Marine partially protected areas: drivers of ecological effectiveness

Abstract

The number of marine protected areas (MPAs) has grown exponentially worldwide over the past decade in order to meet international
targets. Most of these protected areas allow extraction of resources and are therefore designated as “partially protected
areas” (PPAs). However, the effectiveness of PPAs remains unclear due to the high variability of use types permitted. Here, we
carried out what we believe to be the first global meta-analysis of PPAs using a regulation-based classification system for MPAs to
assess their ecological effectiveness. This novel classification allows for unambiguous differentiation between areas according to
allowed use, which is the key feature determining PPA performance. Highly and moderately regulated areas exhibited higher biomass
and abundance of commercial fish species, whereas fish abundance and biomass in weakly regulated areas differed little from
unprotected areas. Notably, the effectiveness of moderately regulated areas can be enhanced by the presence of an adjacent fully
protected area. We concluded that limited and well-regulated uses in PPAs and the presence of an adjacent fully protected area
confer ecological benefits, from which socioeconomic advantages are derived.

Full text

© The Ecological Society of America Front Ecol Environ doi:10.1002/fee.1934
RESEARCH COMMUNICATIONS 1
Front Ecol Environ 2018; 16(7): 1–7, doi:10.1002/fee.1934
Coastal zones around the world are increasingly subjected
to human and environmental pressures and are in need of
strategic management (Halpern et al. 2015). The establishment
of marine protected areas (MPAs) is a commonly used tool for
improving conservation, food security, and fisheries manage-
ment (Gaines et al. 2010). The ecological effects of fully
protected areas (ie no- take areas) are well studied, and the
abundance and size of species are usually enhanced within (eg
Claudet et al. 2008; Edgar et al. 2014) and in some cases out-
side of (eg Caselle et al. 2015) these areas. MPAs also support
the recovery of populations and communities of fish and other
marine taxa and can preserve habitat structure (Sandin et al.
2008).
The establishment of fully protected areas has often resulted
in conflicts between conservation and socioeconomic objec-
tives, especially in areas with numerous users and types of uses
(Fox et al. 2011). As such, the implementation of partially pro-
tected areas (PPAs), in which some extractive activities may be
allowed, has in some cases become a preferable option, given
that PPAs can provide a better balance between social and eco-
logical objectives, and may be easier to implement.
Simultaneously, in response to international agreements and
commitments, more and more MPAs are being established,
most of which are PPAs of one type or another (Lubchenco
and Grorud- Colvert 2015). It is therefore urgent to identify
which forms of partial protection can provide socioeconomic
benefits while still protecting biodiversity.
PPAs are context- dependent, and their regulations vary
with management objectives; in turn, regulations will likely
affect their ecological effectiveness. Only a handful of studies
have examined the effects of different levels of partial protec-
tion (eg Di Franco et al. 2009; Sciberras et al. 2013; Ban et al.
2014), none of which have been based on a systematic classifi-
cation for these different levels, leading to variable results that
are difficult to generalize. Sciberras et al. (2013), for instance,
broadly characterized three types of PPAs based on replies to a
survey questionnaire that included somewhat subjective ques-
tions (eg whether an activity damages the bottom, targets par-
ticular species, or affects other species); moreover, the study
did not account for such factors as aquaculture, bottom
exploitation, and other non- extractive activities (eg anchoring)
that may impact the marine habitat.
Ban et al. (2014) re- analyzed the dataset used by Sciberras
et al. (2013) but used the International Union for Conservation
of Nature (IUCN) categories of protected areas instead (see
Table 1 in Ban et al. 2014); however, the current IUCN classifica-
tion system is based on management objectives that can be mis-
matched to regulations, resulting in considerable uncertainty
when evaluating MPA effectiveness (Horta e Costa et al. 2016).
