthe
Hinchinbrook Island Region. An array of 12 replicate BRUVS
deployments were placed per site. Colours indicate the Great
Barrier Reef Marine Park management zones
123
Rev Fish Biol Fisheries
field of view was significantly impaired. The resulting
dataset for video analysis was reduced to 160 of 204
replicates. Abundance data were recorded as MaxN,
the maximum number of individuals of a given species
observed in a single video frame during the 60 min of
footage. MaxN is a commonly used estimate of
abundance for BRUVS data, and its use prevents over
counting of fishes that may repeatedly move in and out
of the field of view (Cappo et al. 2007,2004). All
fishes were identified to species level where possible,
and to genus where species could not be distinguished.
Coral trout (Plectropomus spp.) were pooled into a
single group, due to the difficulty in distinguishing
species in the video footage, and since they are very
similar species that are all considered primary fishing
targets in the region (Emslie et al. 2015; Webley et al.
2015). The two most abundant coral trout species on
inshore reefs (P. maculatus, and P. leopardus) would
have comprised the majority of individuals, and are
known to hybridise (Frisch and Herwenden 2006). A
third species P. laevis also occurs but is rare on inshore
reefs (Emslie et al. 2017). Given this, the three species
are commonly pooled together for assessments (e.g.
Emslie et al. 2015; Russ et al. 2008), consistent with
our approach.
To compare the size of key fisheries targets
amongst zones, we measured the length of Plectropo-
mus spp. and Lutjanus carponotatus from stereo-
BRUVS footage at the time of MaxN using the
measurement software ‘‘Eventmeasure
TM
(www.
seagis.com.au). Measurements were made from
snout to tail fork (fork length), as specified for Stereo-
BRUVS data, but were converted to total length (snout
to end of tail) using conversion formulae (Kingsford
2009; Mapstone et al. 2004), in order to align with
defined legal fisheries minimum size limits. Due to
low visibility and/ or the sub-optimal positioning of
fishes in some videos, not all individuals could be
accurately measured, which limited the sample size to
47 Plectropomus spp. and 48 L. carponotatus. There
were a number of additional predictor variables
recorded from the BRUVS footage, including visi-
bility (m), topographic complexity scores (called
‘‘complexity’’ hereafter), and the percentage cover of
major substratum types (Table 1). These habitat and
biophysical predictors were used in the subsequent
statistical analyses, and provided an understanding of
how zoning interacted with these processes to influ-
ence fish assemblages.
Statistical analysis
Fish species were categorized as primary targets,
secondary targets, or non-targets, based on their
importance for recreational and/ or commercial fish-
eries. Primary targets were defined as those species
that were highly likely to be sought after and retained.
Secondary targets were of lower value as fisheries
targets, but considered moderately likely to be
retained if caught (Online Resource 1: Table S1).
Assignation of species into these groups was based on
the available literature, discussions with relevant
experts, and an understanding that primarily recre-
ational fishers would target the inshore reef habitat that
occurs within the study area (Webley et al. 2015).
Species were only included as primary or secondary
targets if they are known to grow larger than the
minimum size limit, and could therefore be retained if
caught. These groups were used for subsequent
univariate analysis. Two key fisheries target taxa
(Plectropomus spp. and L. carponotatus) were also
analysed separately, due to their abundance and
importance as key fisheries targets (Williamson et al.
2004).
Generalised Linear Mixed Models (GLMM) were
used to assess how zoning interacts with benthic
habitats and biophysical characteristics. The MaxN
data were over-dispersed or highly skewed, so a
negative binomial distribution was used (Bolker et al.
2009). To best integrate benthic data into the analysis,
we performed a Principal Component Analysis (PCA)
on the benthic data, and used the resulting PC scores as
continuous variables; a scree plot indicated that the
first three PC axes were the most informative, and
explained over 80% of the variation (Espinoza et al.
2014). We predicted that a combination of depth,
visibility, complexity, and the PC axes, would influ-
ence response variables in conjunction with zoning
(Table 1). We sought to produce the most parsimo-
nious model using a combination of these predictors.
To do this, we compared models with different
combinations of predictors, starting with more com-
plex models and reducing down to simpler models,
using corrected Akaike Information Criteria (AICc) as
a basis for model selection (Bolker et al. 2009). The
fixed factor of zone and the random factors of region
and site were included in all models. The significance
of model variables was assessed using p value
estimates from model outputs (p \0.05), along with
123
Rev Fish Biol Fisheries
visual examination of the modelled coefficients and
confidence intervals for the factor ‘‘zone’’. Modelled
effect sizes were calculated for response variables
where ‘‘zone’’ was considered a significant model
predictor. All GLMMs were fitted in R using the
‘lme40package for models, and the ‘effects’ package
for effect sizes (R Core Team 2018).
Boosted Regression Trees (BRTs) were used to
explore the key drivers of fish abundance and species
richness. Boosted regression trees are machine
Table 1 Predictor variables used in statistical analyses, and their overall observed range, mean, and definitions. Values for means
and Standard Error of Mean (SEM) are based on total pooled replicates within each region and zone
Variable Type Range Mean (±SEM) per region and zone Estimation method and definition
Hinchinbrook Dunk
NTMR CPZ NTMR SMA CPZ
Depth (m) Continuous 2–17 5.3
(±0.5)
7.8
(±0.6)
5.9
(±0.5)
6.5
(±0.4)
5.3
(±0.5)
Water depth from surface to
seafloor at location of BRUVS
deployment
Visibility (m) Continuous 1.5–5 2.4
(±0.1)
2.3
(±0.1)
2.5
(±0.2)
2.4
(±0.1)
2.7
(±0.2)
Estimated during video analysis.
Defined as the horizontal
distance that could be seen in the
FOV of each video. Estimated
using the bait arm as a reference,
and calibrated using
measurements in the software
Eventmeasure
Topographic
complexity
score
Categorical 1–4 2.4
(±0.2)
2.4
(±0.1)
2.2
(±0.1)
2.4
(±0.2)
2.3
(±0.2)
Estimated using snapshots of the
benthos from each video based
on the habitat structure in the
FOV. Each replicate was
assigned an overall complexity
score, ranging from flat sandy
seafloors with no structure (score
1), to highly complex habitats
with high structural rugosity
(score 4)
% coral cover Continuous 0–80 22.1
(±3.9)
20.5
(±3.9)
23.8
(±2.9)
30.0
(±5.6)
22.2
(±4.8)
Estimated during video analysis,
by visually estimating the %
cover of each substrate
component. % coral
cover = combined percentage of
live hard and live soft corals, %
algae = all visible algae (mostly
macroalgae of the genus
Sargassum), % bare = sand or
rubble, and %
bedrock = bedrock with no
substrate biota
% algae 0–100 21.5
(±4.5)
16.6
(±3.4)
33.2
(±5.6)
18.75
(±5.1)
41.2
(±7.8)
% bare 0–100 44.1
(±5.6)
40.3
(±5.5)
28.9
(±5.1)
36.6
(±6.4)
15.7
(±3.8)
% bedrock 0–90 11.3
(±3.7)
21.4
(±4.4)
14.1
(±5.6)
14.6
(±5.1)
18.6
(±5.6)
PC1,PC2,
PC3
Continuous NA PC scores, derived from a PCA on
the % cover of substrate
components. Provided three
continuous predictor variables
that describe the overall
variation in substrate
components
FOV Field of view, NTMR No-take Marine Reserve, CPZ Conservation Park Zone, SMA Special Management Area
123
Rev Fish Biol Fisheries
learning models that fit a succession of regression
trees, ‘‘ learning’’ relevant fractions of the data
features in each tree and passing this information to
successive trees (Elith et al. 2008). Boosted Regres-
sion Trees are well suited to exploring complex
relationships, and provide estimates of the relative
influence of each variable on the response variables.
