The inclusion of biodiversity in environmental impact assessment: Policy-related progress limited by gaps and semantic confusion

Abstract

Natural habitat loss and fragmentation, as a result of development projects, are major causes of biodiversity erosion. Environmental impact assessment (EIA) is the most commonly used site-specific planning tool that takes into account the effects of development projects on biodiversity by integrating potential impacts into the mitigation hierarchy of avoidance, reduction, and offset measures. However, the extent to which EIA fully address the identification of impacts and conservation stakes associated with biodiversity loss has been criticized in recent work. In this paper we examine the extent to which biodiversity criteria have been integrated into 42 EIA from 2006 to 2016 for small development projects in the Montpellier Metropolitan territory in southern France. This study system allowed us to question how EIA integrates biodiversity impacts on a scale relevant to land-use planning. We examine how biodiversity inclusion has changed over time in relation to new policy for EIA and how the mitigation hierarchy is implemented in practice and in comparison with national guidelines. We demonstrate that the inclusion of biodiversity features into EIA has increased significantly in relation to policy change. Several weaknesses nevertheless persist, including the continued absence of substitution solution assessment, a correct analysis of cumulative impacts, the evaluation of impacts on common species, the inclusion of an ecological network scale, and the lack of monitoring and evaluation measures. We also show that measures for mitigation hierarchy are primarily associated with the reduction of impacts rather than their avoidance, and avoidance and offset measures are often misleadingly proposed in EIA. There is in fact marked semantic confusion between avoidance, reduction and offset measures that may impair stakeholders' understanding. All in all, reconsideration of stakeholders routine practices associated with a more strategic approach towards impact anticipation and avoidance at a land-use planning scale is now necessary for the mitigation hierarchy to become a clear and practical hierarchy for “no net loss” objectives based on conservation priorities.