In fact, when correlating IUCN categories with the expected
impacts of activities, there is a high degree of variability among,
and overlap between, categories. There is also no clear trend
between the expected cumulative impacts of activities and the
IUCN classification scheme, from more restricted (Ia) to less
restricted (V or VI) categories (Horta e Costa et al. 2016).
A recently published regulation- based classification system
for MPAs, that of Horta e Costa et al. (2016), presents a new
Marine partially protected areas: drivers of
ecological effectiveness
Mirta Zupan1†, Eliza Fragkopoulou1,2†, Joachim Claudet3,4, Karim Erzini2, Bárbara Horta e Costa1,2,3,4, and
Emanuel J Gonçalves1*
The number of marine protected areas (MPAs) has grown exponentially worldwide over the past decade in order to meet interna-
tional targets. Most of these protected areas allow extraction of resources and are therefore designated as “partially protected
areas” (PPAs). However, the effectiveness of PPAs remains unclear due to the high variability of use types permitted. Here, we
carried out what we believe to be the first global meta- analysis of PPAs using a regulation- based classification system for MPAs to
assess their ecological effectiveness. This novel classification allows for unambiguous differentiation between areas according to
allowed use, which is the key feature determining PPA performance. Highly and moderately regulated areas exhibited higher bio-
mass and abundance of commercial fish species, whereas fish abundance and biomass in weakly regulated areas differed little from
unprotected areas. Notably, the effectiveness of moderately regulated areas can be enhanced by the presence of an adjacent fully
protected area. We concluded that limited and well- regulated uses in PPAs and the presence of an adjacent fully protected area
confer ecological benefits, from which socioeconomic advantages are derived.
1MARE Marine and Environmental Sciences Centre, ISPA – Instituto
Universitário, Lisbon, Portugal *(emanuel@ispa.pt); 2Centre of Marine
Sciences, CCMAR, University of Algarve, Campus de Gambelas, Faro,
Portugal; 3National Center for Scientific Research, PSL Université Paris,
CRIOBE, USR 3278 CNRS-EPHE-UPVD, Maison des Océans, Paris,
France; 4Laboratoire d’Excellence CORAIL, Moorea, French Polynesia
†these authors contributed equally to this work

Front Ecol Environ doi:10.1002/fee.1934 © The Ecological Society of America
M Zupan et al.
2 RESEARCH COMMUNICATIONS
way to categorize both MPAs and each type of zone within
them according to allowed commercial and recreational uses
(WebFigure 1). In this system, PPAs are classified based on the
cumulative impacts of allowed activities.
Understanding the ecological responses of various types of
partial protection is essential, since most MPAs are multiple-
use and the ecological effects that each PPA provides are likely
linked to different regulatory regimes (Fox et al. 2011). In this
paper, we present a novel approach to investigate and infer how
varying levels of partial protection lead to varying ecological
effects through a global meta- analysis. We also examine how
design characteristics that are known to influence the effec-
tiveness of no- take areas, such as protected area age and size
(Claudet et al. 2008), or that are specific to multiple- use MPAs,
such as the presence of an adjacent fully protected area, may
also mediate the effectiveness of partial protection.
Methods and materials
Data selection: response variables and covariates
We built our database from studies compiled by Sciberras
et al. (2013) and Horta e Costa et al. (2016), updated with
recent peer- reviewed literature obtained via a database search
following the methods of Sciberras et al. (2013). We limited
our analyses to studies that reported values for abundance
and/or biomass of nsh species targeted by sheries, as
they are directly aected by the protection regimes. In order
to qualify, studies must also have included a comparison
of these ecological variables between PPAs and surrounding
open areas, which we will refer to hereaer as “unprotected
areas”. We only retained studies that reported ecological
responses for a particular PPA when they were compared
to unprotected areas, but not in cases where biological
responses were aggregated for an entire multiple- use MPA
with varied regulations. Studies that reported ecological
responses for PPAs with dierent protection levels within
the same MPA were included separately in the database,
because they represented dierent types of partial protection.