To better understand the nature of specific benthic
drivers, we used the percentage cover of coral, algae,
bedrock, and bare substrate as predictor variables,
rather than using PC scores as per the GLMMs.
Boosted Regression Tree models were fitted using the
‘gbm’ package in R (R Core Team 2018), with the
following parameters: learning rate (contribution of
each tree to the final model) = 0.001, bag fraction
(proportion of data used in each step) = 0.5, and tree
complexity (maximum nodes per tree) = 5 (Elith et al.
2008). The number of trees produced by each model
varied as a function of the above parameters, and was
between 1600 and 4000.
The length distributions of Plectropomus spp. and
L. carponotatus were compared between NTMRs and
PPAs using kernel density estimates (KDEs) in R
using the ‘kde.compare’ function in the ‘sm’ package
(R Core Team 2018). Samples were pooled amongst
the two PPA zones to increase replication. Using the
method described by Langlois et al. (2015), length
frequency distributions were compared between the
PPA and the NTMR zones based on a null model of no
difference and a permutation test. The permutation test
is robust to small sizes, and differences in sample sizes
amongst treatments (Langlois et al. 2015). The KDE
method compares both the shape and location of
length frequency distributions, however, in our case
all tests for shape were found to be not significant
(p [0.05). As such, all subsequent tests compared the
location (i.e. mean) of length distributions only, and
only graphs relating to the test for location are shown
(Langlois et al. 2015; Parker et al. 2016). The
modelled KDE curves for each zone were compared
to the permuted null model visually from the resultant
graphs, and statistically using the permutation test
(Langlois et al. 2015; Parker et al. 2016).
To examine the influence of zoning on species
assemblages, we used Canonical Analysis of Principle
coordinates (CAP), along with a Permutational Mul-
tivariate Analysis of Variance (PERMANOVA), using
the PRIMER statistical package (Anderson 2001;
Anderson and Willis 2003). Data were dispersion
weighted and log transformed to reduce the influence
of highly abundant species, and the Bray Curtis
similarity matrix was used. The CAP method is
described in detail in Anderson and Willis (2003),
and uses an ordination process constrained by a priori
classifications (in this case, ‘‘zone’’). Differences
amongst zones were formally tested using a PERMA-
NOVA, and the components of variation were used to
evaluate the relative influence of region and zone
(within region) on fish assemblages; significant terms
were investigated using posteriori pairwise compar-
isons (Anderson 2017). Percentage of Similarity
(SIMPER) was used to determine the contribution of
species to dissimilarities amongst zones (Clarke et al.
2014).
Results
A total of 180 taxa from 29 fish families were recorded
from the BRUVS drops (Online Resource 1:
Table S1). Of this, 140 were identified to species,
and 40 identified to genera. This included a variable
suite of species, ranging from site-attached coral
associated families such as chaetodontids, to more
transient, roving groups such as carangids and elas-
mobranchs. The most abundant families were school-
ing species of Caesionidae and Lutjanidae. Many
species were rare, and/ or patchy; 37 species were
recorded as a single individual, and the majority of
species (163) occurred in less than 20% of replicates
(Online Resource 1: Table S1). Ten species that were
classified as primary targets, from the Labridae,
Lutjanidae, and Serranidae families, were recorded.
Of these, the most commonly encountered were
stripey snapper (Lutjanus carponotatus), and coral
trout (Plectropomus spp.). Both of these taxa had
relatively low abundance by replicate (MaxN \5),
but high total abundances due to their presence in a
large proportion of replicates (*64 and *66% for
L. carponotatus and Plectropomus spp. respectively;
Online Resource 1: Table S1). Fifteen species were
recorded that were considered secondary targets.
These tended to be smaller and/ or less desirable
lutjanids and lethrinids, as well as tuskfish (Labridae:
Choerodon spp.) and carangids. The non-target
species assemblage was diverse (156 species), how-
ever only 13 species were recorded in more than 20%
of replicates, and 36 species were recorded as a single
123
Rev Fish Biol Fisheries
individual. Nine species of sharks and three species of
rays were recorded. The most abundant of these was
the blacktip reef shark (Carcharhinus melanopterus),
which occurred in 15% of replicates; most other
elasmobranch species were very rare.
Habitat and biophysical parameters varied within
the study area, however there was little difference in
the overall composition or structure of habitats
amongst zones. Habitats in both regions consisted of
a mosaic of coral, macroalgae, and bedrock, with
patches of bare sand/rubble beds (Table 1). There was
variation in the composition of benthic substrates
amongst sites and replicates, with some replicates
having quite high coral cover (up to 80% within the
field of view), and others dominated by algae, bedrock,
or sand/ rubble beds (Table 1). Topographic complex-
ity, depth, and visibility did not vary in a consistent
manner by zone. Mean coral cover ranged from &20
to 30%, with the highest mean value in SMAs, which
only occurred in the Dunk Island region. Of the three
zones in the Dunk Island Region, SMAs had the lowest
mean percentage of algae and the greatest mean
percentage of bare (sand/rubble), however, similar
values for the cover of these substrates were found in
both zones in the Hinchinbrook Island region. Accord-
ingly, there was no evidence that zoning comparisons
(for fishes) would be confounded by differences in
habitat representation (Table 1).
Influence of zoning on the abundance of fish
species and groups
The effect of zone on pooled primary fisheries targets,
and Lutjanus carponotatus was significant, with little
difference amongst zones for other fish groups or
species (Fig. 2). For key target species Plectropomus
spp., and L. carponotatus, there was a trend for
abundance to be lower inside SMAs compared to
CPZs, even in instances where zoning was not a
significant model contributor (Figs. 2and 3). The most
parsimonious model to describe abundance of key fish
species and groups was ‘‘Re-
sponse *zone ?depth ?visibility ?complexity’’.
Model coefficients and significance values for all
models are listed in Table 2.
For primary fishing targets, abundance was greater
inside NTMRs compared to both PPAs (Fig. 2and 3).
Both SMAs and CPZs had significantly lower abun-
dances compared to NTMRs, based on model
coefficients and confidence intervals (Fig. 2and
Table 2). There was no evidence that prohibition of
spearfishing inside SMAs resulted in greater abun-
dance of primary targets; rather there was a trend for
abundance of primary fishing targets to be lower inside
SMAs compared to CPZs; modelled effect sizes
predicted that NTMRs supported *1.4–2.2 9the
abundance of primary fishing targets compared to
CPZs and SMAs respectively (Figs. 2and 3). Primary
targets also had a significant positive relationship with
complexity (Table 2), however, the marginal effects
plot illustrated that the trend of NTMR [CPZ [
SMA for abundance was consistent across complex-
ity scores (Fig. 4). No-take marine reserves had
greater variation in abundance of primary targets
among replicate BRUVS compared to other zones;
this was largely due to clusters of schooling target
species. Particularly high numbers of primary targets
Fig. 2 Summary coefficients plot, showing the effect of the
factor zone in the GLMM for each fish species/ group. The zero
line indicates the modelled data from No Take Marine Reserves
(NTMRs), such that lines that are non-overlapping with zero
indicate where zoning was a significant model contributor.
CPZs = Conservation Park Zones, SMAs = Special Manage-
ment Areas
123
Rev Fish Biol Fisheries
occurred where schools comprising both crimson
snapper (Lutjanus erythopterus) and saddletail snap-
per (L. malabaricus) were encountered inside
NTMRs.