Full text

Research article
The inclusion of biodiversity in environmental impact assessment:
Policy-related progress limited by gaps and semantic confusion
Charlotte Bigard
a
,
c
,
*
, Sylvain Pioch
b
, John D. Thompson
a
a
UMR 5175 Centre d'Ecologie Fonctionnelle et Evolutive, CNRS, 1919 route de Mende, 34293, Montpellier Cedex 5, France
b
UMR 5175, Centre d'Ecologie Fonctionnelle et Evolutive, Universit!
e de Montpellier - Universit!
e Paul-Val!
ery Montpellier, Route de Mende, 34199,
Montpellier Cedex 5, France
c
Montpellier M!
editerran!
ee M!
etropole, 50, place Zeus, CS 39556, 34961, Montpellier Cedex 2, France
article info
Article history:
Received 6 February 2017
Received in revised form
12 May 2017
Accepted 19 May 2017
Available online 26 May 2017
Keywords:
Environmental impact assessment
Mitigation hierarchy
Conservation science
Land use planning
No net loss
abstract
Natural habitat loss and fragmentation, as a result of development projects, are major causes of biodi-
versity erosion. Environmental impact assessment (EIA) is the most commonly used site-specific plan-
ning tool that takes into account the effects of development projects on biodiversity by integrating
potential impacts into the mitigation hierarchy of avoidance, reduction, and offset measures. However,
the extent to which EIA fully address the identification of impacts and conservation stakes associated
with biodiversity loss has been criticized in recent work. In this paper we examine the extent to which
biodiversity criteria have been integrated into 42 EIA from 2006 to 2016 for small development projects
in the Montpellier Metropolitan territory in southern France. This study system allowed us to question
how EIA integrates biodiversity impacts on a scale relevant to land-use planning. We examine how
biodiversity inclusion has changed over time in relation to new policy for EIA and how the mitigation
hierarchy is implemented in practice and in comparison with national guidelines. We demonstrate that
the inclusion of biodiversity features into EIA has increased significantly in relation to policy change.
Several weaknesses nevertheless persist, including the continued absence of substitution solution
assessment, a correct analysis of cumulative impacts, the evaluation of impacts on common species, the
inclusion of an ecological network scale, and the lack of monitoring and evaluation measures. We also
show that measures for mitigation hierarchy are primarily associated with the reduction of impacts
rather than their avoidance, and avoidance and offset measures are often misleadingly proposed in EIA.
There is in fact marked semantic confusion between avoidance, reduction and offset measures that may
impair stakeholders' understanding. All in all, reconsideration of stakeholders routine practices associ-
ated with a more strategic approach towards impact anticipation and avoidance at a land-use planning
scale is now necessary for the mitigation hierarchy to become a clear and practical hierarchy for “no net
loss”objectives based on conservation priorities.
©2017 Elsevier Ltd. All rights reserved.
1. Introduction
Natural habitat destruction by development projects (e.g. linear
infrastructures, urbanisation, commercial centres, quarries, etc.)
has continued to cause the loss of genetic and species diversity, the
fragmentation of natural habitats and the degradation of ecosystem
function (Fahrig, 2003; McKinney, 2008; MEA, 2005). Many coun-
tries have thus developed instruments that attempt to ensure a «no
net loss »(henceforth NNL) of biodiversity with measures to
attenuate and mitigate the loss of biodiversity in the face of land
development (Bull et al., 2016; Hassan et al., 2015; Maron et al.,
2016). The development of the NNL paradigm, and its application
in land-use planning, has however encountered difficulties due to
inconsistencies in the way its underlying concepts are framed
(Apostolopoulou and Adams, 2015; Bull et al., 2016; Gordon et al.,
2015) and how impacts are compared with a baseline to assure
NNL (Bull et al., 2014; Maron et al., 2016, 2015). Indeed, in practice,
NNL appears to be impossible, there is nearly always some form of
*Corresponding author. UMR 5175 Centre d'Ecologie Fonctionnelle et Evolutive,
CNRS, 1919 route de Mende, 34293, Montpellier Cedex 5, France. Tel.: þ33
0467613345.
E-mail addresses: charlotte.bigard@cefe.cnrs.fr (C. Bigard), sylvain.pioch@gmail.
com (S. Pioch), john.thompson@cefe.cnrs.fr (J.D. Thompson).
Contents lists available at ScienceDirect
Journal of Environmental Management
journal homepage: www.elsevier.com/locate/jenvman
http://dx.doi.org/10.1016/j.jenvman.2017.05.057
0301-4797/©2017 Elsevier Ltd. All rights reserved.
Journal of Environmental Management 200 (2017) 35e45