In cases where more than one study investigated the eects
of protection, only the most recent was retained, unless
dierent metrics were used among the studies. Although it
would have been important to assess eects on the overall
biodiversity of these areas, data for non- target species were
not suciently available across studies to allow for a detailed
analysis.
The studies had to report the mean of the response variable
(abundance and/or biomass), sample size (eg number of tran-
sects), and an appropriate error measure (eg variance). If the
study assessed abundance and biomass of targeted fish species
over some other variables (eg depth, habitat types), data were
averaged for each variable. When data were collected over time,
only the most recent results were extracted, as they represented
the longest duration of protection; however, when data were
reported several times within a year, results were averaged for
that year to minimize seasonal effects associated with sampling
period. Similarly, when data were reported for multiple tar-
geted species (k), we calculated the overall mean (
̄
X
) and
standard deviation (SD) for the study as:
and
where
̄x
is the mean biomass or abundance for species j,
and SD and ni are the standard deviation and sample sizes
(eg number of transects) associated with ̄
xj
.
As mentioned, we classified each PPA based on the system
described by Horta e Costa et al. (2016), in which each area
type allows different activities. Five classes of PPAs were iden-
tified: (1) highly regulated, (2) moderately regulated, (3)
weakly regulated, (4) very weakly regulated, and (5) unregu-
lated (WebFigure 1). Highly regulated areas were defined as
those allowing only a limited number (five maximum) of low-
impact types of fishing gear (eg lines, octopus trap), moder-
ately regulated areas were defined as those that allow more (up
to ten) low- to medium- impact fishing gear types (eg gillnets),
and weakly regulated areas were defined as those in which
higher- impact gear types (eg beach seines, bottom trawling,
trammel nets) were permitted.
We recorded the age (years since establishment) and size of
each PPA, as well as the presence or absence of an adjacent
fully protected area (when side by side with a PPA and part of
a multiple- use MPA). We also scored the capacity to imple-
ment regulations using an index for fisheries management
effectiveness (Mora et al. 2009) at the national level as a proxy
for enforcement of fishing regulations in MPAs. Values ranged
from 0 to 1, with 0 representing low enforcement capacity and
1 representing high enforcement capacity.
The final database consisted of 26 peer- reviewed research
articles and 49 case studies worldwide (WebTable 1). Of the
PPAs included in the 49 case studies, 24 were characterized as
highly regulated, 17 as moderately regulated, seven as weakly
regulated, and one as very weakly regulated. We restricted our
analysis to the first three classes.
Meta- analysis
We used a weighted random- eects meta- analysis to assess
the ecological eectiveness of PPAs. e eect size Ri for
each area i was modeled as a natural logarithm (ln) response
ratio of the mean (
̄
Xi
) abundance or biomass estimates
measured within and outside the PPA (Osenberg et al. 1997;
Hedges et al. 1999):
(Eq 1)
̄
X
=
∑
k
j=1nj
̄xj
∑
k
j
=
1
nj
(Eq 2),
SD
=
√
1
k
2
∑
k
j=1
SD2
j
(Eq 3),
R
i=ln
(̄
XPPAi
̄
XUPA
i)

© The Ecological Society of America Front Ecol Environ doi:10.1002/fee.1934
Marine partially protected areas RESEARCH COMMUNICATIONS 3
where
̄
XPPA
and
̄
XUPA
are the mean abundance/biomass within
and outside the PPA of study i, respectively. e variance
vi of the eect sizes (ie the within- study variance) was cal-
culated as follows:
where
̄
XPPA
and
̄
XUPA
are the mean abundance/biomass within
and outside the PPA of study i, respectively; SDPPA and
SDUPA are the standard deviations associated with
̄
XPPA
and
̄
XUPA
of study i, respectively; and nPPA and nUPA are the
sample sizes of study i for the estimation of the mean (eg
number of transects). As in traditional random- eects meta-
analyses, our weights wi included both the within- and
among- study variances, and were calculated as follows:
where vi is dened as above and vA is the among- study
variance.