There was a trend for lower abundance for both
Plectropomus spp. and L. carponotatus inside fished
zones, however zone was only a significant model
predictor for L. carponotatus (Figs. 2and 3, Table 2).
0
1
2
3
4
5
0
5
10
15
20
0
1
2
3
4
5
6
Primary targets Plectropomus spp. L.carponotatus
0
2
4
6
8
10
12
22
23
24
0
1
2
3
4
5
6
0
1
2
3
4
5
6
Abundance (MaxN)
Hinchinbrook Island Region
Dunk Island Region
Abundance (MaxN)
No-take Marine Reserve Special Management Area Conservation Park Zone
Fig. 3 Boxplots (Tukey) showing the abundance (MaxN) of
summed primary targets, Plectropomus spp., and Lutjanus
carponotatus among zones in the Hinchinbrook and Dunk Island
Regions. Raw abundance values are shown as open circles; these
are MaxN numbers from each replicate BRUVS drop for the fish
species/ group
Table 2 Model coefficients for the GLMM model Response *Zone ?depth ?visibility ?complexity, and corresponding p
values (bold where p \0.05); p values for SMA and CPZ zones indicate significance values as compared to NTMRs
Fish group Zone: SMA Zone: CPZ Depth Visibility Complexity
Coefficient p Coefficient p Coefficient p Coefficient p Coefficient p
Primary targets -0.779 0.001 -0.365 0.021 0.092 0.202 0.104 0.120 0.159 0.020
Coral trout
(Plectropomus
spp.)
-0.449 0.979 -0.145 0.135 0.020 0.882 0.325 < 0.001 0.068 0.404
Stripey snapper
(Lutjanus
carponotatus)
-0.574 0.035 -0.194 0.251 -0.093 0.283 0.100 0.198 0.107 0.196
Secondary targets 0.226 0.552 -0.205 0.485 0.051 0.703 0.088 0.459 0.253 0.017
Non-targets 0.042 0.928 0.073 0.812 -0.036 0.701 0.209 0.014 0.197 0.016
Species richness -0.140 0.595 -0.003 0.988 -0.002 0.985 0.122 0.007 0.162 < 0.001
NTMR No-take Marine Reserve, SMA Special Management Area, CPZ Conservation Park Zone
123
Rev Fish Biol Fisheries
Both taxa had a trend of lowest abundances inside
SMAs, similarly to total primary targets. For coral
trout, zoning trends differed by region; with little
difference in abundance between zones in the Hinch-
inbrook Island region, and a pattern of NTMR [
CPZ [SMA in the Dunk Island region. Abundances
of coral trout tended to be greater overall in the Dunk
Island region, and variation due to the random effect of
region was more than twice as great as that of site
(within region and zone) in the GLMM model (Fig. 3).
Visibility had a strong positive effect on coral trout
abundance, and was the only significant model
predictor (Table 2). For L. carponotatus, abundances
were lower inside both PPAs compared to NTMRs,
however only SMAs had significantly lower abun-
dances compared to NTMRs (Table 2). No-take
marine reserves were predicted to sup-
port *1.2–1.8 9the abundance of L.carponotatus
compared to CPZs and SMAs respectively.
There was no evidence of a fishing effect on
secondary fishing targets, and zone was not a signif-
icant model predictor (Figs. 2and 5). Secondary
targets had a significant positive relationship with
complexity, similarly to other fish groups (Table 2).
Particularly high abundances of secondary targets
were recorded at certain sites with the presence of
schooling pelagic fishes including Carangoides gym-
nostethus and Scomberomorus sp., as well as mixed
schools of lutjanids (Lutjanus lemniscatus/ L. russelli,
Fig. 5). There was no significant effect of zone on non-
target fishes, or species richness, which both had high
site and replicate level variation, and significant
positive relationships with visibility and complexity
(Table 2). For non-target fishes, there was a trend for
higher abundances inside CPZs compared to NTMRs
in the Dunk Island region, however this was not
consistent in the Hinchinbrook Island region, and
zoning was not a significant model predictor (Fig. 5).
Both regions had a similar range of species richness;
although maximum values were greater in the Dunk
Island region (Fig. 5).
Variation in size of key fisheries targets by zone
There were significant differences in size distributions
between NTMR and combined PPA zones for Plec-
tropomus spp., but not L. carponotatus (Fig. 6). For
Plectropomus spp., a smaller mean size (300 mm TL,
n = 19) inside NTMRs compared to PPAs (369 mm
TL, n = 28) resulted in a significant difference in the
location of length frequency distributions. This is
illustrated by the KDE function for each zone extend-
ing beyond the null distribution, and supported by the
results of the permutation test (p \0.05; Fig. 6). The
proportion of Plectropomus spp. larger than the
minimum legal length of 380 mm was much greater
in PPAs (32%), compared to NTMRs (11%). For
L. carponotatus, the mean size was fairly similar
between NTMRs (229 mm TL, n = 21) and PPAs
(247 mm TL, n = 27), and the proportion of fish above
the minimum legal length of 250 mm was 29% for
both zones. There were no significant differences in
the location of KDE curves for L. carponotatus, which
fell within the modelled null distribution (Fig. 6).
Key drivers of abundance
Drivers of abundance varied greatly amongst species/
groups, and zone was a more important predictor for
targeted versus non-target fishes (Fig. 7). Zone was
the third most important predictor for primary targets
(12.9% relative influence), and the fourth ranked
predictor for L. carponotatus (9.1% relative influ-
ence), but was an unimportant predictor for secondary
targets, non-targets and species richness (Fig. 7). For
total primary targets and L. carponotatus, depth had
the strongest relative influence, followed by coral
Fig. 4 Marginal effects plot showing the predicted relationship
between abundance of primary targets and topographic com-
plexity, separated by zone. Data are predicted values based on
the model estimates, and their corresponding confidence
intervals, calculated by holding other fixed factors in the model
constant, to show the marginal effect of significant model
predictors
123
Rev Fish Biol Fisheries
0
5
10
15
20
25
30
0
2
4
6
8
10
0
10
20
30
40
50
60
70
80
Secondary targets Non-targets Species richness
0
2
4
6
8
10
12
14
16
0
20
40
60
80
100
120
140
0
5
10
15
20
25
30
35
Abundance (MaxN) Abundance (MaxN)
No-take Marine Reserve Special Management Area Conservation Park Zone
Hinchinbrook Island Region
Dunk Island Region
Fig. 5 Boxplots (Tukey) showing the abundance (MaxN) of
summed secondary targets, non-targets, and species richness
among zones in the Hinchinbrook and Dunk Island Regions.
Raw values are shown as open circles: these are MaxN or
species richness numbers from each replicate BRUVS drop for
the fish species/ group
Fig. 6 Kernal density estimates (KDE) and probability density
functions for Plectropomus spp. and L.carponotatus. Lines
represent KDE functions approximating length data from each
zone, with data from Special Management Areas (SMA) and
Conservation Park Zones (CPZ) pooled and represented
together as ‘‘PPAs’’. Grey bands represent one standard error
either side of the null model of no difference between KDEs
between zones; p values are given for the KDE permutation test
123
Rev Fish Biol Fisheries
cover (Fig. 7). Total primary targets, Plectropomus
spp., and L. carponotatus all had positive relation-
ships with coral cover, which was asymptotic at
around 20%–30% (Online Resource 1: Table S2).
Consistent with the GLMMs, zone was not an
important predictor for Plectropomus spp., but there
was a strong positive relationship with visibility
(25.7% relative influence; Fig. 7and Online Resource
1: Table S2). Complexity was a very strong positive
predictor for secondary targets (39.3% relative influ-
ence); the second and third most influential predictors
were depth and algae, which had a positive and
negative influence respectively (Fig. 7and Online
Resource 1: Table S2). For non-targets and species
richness, the most influential predictor was complex-
ity, and the relative importance of all other predictors
was quite even, apart from a markedly low importance
for zone (4–5%). This likely reflects sampled habitats
covering the range of preferred habitats for the suite of
species included in these groups, with little impact due
to zone (Fig. 7).
Variation in species assemblages by management
zone
There were significant differences in species assem-
blages amongst zones, which were driven primarily by
differences between NTMRs and SMAs (Fig. 8,
Table 3). The CAP ordination showed separation by
zone, particularly between NTMRs and SMAs; CPZs
shared properties of both zones. Overlaid bubble plots
on the CAP ordination indicated that variation in the
abundance of primary fishing targets was a large
component of the differentiation by zone. This was
supported by the vector overlays, which showed a
number of primary target species (e.g.
depth
%c
oral
zone
%a
lgae
complexity
visibility
%ba
re
location
%be
drock
0
10
20
30
40
50
Relative influence (%)
Secondary targets
Non-targets
Species richness
depth
%coral
zone
%a
lgae
complexity
visibility
%ba
re
location
%be
drock
0
10
20
30
40
Relative influence (%)
Plectropomus spp.
Primary targets
L.carponotatus
a
b
Fig. 7 Results from
Boosted Regression Trees
(BRTs), showing the
percentage relative
influence of each predictor
variable in the BRT model,
for asummed primary
fisheries targets, and key
target species Plectropomus
spp. and Lutjanus
carponotatus, and
bSecondary targets, non-
targets, and species richness.
Predictor variables on the x
axis are ordered in both
graphs by relative
importance for primary
targets, to allow comparison
amongst groups
123
Rev Fish Biol Fisheries
L. carponotatus, L. erythopterus, L. malabaricus and
Lethrinus nebulosis) strongly correlated with NTMRs
(Fig. 8). Zone and region were both significant in the
PERMANOVA, and region explained a larger com-
ponent of the modelled variation (17.7%) compared to
zone (6.7%; Table 3). Pair-wise comparisons indicated
that although assemblages differed significantly by
zone, this trend was driven only by significant
differences between NTMRs and the SMAs in the
Dunk Island region. A SIMPER comparison between
these two zones in the Dunk Island region showed the
overall 73% dissimilarity between these two zones
Fig. 8 Canonical Analysis of Principle Coordinates (CAP) of
fish assemblages in relation to management zones, by site level
centroids. awith overlaid bubble plot indicating the abundance
of primary fishing targets (site means range from 2 to 20), and
bwith overlaid vectors showing the contribution of fish species
to differences amongst zones. Only species with Pearson
correlations of [0.4 are shown on vectors; primary targets
are indicated in bold
Table 3 Results of PERMANOVA main effects and post-hoc pairwise comparison tests on fish assemblage data using site level
centroids; p (perm) values were obtained using 9999 permutations under a reduced model
Main effects tests
Source df MS p(perm) Components of variation
Estimate SD % of total
Region 1 3945.3 0.001 186.1 13.64 17.7
Zone (Region) 3 3776.5 0.016 69.89 8.36 6.7
Residual 29 22,994.0 – 792.91 28.16 75.6
Pair- wise comparisons for zone
Region Zone comparisons p(perm)
Hinchinbrook Island region NTMR versus CPZ 0.136
Dunk Island region NTMR versus CPZ 0.073
NTMR versus SMA 0.017
CPZ versus SMA 0.271
Components of variation give an indication of the contribution of each source to total variation in the permuted data; SD is an
indication of error for each estimate
NTMR No-take Marine Reserve, CPZ Conservation Park Zone, SMA Special Management Area
123
Rev Fish Biol Fisheries
was not driven strongly by any single species, rather
by a suite of species contributing incrementally. The
top three most influential species were Plectropomus
spp. and L. carponotatus, which were more abundant
on average inside NTMRs, and Scolopsis mono-
gramma (Nemipteridae), which was more abundant
on average inside SMAs. Each of these three species
singularly contributed only 6%–7% of total dissimi-
larity between zones.
Discussion
This study provides novel data on how fishing and
marine park zoning interact with a range of drivers to
influence fish communities on inshore reefs. Inshore
marine habitats are often overlooked in long term
monitoring studies, possibly due to their turbid waters
and the associated difficulties with conducting fish and
benthic surveys (Bradley et al. 2017). Such inshore
areas may have unique species assemblages, and are
subject to strong anthropogenic influence from coastal
processes, making them important systems for con-
sideration in conservation and management (Wenger
et al. 2016). Furthermore, their proximity to coastal
towns make them important areas supporting fishing
opportunities, as evidenced by the prevalence of PPAs
in coastal and near coastal areas within the GBRMP
(GBRMPA 2003; Webley et al. 2015). Consistent with
our hypothesis, our results demonstrate there are
reduced abundances of primary target fishes occurring
on inshore reefs inside two PPA zones in the study
region, which are likely caused by fishing. It is
notable, however, that moderate numbers of primary
targets were found inside PPAs and there were
negligible differences observed inside PPAs and
NTMRs for non-target species and groups. Given the
limited impacts of fishing observed in the study area
(where fishing is an important activity), there is great
potential for PPAs to contribute to conservation and
fisheries management goals.
Response of key fishing targets to zoning
Zoning effects inside PPAs were strongest and most
consistent for primary targets overall. We were unable
to make comparisons between PPAs and open fishing
zones in this study, since there are no such zones in
similar habitats within the study region. Our approach,
therefore, was to scale PPAs against NTMRs, in order
to best understand how PPAs may work in comparison
to NTMRS. Modelled effect sizes predicted that
NTMRs supported 1.4 times the abundance of primary
targets compared to CPZs, and 2.2 times the abun-
dance compared to SMAs. From this we can estimate
that PPAs in this region support 46%–69% of the
abundance of primary targets expected inside a
NTMR. Although there have been many previous
studies investigating the effects of zoning on fishes in
the GBRMP, the majority of these focus primarily on
coral trout, due to their importance to recreational
fishers (Webley et al. 2015) and high value in
commercial fisheries, where they constitute a large
proportion of the overall harvest (Emslie et al. 2015;
Frisch et al. 2012; Williamson et al. 2019,2004).
Furthermore, many studies pool together PPAs with
less restricted ‘‘open fishing’’ (Habitat Protection or
General Use) zones (Emslie et al. 2017; Williamson
et al. 2019,2004). As such, there are few comparable
studies with which we can directly compare the
magnitude of PPA zoning effects for primary fishing
targets. Comparable studies in the region have
reported a varying magnitude of zoning effects for
PPAs compared to NTMRs. Cappo et al. (2010)
reported more than five times the abundance of targets
inside NTMRs versus CPZs when comparing fishes on
shoal habitats in the Hinchinbrook and Dunk Islands
regions using BRUVS. This study was focused on
comparison of discrete deeper shoal habitats (15 m–
25 m), and the CPZ investigated was a known and
highly popular fishing location, which may explain the
larger magnitude of zoning effects reported. A study
by Boaden and Kingsford (2015) in the nearby inshore
Palm Islands area found that SMAs supported &52%
of the abundance of targeted predators found in
adjacent NTMRs, whereas open fishing zones only
supported 34% relative abundance, indicating an
intermediate effect of SMAs. Although our study did
not include open fishing zones, the range of effect
sizes we observed for PPAs (46%–69%), is similar or
above that observed by Boaden and Kingsford (2015).