The overall effect of partial protection was calculated as a
weighted average of the effect sizes:
where wi and Ri are dened above. e overall heterogeneity
(Qt) was calculated as:
and its signicance was tested against the χ2 distribution
with ni – 1 degrees of freedom.
We used weighted general linear (mixed- effects) models to
examine how different features impact the ecological effective-
ness of PPAs. We first investigated if different types of areas
exhibited different levels of ecological responses. For a given class
category, weighted cumulative effect sizes were calculated as:
where nc is the number of PPAs belonging to class c, and Ri
and wi are dened as above. e heterogeneity of the model
explained by the class (Qm) was calculated as follows:
where m is the number of classes
̄
R
, and
̄
Rc
is calculated
as above. e signicance of Qm was tested against the χ2
distribution with nc – 1 degrees of freedom.
In addition, we ran models to assess if different features were
mediating the response to protection, namely (1) the age of the
protected area, (2) the size of the protected area (measured in
square kilometers and log- transformed in the analyses), (3) the
capacity to implement regulations, and (4) the presence/
absence of an adjacent fully protected area. We ran mixed-
effects categorical analyses for categorical variables and applied
meta- analytic regression through linear mixed- effects models
to the continuous variables. In addition, interaction models
between classes and each of the features were also tested
(WebTable 2). All statistical analyses were performed with R (R
Core Team 2016).
Results
Abundance and biomass of targeted sh species were sig-
nicantly higher overall within PPAs than in unprotected
areas (on average 2.4 and 2.9 times higher, respectively;
Figure 1). PPA eectiveness was, however, variable across
studies, both in terms of abundance (Ri = 0.89, Qt = 961,
df [degrees of freedom] = 35, P < 0.001) and biomass (Ri
= 1.08, Qt = 2197, df = 38, P < 0.001), with dierent classes
exhibiting dierent levels of eectiveness (abundance Qm =
11.35, P = 0.0034; biomass Qm = 6.6636, P = 0.048). When
compared to unprotected areas, highly regulated PPAs sup-
ported 2.9 times higher sh abundance (Rk = 1.1) and 3
times higher sh biomass (Rk = 1.12), and moderately reg-
ulated PPAs supported 2.9 times higher sh abundance (Rk
(Eq 4),
v
i=
SD
2
PPAi
nPPA
i
∗
̄
X2
PPA
i
+
SD
2
UPAi
nUPA
i
∗
̄
X2
UPA
i
(Eq 5),
w
i=
1
vi
+
vA
(Eq 6),
̄
R
=
∑n
i
i=1wiRi
∑
ni
i=1
w
i
(Eq 7),
Q
t=
∑n
i
i=1
wi(Ri−
̄
R)
2
(Eq 8),
̄
R
c=
∑n
c
i=1wiRi
∑
nc
i=1
w
i
(Eq 9),
Q
m=
∑m
j
=1
∑n
c
i
=1wij(
̄
Rc−
̄
R)
2
Figure1. Ecological effectiveness of partially protected areas (PPAs) for
(a) abundance and (b) biomass of targeted fish species for all PPAs com-
bined and for PPAs grouped by class (sensu Horta e Costa et al. 2016). The
horizontal dotted line at 1 represents equal fish abundance or biomass
within and outside the PPA; values greater than 1 indicate more fish (or
more biomass) within the PPA; values below 1 indicate fewer fish (or less
biomass) within the PPA. The bars represent 95% confidence intervals.
Sample sizes for each group are shown.
(a) (b)

Front Ecol Environ doi:10.1002/fee.1934 © The Ecological Society of America
M Zupan et al.
4 RESEARCH COMMUNICATIONS
= 1.07) and 4.2 times higher sh biomass (Rk = 1.42).
However, sh abundance (Rk = –0.13) and biomass (Rk =
0.18) in weakly regulated PPAs did not dier from that in
surrounding unprotected areas (Figure 1).