Together these studies support the concept that PPAs
can have conservation benefits, and may act as a
fisheries management tool in supporting moderate
numbers of targeted fishes.
Pivotal next steps in this field of research will be
determining the geographical generality of observed
zoning trends on both inshore and offshore reefs, and
123
Rev Fish Biol Fisheries
any underlying factors that promote the success or
otherwise of PPAs as a conservation and/ or fisheries
management tool. An important consideration is non-
compliance within PPAs and adjacent NTMRs, which
could influence the detectability of zoning effects. In
the GBRMP, recent reports have raised concern
regarding non-compliance by recreational fishers,
however there is comparatively little known about
compliance inside PPAs, since research and compli-
ance and enforcement efforts are often focused on
poaching inside NTMRs (Bergseth et al. 2017; Thiault
et al. 2020; Weekers and Zahnow 2019). The ability to
detect non-compliance inside PPAs is also likely to be
lower, given the more subtle nature of fishing restric-
tions compared to NTMRs. Given this level of
uncertainty, it is possible that non-compliance may
have influenced our results, however there is no public
information available to suggest that compliance
levels inside the Marine Park vary by zone. Never-
theless, a greater understanding of levels of non-
compliance inside PPAs would provide improved
understanding of the overall efficacy of PPAs.
Although coral trout (Plectropomus spp.) are
undoubtedly one of the most highly prized and
targeted fishes on the GBR, there was no evidence of
a reduction in their abundance or size inside PPAs,
compared to NTMRs, as would have been expected
from a significant fishing effect. Rather, PPAs sup-
ported larger coral trout on average, with a greater
proportion of fish above the minimum legal length.
This trend is suggestive of more successful recruit-
ment inside NTMRs, which may act as a nursery area
for younger/ smaller coral trout, however repeated
seasonal surveys would be required to further explore
this. Our results for coral trout abundance contrast
strongly with a number of studies in the GBRMP that
have demonstrated strong fishing effects, with up to
fivefold differences between NTMR and fished zones,
although it is notable that many of these studies did not
explicitly consider PPAs (Boaden and Kingsford
2015; Emslie et al. 2015; McCook et al. 2010;
Williamson et al. 2004). A study by Frisch et al.
(2012) in the nearby Palm Islands group showed
NTMRs supported &1.3 -3.3 times the abundance of
coral trout compared to adjacent PPAs, which also
contrasts with our results.
In the present study, detection of zoning effects
may have been influenced by the strong relationship
between coral trout abundance and water visibility.
Coral trout were one of the most prevalent taxa
observed (second to stripey snapper), however their
observed timid behaviour and tendency to hover at the
periphery of the field of view may have reduced their
detectability in lower visibility replicates. Unlike
lutjanids and lethrinids, coral trout rarely directly
approached or fed from the bait bag, and as such were
more difficult to detect in lower visibility replicates
since they were further from the camera. This
behaviour was consistent by zone, and as such is
unlikely to be due to socially learned ‘‘shyness’’ as a
response to fishing activity, as has been proposed for
coral trout in fished areas of the GBR (Leigh et al.
2014). Since neither visibility nor behaviour of coral
trout differed by zone, there was no artificial skewing
of zoning results due to limitations in their detectabil-
ity. However, the strong relationship with visibility
may have nevertheless reduced the ability to detect
differences amongst zones in analyses, resulting in the
non-significant zoning effects observed. This is an
unfortunate albeit unavoidable outcome of conducting
visual surveys in low visibility environments, where
diver conducted underwater visual surveys are not
feasible. These results highlight the importance of
explicitly including visibility as a covariate in com-
parisons of survey data, especially in inshore regions.
In contrast to coral trout, stripey snapper (L. car-
ponotatus) were less abundant inside both fished
zones, although only significantly so inside SMAs.
Stripey snapper were the most prevalent species
recorded, and along with coral trout, were the only
species occurring in greater than 60% of replicates.
Modelled effect sizes predicted that SMAs (i.e.
without spearfishing) supported &56% of the abun-
dance of L. carponotatus found inside NTMRs, how-
ever this proportion was much greater (&83%) for
CPZs, where spearfishing is permitted. This suggests
that L. carponotatus are targeted by fishers in the
region, but are unlikely to be an important spearfishing
target. These results concur with a study by Frisch
et al. (2012), which also found no effect of spearfish-
ing on the abundance or size of L. carponotatus in the
nearby Palm Islands region. Lutjanus carponotatus are
a highly abundant mid-sized predator on inshore and
mid-shelf reefs. It is unlikely they would be targeted
frequently by commercial fishers in the study area, but
it is expected they would be commonly caught and
retained by recreational fishers. (Evans and Russ 2004;
Kritzer 2004; Newman and Williams 1996;
123
Rev Fish Biol Fisheries
Williamson et al. 2004). Given the zoning effects we
detected for L. carponotatus abundance, our results
support this expectation. It is notable, however, that
their size distributions and the proportion of legal-
sized fish were virtually identical amongst zones.
Lutjanus carponotatus have relatively fast growth
rates, reaching the minimum legal catch size of
250 mm in four to five years, with an age maximum
of 18 years (Kingsford 2009). As such, fish that are
captured and retained could be of a range of ages, with
remaining younger fish able to grow fairly quickly to
legal size to replenish fishable populations. Given this,
along with their overall high abundance, it is likely
that this species is relatively resilient to moderate
levels of fishing pressure that would be expected
inside PPAs (Kingsford 2009; Kritzer 2004).
Special Management Areas
In contrast to our hypothesis, we found no evidence to
support a benefit to targeted fishes from the prohibition
of spearfishing inside SMAs. Rather, there was a trend
for lower abundances of targets inside SMAs com-
pared to CPZs where spearfishing is allowed.
Although SMAs may be expected to have a lower
abundance of targets compared to unfished NTMRs,
that they held lower abundances compared to the
CPZs, which are open to spearfishing, was an unex-
pected result. A number of studies in temperate and
tropical reef systems have demonstrated clear benefits
from the prohibition of spearfishing inside reserves,
(Curley et al. 2013; Frisch et al. 2012; Jouvenel and
Pollard 2001; Skinner et al. 2019). Our results may
have occurred for a variety of reasons: firstly, it is
possible that the areas surveyed within the SMA zones
constituted overall a poorer habitat quality compared
to CPZs, and therefore supported reduced abundances
of target fishes. However, we found no evidence to
indicate that habitat quality was systematically poorer
inside SMAs, indeed SMAs tended to be more coral
dominated and less algae dominated than other zones.
Algae in the region was mostly macroalgae of the
genus Sargassum, which is often abundant on inshore
reefs, but is typically considered to be of low habitat
value compared to corals (Ceccarelli et al. 2018;
Chong-Seng et al. 2012). Results from Boosted
Regression Trees also indicated the importance of
coral cover as a driver of primary target abundance.
Given this, it is expected that SMAs provided equal or
greater habitat quality compared to other zones.
A second possibility is that abundances of primary
targets were lower inside the SMA regions before the
revised zoning was implemented in 2004. However, it
is notable that spearfishing was already prohibited by
state fisheries legislation in parts of the current SMAs
prior to the rezoning, so the SMA areas have had a
long history of protection from spearfishing. Given
this, along with the fast growth rates of most target
species, it is unlikely that the SMA regions had
historically low abundances of fishing targets com-
pared to adjacent CPZs. Alternatively, it is possible
that there is little spearfishing occurring inside the
CPZs, rendering the overall fishing pressure similar
inside both PPA zones. If this is the case, then the
lower abundances of key targets observed inside
SMAs may simply demonstrate variation in the
magnitude of fishing effects inside PPAs, arising
mostly from site level variation. Greater spatial
resolution on the extent and type of recreational
fishing activities occurring within the GBRMP would
allow this hypothesis to be tested. Ultimately, our
results indicate that spearfishing is either rare, or of
little impact to target species in the region and suggest
that restrictions on spearfishing are not an important
component of zoning or fisheries management aimed
to benefit targeted fishes in the region.
Key drivers of abundance and species assemblages
As well as top-down effects from fishing, fish species
and assemblages are likely to be influenced by a
variety of bottom-up effects including variations in
habitat composition, depth, and topography (Jones
et al. 2004; Srinivasan 2003; Wilson et al. 2008). As
such, impacts from zoning are bound to occur in the
context of other important drivers, and acknowledging
and understanding these interactions is critical for
responsible management. Utilizing Boosted Regres-
sion Trees (BRTs), we demonstrated that habitat
characteristics were important drivers of abundance,
but that zone had a relatively greater influence on
targeted species compared to non-target species and
groups. These outcomes are consistent with our
predictions, and highlight clear differences in the
drivers of abundance for target and non-target species.
Coral cover, depth, and complexity were the most
prevalent drivers of fish abundance, which is
123
Rev Fish Biol Fisheries
unsurprising given the importance of habitat quality
and rugosity to the abundance and diversity of reef
fishes (Kingsford et al. 2019; Nash et al. 2013). These
results highlight the importance of healthy benthic
habitats as critical underlying structures for effective
marine reserve systems (Jones et al. 2004). Species
assemblages were influenced by zoning, however
zoning trends were weak in comparison to regional
differences, and significant differences between zones
were restricted to NTMRs vs. SMAs in the Dunk
Island region. The species driving zoning differences
were mostly primary targets, but also included
secondary targets such as Lutjanus lemniscatus, and
non-target species such as Epinephelus quoyanus,
which were both strongly associated with NTMRs.
Lutjanus lemniscatus is considered a secondary target,
and although E. quoyanus is not targeted, they are
commonly caught as bycatch and may suffer post-
release mortality.
Management implications
This study provides a valuable contribution to our
understanding of how PPAs operate and affect fish
populations on inshore reefs. It provides some of the
first empirical data on how PPAs within the GBRMP
contribute to intended conservation and/or fishery
management goals. Importantly, our results provide
estimates of the magnitude of zoning effects that allow
the contribution of PPAs to be assessed against the
benchmark of no-take zones, and enable an under-
standing of how PPAs may be used as a complemen-
tary zoning tool. As highlighted by two global reviews
(Sciberras et al. 2015; Zupan et al. 2018), PPAs are
most likely to be effective if they are highly regulated,
with stringent limits on fishing activities, and are
located adjacent to NTMRs. Although there are
concerns regarding the level of compliance in some
regions, the GBRMP largely meets all of the above
criteria, making it an excellent candidate as a model
system in understanding the conservation significance
of PPAs.
Although we did find reductions in the abundance
of highly targeted species inside fished areas, com-
parisons of effect sizes both within this study and with
other previous studies in the region highlight the value
of PPAs in providing intermediate conservation ben-
efits. Furthermore, our results concur with a number of
studies from temperate and tropical Australia, as well
as globally, which have highlighted the potential for
PPAs to act as a ‘‘conservation middle ground’’
(Boaden and Kingsford 2015; Harasti et al. 2018;
Sciberras et al. 2015; Zupan et al. 2018). Partially
protected zones can make an important contribution to
fisheries management, if the implemented restrictions
act to limit effective fishing effort and therefore
support more conservative fish harvest. Our results
suggest that PPAs in the region support populations of
fisheries target species which, although significantly
lower in abundance compared to no-take zones, were
still present in moderate numbers. Given that they
exist alongside NTMRs, PPAs in the GBRMP can be
considered an effective complementary zone type, and
also make an important contribution towards sustain-
able fishing practices in the region.
Our results provide novel empirical data to support
evidence-based conservation and fisheries manage-
ment decisions. This study was, however, limited to a
single region on the GBR, and given the influence of
latitude and cross-shelf distance on fish assemblages
(Russ 1984; Williams 1982), we cannot infer explicit
management outcomes across large spatial scales.
Nevertheless, in marine systems where PPAs may be a
more feasible management option instead of NTMRs,
or complementary to small NTMRs, we encourage
their implementation and investigation of their con-
tribution to fishes, dependent fisheries and associated
marine systems. Important next steps in this field of
research will be to spatially expand our understanding
of PPAs, in order to explicitly assess their performance
against a range of management zones, and in the
context of key environmental and ecological drivers.
Further understanding of how different zoning types
can help balance socio-economic outcomes in marine
systems is and will continue to be a critical component
of marine conservation and fisheries management.
Acknowledgements The authors would like to thank Mark
O’Callaghan for assistance with field work, Marcus Stowar and
Colin Simpindorfer for the provision of research equipment, and
Mike Cappo, Leanne Currey, and Gavin Ericson for technical
expertise, software, and statistical advice. We thank Rhonda
Banks for assistance with mapping.
Author contributions All authors contributed towards
conceiving the project and sampling design. AH and MK
collected and analysed field data. All authors contributed
towards writing and editing the paper, with AH as the primary
and corresponding author.
123
Rev Fish Biol Fisheries
Funding Funding was provided from an Advance Queensland
Research Fellowship to A.E. Hall, in conjunction with funding
from the Great Barrier Reef Marine Park Authority, and James
Cook University. The project was co-funded by an ARC Centre
of Excellence for Coral Reef Studies fund provided to M.J.
Kingsford.
Data availability and material The datasets generated during
and/or analysed during the current study are available from the
corresponding author on reasonable request.
Code availability Code for R statistical procedures are
available from the corresponding author on reasonable request.
Declarations
Conflict of interest The authors declare no conflict of interest.
Ethical approval This research was conducted with approval
from the James Cook Univiersity Animal Ethics Committee
#A2438.
References
Anderson MJ (2001) A new method for non-parametric multi-
variate analysis of variance. Austral Ecol 26:32–46
Anderson MJ (2017) Permutational Multivariate Analysis of
Variance (PERMANOVA). Wiley StatsRef: Statistics
Reference Online:1–15 doi:https://doi.org/10.1002/
9781118445112.stat07841
Anderson MJ, Willis TJ (2003) Canonical analysis of principal
coordinates: a useful method of constrained ordination for
ecology. Ecology 84:511–525. https://doi.org/10.1890/
0012-9658(2003)084[0511:caopca]
Ban NC, McDougall C, Beck M, Salomon AK, Cripps K (2014)
Applying empirical estimates of marine protected area
effectiveness to assess conservation plans in British
Columbia. Canada Biol Conserv 180:134–148. https://doi.
org/10.1016/j.biocon.2014.09.037
Bergseth BJ, Williamson DH, Russ GR, Sutton SG, Cinner JE
(2017) A social–ecological approach to assessing and
managing poaching by recreational fishers. Front Ecol
Environ 15:67–73. https://doi.org/10.1002/fee.1457
Boaden AE, Kingsford MJ (2015) Predators drive community
structure in coral reef fish assemblages. Ecosphere 6:art46
doi:https://doi.org/10.1890/ES14-00292.1
Bobiles RU, Nakamura Y (2019) Partially protected marine
areas as a conservation tool for commercially important
fishes in the Philippines: Do age, size, and design matter?