Ecological effectiveness increased with both the age and size
of PPAs, and with the capacity to implement regulations
(Figure2; WebFigure 2; WebTable 2a). Abundance and biomass
of targeted fish species increased on average by 5.1% and 4.6%
annually, respectively, in protected areas relative to unprotected
areas following implementation. For every tenfold increase in
the size of a PPA, fish abundance and biomass increased by 37%
and 46%, respectively. Furthermore, increasing the implementa-
tion capacity by 10% resulted in 4.3- and 6.4- fold higher abun-
dance and biomass of targeted fish species, respectively. The
effect of age, size, and capacity to implement regulations varied
across the three PPA classes, yet these interactions were signifi-
cant only for targeted fish species abundance and not biomass
(WebTable 2b). Targeted species within moderately and highly
regulated areas were positively affected by age, size, and the
capacity to implement regulations, whereas no significant effect
was detected for targeted species within weakly regulated areas
(Figure2; WebFigure 2).
Interestingly, the presence of a fully pro-
tected area adjacent to a PPA played a role in
enhancing the ecological effectiveness of par-
tial protection (abundance Qm = 2.05, P = 0.15;
biomass Qm = 5.47, P = 0.082). Fish abundance
and biomass were on average 1.6 and 2.1 times
higher, respectively, within PPAs that were
adjacent to a fully protected area (Figure 3).
This effect varied across the three classes
(abundance Qm = 22.07, P = 0.0005; biomass
Qm = 12.59, P = 0.096), with some moderately
regulated areas showing positive ecological
benefits only when adjacent to a fully protected
area (Figure3; WebTable 2b) and weakly regu-
lated areas not showing any benefit.
Discussion
We provide what is, to our knowledge, the
rst global assessment of the performance of
marine PPAs based on a regulation- based MPA
classication system (Horta e Costa et al. 2016).
We show that the ecological eectiveness of
partial protection depends on specic compo-
nents: (1) their type (classied according to
allowed uses; WebFigure 1), (2) the presence
of an adjacent fully protected area that might
inuence their eectiveness, (3) the capacity
to enforce regulations, and their (4) age and
(5) size. ese results help to clarify the pre-
viously reported mixed responses to protection
in PPAs (eg Lester and Halpern 2008; Di
Franco et al. 2009; Sciberras et al. 2013).
Our most notable finding is that regulations are the key fea-
ture determining the ecological effectiveness of PPAs.
Moderately and highly regulated areas are effective at harbor-
ing greater abundances and biomass of targeted fish species as
compared to unprotected areas, whereas no ecological benefits
were detected in weakly regulated areas. Highly and moder-
ately regulated PPAs permit some extractive uses (maximum
of five and 10 fishing gears, respectively) that have low (eg lines
and traps) or moderate (eg gillnets) impacts on ecosystems.
Weakly regulated areas permit more types of fishing gear and/
or types that have greater negative environmental impacts (eg
trawling; Horta e Costa et al. 2016). Fernández- Chacón et al.
(2015) demonstrated empirically that the exclusion of several
fishing gears within PPAs resulted in fish species targeted by
those gears benefiting from protection as compared to popula-
tions in unprotected areas.
In addition, we show that combining a fully protected area
with moderately regulated ones confers positive benefits
(Figure3), with the full range of response always above 1 (non-
significant differences between partial protection and open
areas are shown when response overlaps 1). As this class (ie
moderately regulated) is a common choice of MPA design,
Figure2. Ecological effectiveness of the classes of PPAs as mediated by PPA age (a and b)
and size (c and d) for abundance (top panel) and biomass (bottom panel) of targeted fish spe-
cies. The horizontal dotted line at 1 represents equal fish abundance or biomass inside and
outside the PPA; values greater than 1 indicate more fish (or more biomass) within the PPA;
values below 1 indicate fewer fish (or less biomass) within the PPA. The fitted lines are regres-
sions of each PPA class and the corresponding feature (solid line: significant regression, P <
0.05; dashed line: non- significant regression, P > 0.05).