Reg Stud Mar Sci. https://doi.org/10.1016/j.rsma.2018.
100459
Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR,
Stevens MHH, White J-SS (2009) Generalized linear
mixed models: a practical guide for ecology and evolution.
Trends Ecol Evol 24:127–135. https://doi.org/10.1016/j.
tree.2008.10.008
Bradley M, Baker R, Sheaves M (2017) Hidden components in
tropical seascapes: deep-estuary habitats support unique
fish assemblages. Estuar Coasts: J ERF 40:1195–1206.
https://doi.org/10.1007/s12237-016-0192-z
Cappo M, De’ath G, Speare P (2007) Inter-reef vertebrate
communities of the Great Barrier Reef Marine Park
determined by baited remote underwater video stations.
Mar Ecol Prog Ser 350:209–221
Cappo M, Speare P, De’ath G (2004) Comparison of baited
remote underwater video stations (BRUVS) and prawn
(shrimp) trawls for assessments of fish biodiversity in inter-
reefal areas of the Great Barrier Reef Marine Park. J Exp
Mar Biol Ecol 302:123–152. https://doi.org/10.1016/j.
jembe.2003.10.006
Cappo M, Stowar M, MacNeil M (2010) The influence of zoning
(closure to fishing) on fish communities of the shoals and
reef bases of the Great Barrier Reef Marine Park. Results of
repeated surveys of the southern banks and Cardwell
shoals, and an overview with regional comparisons. Report
to the Australian Government’s Marine and Tropical Sci-
ences Research Facility (MTSRF). Australian Institute of
Marine Science (AIMS), Townsville
Cappo M, Stowar M, Syms C, Johansson C, Cooper T (2011)
Fish-habitat associations in the region offshore from James
Price Point: a rapid assessment using Baited Remote
Underwater Video Stations (BRUVS). J Royal Soc West
Aust 94:303–321
Ceccarelli DM et al (2018) Rehabilitation of coral reefs through
removal of macroalgae: state of knowledge and consider-
ations for management and implementation. Restor Ecol
26:827–838. https://doi.org/10.1111/rec.12852
Chong-Seng KM, Mannering TD, Pratchett MS, Bellwood DR,
Graham NAJ (2012) The influence of coral reef benthic
condition on associated fish assemblages. PLoS ONE 7:10.
https://doi.org/10.1371/journal.pone.0042167
Clarke KR, Gorley RN, Somerfield RN, Warwick RM (2014)
Change in marine communities: an approach to statistical
analysis and interpretation: 3rd Edition. In. Plymouth,
Coleman MA, Palmer-Brodie A, Kelaher BP (2013) Conser-
vation benefits of a network of marine reserves and par-
tially protected areas. Biol Conserv 167:257–264. https://
doi.org/10.1016/j.biocon.2013.08.033
Curley BG, Glasby TM, Curley AJ, Creese RG, Kingsford MJ
(2013) Enhanced numbers of two temperate reef fishes in a
small, partial-take marine protected area related to spear-
fisher exclusion. Biol Conserv 167:435–445. https://doi.
org/10.1016/j.biocon.2013.07.031
Day JC, Kenchington RA, Tanzer JM, Cameron DS (2019)
Marine zoning revisited: How decades of zoning the Great
Barrier Reef has evolved as an effective spatial planning
approach for marine ecosystem-based management. Aquat
Conserv 29:9–32. https://doi.org/10.1002/aqc.3115
Denny CM, Babcock RC (2004) Do partial marine reserves
protect reef fish assemblages? Biol Conserv 116:119–129.
https://doi.org/10.1016/S0006-3207(03)00183-6
Department of Agriculture F, and Forestry (DAFF) (2020)
QFish Commercial Line Fishery Data. https://qfish.
fisheries.qld.gov.au/. Accessed 18/06/2020 2020
Edgar GJ et al(2014) Global conservation outcomes depend on
marine protected areas with five key features. Nature
506:216-. https://doi.org/10.1038/nature13022
123
Rev Fish Biol Fisheries
Elith J, Leathwick JR, Hastie T (2008) A working guide to
boosted regression trees. J Anim Ecol 77:802–813. https://
doi.org/10.1111/j.1365-2656.2008.01390.x
Emslie MJ, Cheal AJ, Logan M (2017) The distribution and
abundance of reef-associated predatory fishes on the Great
Barrier Reef. Coral Reefs 36:829–846. https://doi.org/10.
1007/s00338-017-1573-x
Emslie MJ et al (2015) Expectations and outcomes of reserve
network performance following re-zoning of the Great
Barrier Reef Marine Park. Curr Biol 25:983–992. https://
doi.org/10.1016/j.cub.2015.01.073
Espinoza M, Cappo M, Heupel MR, Tobin AJ, Simpfendorfer
CA (2014) Quantifying shark distribution patterns and
species-habitat associations: implications of marine park
zoning. PLoS ONE 9:e106885. https://doi.org/10.1371/
journal.pone.0106885
Evans RD, Russ GR (2004) Larger biomass of targeted reef fish
in no-take marine reserves on the Great Barrier Reef,
Australia. Aquat Conserv 14:505–519. https://doi.org/10.
1002/aqc.631
Fox HE et al (2012) Reexamining the science of marine pro-
tected areas: linking knowledge to action. Conserv Lett
5:1–10. https://doi.org/10.1111/j.1755-263X.2011.00207.
x
Frisch AJ, Cole AJ, Hobbs J-PA, Rizzari JR, Munkres KP
(2012) Effects of spearfishing on reef fish populations in a
multi-use conservation area. PLoS ONE 7:e51938. https://
doi.org/10.1371/journal.pone.0051938
Gaines SD, White C, Carr MH, Palumbi SR (2010) Designing
marine reserve networks for both conservation and fish-
eries management. Proc Natl Acad Sci 107:18286–18293.
https://doi.org/10.1073/pnas.0906473107
GBRMPA (2003) Report on the Great Barrier Reef Marine Park
Zoning Plan. Great Barrier Reef Marine Park Authority
Gell FR, Roberts CM (2003) Benefits beyond boundaries: the
fishery effects of marine reserves. Trends Ecol Evol
18:448–455. https://doi.org/10.1016/s0169-
5347(03)00189-7
Halpern BS, Lester SE, Kellner JB (2009) Spillover from marine
reserves and the replenishment of fished stocks. Environ
Conserv 36:268–276. https://doi.org/10.1017/
s0376892910000032
Harasti D, Williams J, Mitchell E, Lindfield S, Jordan A (2018)
Increase in relative abundance and size of snapper Chry-
sophrys auratus within partially-protected and no-take
areas in a temperate marine protected area. Front Mar Sci.
https://doi.org/10.3389/fmars.2018.00208
Harrison HB et al (2012) Larval export from marine reserves
and the recruitment benefit for fish and fisheries. Curr Biol
22:1023–1028
Hughes TP, Bellwood DR, Folke CS, McCook LJ, Pandolfi JM
(2007) No-take areas, herbivory and coral reef resilience.
Trends Ecol Evol 22:1–3
Hurlbert, (1984) Pseudoreplication and the design of ecological
field experiments. Ecol Monogr 54:187–211
Jones GP, McCormick ML, Srinivasan M, Eagle JV (2004)
Coral decline threatens fish biodiversity in marine reserves.
P Natl Acad Sci 101:8251–8253. https://doi.org/10.1073/
pnas.0401277101
Jouvenel JY, Pollard DA (2001) Some effects of marine reserve
protection on the population structure of two spearfishing
target-fish species, Dicentrarchus labrax (Moronidae) and
Sparus aurata (Sparidae), in shallow inshore waters, along
a rocky coast in the northwestern Mediterranean. Sea
Aquat Conserv-Mar Freshw Ecosyst 11:1–9. https://doi.
org/10.1002/aqc.424
Kingsford M, Syms C, Srinivasan M, Jones GP (2019) Coral reef
habitats and their influence on reef assemblages. In:
Hutchings P, Kingsford M, Hoegh-Guldberg O (eds) Great
Barrier Reef: Biology, Environment and Management.
CSIRO Publishing and CRC Press, Melbourne, Australia,
Kingsford MJ (2009) Contrasting patterns of reef utilization and
recruitment of coral trout (Plectropomus leopardus) and
snapper (Lutjanus carponotatus) at One Tree Island,
southern Great Barrier Reef. Coral Reefs 28:251–264.
https://doi.org/10.1007/s00338-008-0421-4
Kritzer JP (2004) Sex-specific growth and mortality, spawning
season, and female maturation of the stripery bass (Lut-
janus carponotatus) on the Great Barrier Reef. Fish Bull
102:94–107
Langlois TJ, Newman SJ, Cappo M, Harvey ES, Rome BM,
Skepper CL, Wakefield CB (2015) Length selectivity of
commercial fish traps assessed from in situ comparisons
with stereo-video: Is there evidence of sampling bias? Fish
Res 161:145–155. https://doi.org/10.1016/j.fishres.2014.
06.008
Leigh GM, Campbell AB, Lunow CP, O’Neil MF (2014) Stock
assessment of the Queensland east coast common coral
trout (Plectropomus leopardus) fishery. Department of
Agriculture Fisheries and Forestry, Brisbane
Lester SE, Halpern BS (2008) Biological responses in marine
no-take reserves versus partially protected areas. Mar Ecol
Prog Ser 367:49–56. https://doi.org/10.3354/meps07599
Malcolm HA et al (2018) Targeted fishes are larger and more
abundant in ‘‘no-take’’ areas in a subtropical marine park.
Estuar Coast Shelf Sci 212:118–127. https://doi.org/10.
1016/j.ecss.2018.07.003
Mapstone BD, Carlos G, Lunow CP, Reid AE (2004) Uncer-
tainty in length measurements of live coral trout: impli-
cations for compliance to and enforcement of minimum
legal size limits vol technical report No 54. CRC Reef
Research Centre, Townsville
McCook LJ et al (2010) Adaptive management of the Great
Barrier Reef: a globally significant demonstration of the
benefits of networks of marine reserves. P Natl Acad Sci
107:18278–18285. https://doi.org/10.1073/pnas.
0909335107
Nash KL, Graham NAJ, Wilson SK, Bellwood DR (2013)
Cross-scale habitat structure drives fish body size distri-
butions on coral reefs. Ecosystems 16:478–490. https://doi.
org/10.1007/s10021-012-9625-0
Newman SJ, Williams DM (1996) Variation in reef associated
assemblages of the Lutjanidae and Lethrinidae at different
distances offshore in the central Great Barrier Reef.
Environ Biol Fish 46:123–138
Parker D, Winker H, Bernard ATF, Heyns-Veale ER, Langlois
TJ, Harvey ES, Gotz A (2016) Insights from baited video
sampling of temperate reef fishes: how biased are angling
surveys? Fish Res 179:191–201. https://doi.org/10.1016/j.
fishres.2016.02.025
123
Rev Fish Biol Fisheries
R Core Team (2018) R: A language and environment for sta-
tistical computing. . R Foundation for Statistical Comput-
ing. URL https://www.R-project.org/
Russ GR (1984) Distribution and abundance of herbivorous
grazing fishes in the central Great Barrier Reef. I. Levels of
variability across the entire continental shelf. Mar Ecol
Prog Ser 20:23–34
Russ GR, Alcala AC, Maypa AP, Calumpong HP, White AT
(2004) Marine reserve benefits local fisheries. Ecol Appl
14:597–606
Russ GR, Cheal AJ, Dolman AM, Emslie MJ, Evans RD, Miller
I, Sweatman H, Williamson DH (2008) Rapid increase in
fish numbers follows creation of world’s largest marine
reserve network. Curr Biol 18:R514–R515
Sciberras M, Jenkins SR, Mant R, Kaiser MJ, Hawkins SJ, Pullin
AS (2015) Evaluating the relative conservation value of
fully and partially protected marine areas. Fish Fish
16:58–77. https://doi.org/10.1111/faf.12044
Skinner C, Newman SP, Box S, Narozanski A, Polunin NVC
(2019) Chronic spearfishing may indirectly affect reef
health through reductions in parrotfish bite rates. J Fish
Biol 94:585–594. https://doi.org/10.1111/jfb.13939
Srinivasan M (2003) Depth distributions of coral reef fishes: the
influence of microhabitat structure, settlement, and post-
settlement processes. Oecologia 137:76–84. https://doi.
org/10.1007/s00442-003-1320-6
Stowar M, De’ath G, Doherty P, Johansson C, Speare P,
Venebles B (2008) Influence of zoning on midshelf shoals
of the southern great barrier reef report to the marine and
tropical sciences research facility. Reef and Rainforest
Research Centre Limited, Cairns
Thiault L et al (2020) Predicting poaching risk in marine pro-
tected areas for improved patrol efficiency. J Environ
Manage 254:9. https://doi.org/10.1016/j.jenvman.2019.
109808
Webley J, McInnes K, Tiexeira D, Lawson A, Quinn R (2015)
Statewide Recreational Fishing Survey 2013–14. Depart-
ment of Agriculture and Fisheries,
Weekers DP, Zahnow R (2019) Risky facilities: Analysis of
illegal recreational fishing in the Great Barrier Reef Marine
Park, Australia. Aust N Z J Criminol 52:368–389. https://
doi.org/10.1177/0004865818804021
Wenger AS, Williamson DH, da Silva ET, Ceccarelli DM,
Browne NK, Petus C, Devlin MJ (2016) Effects of reduced
water quality on coral reefs in and out of no-take marine
reserves. Conserv Biol 30:142–153. https://doi.org/10.
1111/cobi.12576
Williams DM (1982) Patterns in the distribution of fish com-
munities across the Central Great Barrier Reef. Coral Reefs
1:35–43. https://doi.org/10.1007/bf00286538
Williamson DH, Ceccarelli DM, Jones GP, Russ GR (2019)
Assessing the ecological effects of management zoning on
inshore reefs of the great barrier reef marine park reef 2050
integrated monitoring and reporting program milestone
report 2. Great Barrier Reef Marine Park Authority,
Townsville
Williamson DH, Russ GR, Ayling AM (2004) No-take marine
reserves increase abundance and biomass of reef fish on
inshore fringing reefs of the Great Barrier Reef. Environ
Conserv 31:149–159. https://doi.org/10.1017/
s0376892904001262
Wilson SK et al (2008) Exploitation and habitat degradation as
agents of change within coral reef fish communities. Glob
Change Biol 14:2796–2809
Zupan M, Fragkopoulou E, Claudet J, Erzini K, Costa BHE,
Goncalves EJ (2018) Marine partially protected areas:
drivers of ecological effectiveness. Front Ecol Environ
16:381–387. https://doi.org/10.1002/fee.1934
Publisher’s Note Springer Nature remains neutral with
regard to jurisdictional claims in published maps and
institutional affiliations.
123
Rev Fish Biol Fisheries