(a) (b)
(c) (d)

© The Ecological Society of America Front Ecol Environ doi:10.1002/fee.1934
Marine partially protected areas RESEARCH COMMUNICATIONS 5
placing these areas adjacent to fully protected
areas is an important option to consider, since
doing so can enhance their ecological benefits.
Highly and weakly regulated PPAs may be less
sensitive to the presence of an adjacent fully
protected area for different reasons. For highly
regulated areas, this is likely due to the limited
amount of extractive activities permitted
within them, which already confers high con-
servation benefits, whereas weakly regulated
areas may be less influenced by an adjacent
fully protected area due to the large number of
activities with substantial impacts that occur in
these areas. In moderately regulated areas, reg-
ulations alone may be insufficient to greatly
enhance populations of targeted fish species;
moreover, spillover effects from an adjacent
no- take area may increase their ecological
effectiveness (eg Hackradt et al. 2014). Spillover
effects from highly regulated PPAs may benefit
adjacent areas with weaker regulations, but
more research is needed to test this. Future
studies should assess how designing MPAs
with different combinations of protection lev-
els affects ecological responses.
We also show that the effectiveness of pro-
tection is positively correlated with both age
and size, demonstrating that these variables
matter not only for no- take areas but also for
PPAs (Claudet et al. 2008; Edgar et al. 2014).
Moreover, we found that the higher the capac-
ity to implement regulations, the greater the
ecological effectiveness, confirming that
investment in control and enforcement mecha-
nisms should be a high priority when establishing and manag-
ing MPAs (Guidetti et al. 2008; Mora et al. 2009; Edgar et al.
2014). The positive ecological effects associated with larger,
older, and better- enforced PPAs decline, however, with the
number of extractive activities allowed.
Our findings suggest that well- regulated, well- enforced, large,
and longer- established PPAs can provide substantial ecological
benefits, which are enhanced in some cases by the presence of an
adjacent fully protected area (Figure4). Enforcement, age, and
size are key components of success (Edgar et al. 2014). Several
studies have compared the effects of full and partial protection
to unprotected areas, demonstrating that, overall, full protection
provides more ecological benefits than partial protection (eg
Lester and Halpern 2008; Sciberras et al. 2013; Giakoumi et al.
2017). Here, however, we demonstrate that MPAs do not have to
be strictly no- take (Edgar et al. 2014) to provide ecological ben-
efits. Highly regulated PPAs can be effective and sometimes a
preferable option in complex socioecological systems where full
protection is difficult to implement, or as a complement to full
protection in multiple- use MPAs. Moderately regulated areas
can be combined with adjacent fully protected areas to further
enhance ecological benefits. However, the overall ecological ben-
efits of highly regulated PPAs, when compared to full protection,
are much lower; there is 300% more fish biomass and density
within those PPAs than in unprotected areas, but Sala and
Giakoumi (2018) reported 670% higher fish biomass within
fully protected areas than in unprotected areas; Sciberras et al.
(2013) reported 92% higher biomass in no- take areas than in
PPAs; and Gill et al. (2017) found a twofold difference in bio-
mass between no- take areas and PPAs.
The case studies included in our analysis are global in scope,
with most fish biomass and density data being measured on
relatively shallow (less than 30 m) reefs. Mora et al. (2011) and
Cinner et al. (2013) have shown that social factors can influ-
ence the biomass of reef fishes in coastal areas; coastal devel-
opment and land use, human population density (Mora et al.
2011), distance to market, and economic development (Cinner
et al. 2013) can all greatly influence the structure of reef fish
biomass. Future studies should incorporate these correlates
when enough information is available for the different classes
of PPAs. Most of the studies included in our analysis were for
partial protection classes where extraction is limited (highly
Figure3. Ecological effectiveness of classes of PPAs for the (a) abundance and (b) biomass of
targeted fish species as affected by the presence of an adjacent fully protected area (open
symbols). The horizontal dotted line at 1 represents equal fish abundance inside and outside of
the PPA; values greater than 1 indicate more fish (or more biomass) within the PPA; values
below 1 indicate fewer fish (or less biomass) within the PPA. The bars represent 95% confi-
dence intervals. Sample sizes for each group are shown.
(a)
(b)

Front Ecol Environ doi:10.1002/fee.1934 © The Ecological Society of America
M Zupan et al.
6 RESEARCH COMMUNICATIONS
and moderately regulated areas) and therefore stronger
responses are to be expected, whereas only a handful of studies
reported results for areas with lower levels of protection
(weakly and very weakly regulated areas). Publication bias (ie
scientists tend to sample where an effect is likely to be detected
and journals tend to favor the publication of positive results)
can partially explain why we were only able to locate detailed
information for 47 case studies despite there being more than
11,000 MPAs listed globally (MPA Atlas; www.mpatlas.org).
Therefore, we have very likely captured the most effective
PPAs, potentially leading to an overestimation of the average
effects.
The implementation of MPAs requires the integration of
conservation, social, economic, and political goals, and MPA
design should be driven by the particular management objec-
tives. A regulation- based classification system such as the one
used in this study (Horta e Costa et al. 2016) provides an ade-
quate tool to test not only aspirational goals, based on objec-
tives, but also concrete impacts as predicted by regulations of
uses. Our results can assist policy makers and managers in
determining the appropriate levels of protection to reach spe-
cific goals by accounting for the type of regulations adopted in
each MPA.
Acknowledgements
We thank CW Osenberg for fruitful discussion in the early
stages of this manuscript. is research was funded by the
ERA- Net BiodivERsA project “BUFFER – Partially protected
areas as buers to increase the linked social–ecological re-
silience”, with the national funders ANR (France), FCT
(Portugal), FOR- MAS and SEPA (Sweden), and RCN
(Norway). BHC was supported by a grant under the project
BUFFER, a FCT grant (SFRH/BPD/100377/2014), and a
Fernand Braudel IFER fellowship (Fondation Maison des
Sciences de l’Homme). FCT supported this work under the
strategic project UID/MAR/04292/2013. Authors BHC and
EJG share joint senior authorship.
References
Ban NC, McDougall C, Beck M, et al. 2014. Applying empirical esti-
mates of marine protected area effectiveness to assess conserva-
tion plans in British Columbia, Canada. Biol Conserv 180: 134–48.
Caselle JE, Rassweiler A, Hamilton SL, et al. 2015. Recovery trajec-
tories of kelp forest animals are rapid yet spatially variable across
a network of temperate marine protected areas. Sci Rep- UK
5: 14102.
Cinner JE, Graham NAJ, Huchery C, et al. 2013. Global effects of
local human population density and distance to markets on the
condition of coral reef fisheries. Conserv Biol 27: 453–58.
Claudet J, Osenberg CW, Benedetti-Cecchi L, et al. 2008. Marine
reserves: size and age do matter. Ecol Lett 11: 481–89.
Di Franco A, Bussotti S, Navone A, et al. 2009. Evaluating effects of
total and partial restrictions to fishing on Mediterranean rocky-
reef fish assemblages. Mar Ecol- Prog Ser 387: 275–85.
Edgar GJ, Stuart-Smith RD, Willis TJ, et al. 2014. Global conservation
outcomes depend on marine protected areas with five key features.
Nature 506: 216–20.
Fernández-Chacón A, Moland E, Espeland SH, et al. 2015.
Demographic effects of full vs partial protection from harvesting:
inference from an empirical before–after control–impact study on
Atlantic cod. J Appl Ecol 52: 1206–15.
Fox HE, Mascia MB, Basurto X, et al. 2011. Reexamining the science
of marine protected areas: linking knowledge to action. Conserv
Lett 5: 1–10.
Gaines SD, White C, Carr MH, et al. 2010. Designing marine reserve
networks for both conservation and fisheries management. P Natl
Acad Sci USA 107: 18286–93.
Giakoumi S, Scianna C, Plass-Johnson J, et al. 2017. Ecological effects
of full and partial protection in the crowded Mediterranean Sea: a
regional meta- analysis. Sci Rep- UK 7: 8940.
Figure4. Ecological effectiveness of highly regulated PPAs can be high,
as shown in these two photographs. (a) Seabreams (Diplodus sargus and
Diplodus vulgaris) respond strongly to protection, increasing in both abun-
dance and size in areas with high levels of protection, thereby boosting
tourism operations such as diving (here, in Arrábida Marine Park,
Portugal); (b) the endangered dusky grouper (Epinephelus marginatus) is a
long- lived hermaphroditic fish that is highly vulnerable to both recreational
and commercial fishing, but this species thrives in areas with strong pro-
tection (here, in the Formigas Islets, a nature reserve in the Azores).
(a)
(b)

© The Ecological Society of America Front Ecol Environ doi:10.1002/fee.1934
Marine partially protected areas RESEARCH COMMUNICATIONS 7
Gill DA, Mascia MB, Ahmadia GN, et al. 2017. Capacity shortfalls
hinder the performance of marine protected areas globally. Nature
543: 665–69.
Guidetti P, Milazzo M, Bussotti S, et al. 2008. Italian marine
reserve effectiveness: does enforcement matter? Biol Conserv
141: 699–709.
Hackradt CW, Garcia-Charton JA, Harmelin-Vivien M, et al. 2014.
Response of rocky reef top predators (Serranidae: Epinephelinae)
in and around marine protected areas in the Western
Mediterranean Sea. PLoS ONE 9: e98206.
Halpern BS, Frazier M, Potapenko J, et al. 2015. Spatial and temporal
changes in cumulative human impacts on the world’s ocean. Nat
Commun 6: 7615.
Hedges LV, Gurevitch J, Curtis PS, et al. 1999. The meta- analysis of
response ratios in experimental ecology. Ecology 80: 1150–56.
Horta e Costa B, Claudet J, Franco G, et al. 2016. A regulation- based
classification system for marine protected areas (MPAs). Mar
Policy 72: 192–98.
Lester S and Halpern B. 2008. Biological responses in marine no- take
reserves versus partially protected areas. Mar Ecol- Prog Ser 367:
49–56.
Lubchenco J and Grorud-Colvert K. 2015. Making waves: the science
and politics of ocean protection. Science 350: 382–83.
Mora C, Aburto-Oropeza O, Ayala Bocos A, et al. 2011. Global
human footprint on the linkage between biodiversity and ecosys-
tem functioning in reef fishes. PLoS Biol 9: e1000606.
Mora C, Myers RA, Coll M, et al. 2009. Management effectiveness of
the world’s marine fisheries. PLoS Biol 7: e1000131.
Osenberg CW, Sarnelle O, and Cooper SD. 1997. Effect size in ecolog-
ical experiments: the application of biological models in meta-
analysis. Am Nat 150: 798–812.
R Core Team. 2016. R: a language and environment for statistical com-
puting. Vienna, Austria: Foundation for Statistical Computing.
Sala E and Giakoumi S. 2018. No- take marine reserves are the most
effective protected areas in the ocean. ICES J Mar Sci 75: 1166–68.
Sandin SA, Smith JE, DeMartini EE, et al. 2008. Baselines and degrada-
tion of coral reefs in the northern Line Islands. PLoS ONE 3: e1548.
Sciberras M, Jenkins SR, Mant R, et al. 2013. Evaluating the relative
conservation value of fully and partially protected marine areas.
Fish Fish 16: 58–77.
Supporting Information
Additional, web-only material may be found in the online
version of this article at http://onlinelibrary.wiley.com/
doi/10.1002/fee.1934/suppinfo