decline in biodiversity - a sort of generalised net loss impossible to
avoid, but never explicitly presented (Aronson and Moreno-
Mateos, 2015; Maron et al., 2012; Moreno-Mateos et al., 2015). In
relation to these difficulties, many countries have developed two
main instruments to apply NNL policy in their land-use planning
procedures.
The first of these instruments concerns Environmental Impact
Assessment (henceforth EIA) that developed during the 1970's to
become a key instrument in site-specific planning for biodiversity
(Mandelik et al., 2005) and environmental management (Morgan,
2012). EIA contributes to the assessment and anticipation of
development projects and their impacts on environment and to the
adoption of pro-active policy to mitigate the impacts of such pro-
jects. However, many authors have pointed out recurrent weakness
in the identification of impacts and the conservation stakes asso-
ciated with biodiversity and landscape ecological context (Byron
et al., 2000; Drayson et al., 2015; Gontier et al., 2006; Thompson
et al., 1997; Treweek and Thompson, 1997). EIA has also been
criticised because choices among alternative options for develop-
ment projects are more often based on socio-economic consider-
ations than on ecological arguments (Bonthoux et al., 2015), the
delimitation of the area used to assess impacts is often made on a
non-ecological basis (Geneletti, 2006), measurable indicators or
quantitative predictions are rarely used (Mandelik et al., 2005;
Samarakoon and Rowan, 2008), and the relevance of an impact is
unclear (Atkinson et al., 2000; Khera and Kumar, 2010). In addition,
the study scope is often poorly defined or too narrow; many studies
only assess biodiversity in terms of species' populations with little
attention paid to understanding of effects on ecological processes,
ecosystem function or genetic variation (Atkinson et al., 2000;
Gontier et al., 2006; Khera and Kumar, 2010). Finally, an absence
of precise definitions and correct understanding of ecological pro-
cesses makes the identification of what represents a “significant”
impact difficult (Briggs and Hudson, 2013; Geneletti, 2006).
EIA provides basic information for the identification of NNL
objectives within the context of a second major policy instrument,
the so-called mitigation hierarchy. This hierarchy provides a policy
framework to identify the process by which environmental impacts
from development can be “avoided”, unavoidable impacts
“reduced”, and residual impacts “offset”(Maron et al., 2016). This
mitigation hierarchy has also become a subject of concern in terms
of its environmental efficiency, social implications and ethical basis
(Gobert, 2015; Gordon et al., 2015; Levrel et al., 2015; Lucas, 2009;
Maron et al., 2016; Moreno-Mateos et al., 2015). Despite high sci-
entific tractability, it begets only moderate implementation trac-
tability, and clear-cut rules on how to classify certain impacts
within the mitigation hierarchy barely exist (Martin, 2015; Bull
et al., 2016; Maron et al., 2016). In addition, the common reliance
on offsetting to achieve NNL has received serious criticism due to
the fact that offsets are rarely adequate, complete offsetting may be
illusory due to the complexity of ecological processes (Gardner
et al., 2013; Moreno-Mateos et al., 2015) and weak institutional
organisation of the mitigation hierarchy impairs attempts to ach-
ieve NNL (Jacob et al., 2015; Lucas, 2009). Problems associated with
identifying ecological equivalence and the absence of a systematic
regional approach further undermine the efficiency of the mitiga-
tion hierarchy (Habib et al., 2013; Kujala et al., 2015).
The objective of this study is to examine how biodiversity is
integrated into EIA and defined and treated in the mitigation hi-
erarchy. We examine this issue in relation to recent changes in
French policy aimed at improving the EIA procedure and the
implementation of the mitigation hierarchy. In this context, our
study addresses four main questions. First, how are impacts on
biodiversity taken into account in a large sample of EIAs, all elab-
orated within a single territory? Second, is there a significant effect
of new policy that proposes to make a more detailed analysis of
biodiversity features and their inclusion in EIA? Third, how are
cumulative impacts taken into account in the study area? Finally,
how well do measures proposed in the EIA for the different ele-
ments in the mitigation hierarchy fit French national guidelines and
definitions of the mitigation hierarchy?
2. Methods
2.1. Case study
To undertake this study we analysed 42 EIAs associated with
projects in a single territory, that of the Montpellier Metropolitan
Territory (31 municipalities) and nine adjoining municipalities in
southern France (Fig. 1). This form of territorial grouping allows the
different local municipalities to mutualise their objectives and
obligations (waste treatment, sanitation, economic development
…) and to develop coherent urban land-use planning strategies.
The territory contains a patchwork of semi-natural Mediterranean-
type habitats rich in biodiversity, various agricultural areas and is
one of the fastest developing metropolitan territories in France.
The 42 EIAs we studied represent a large number of small-scale
projects each of which has impacts primarily on common species
and habitats and, to a lesser extent on protected habitats and
species. The EIAs for the 42 projects were elaborated between 2006
and 2016. Two major infrastructure projects that had EIA docu-
ments elaborated during this time period were not used in the
initial analyses because their impact concerned several munici-
palities and different types of ecosystem. Hence, the amount of
money and time invested in the EIA productionwas way above that
of all the other 42 projects. The two infrastructure projects are thus
not comparable with the 42 small-scale projects. We thus only used
the information in these two EIAs in the analysis of cumulative
impacts on biodiversity (see below). Thirty-nine of the develop-
ment projects are small-scale development zones or housing pro-
jects, there is one photovoltaic solar power plant project and two
short sections of local road construction. The EIA of each project
was obtained from the archives of the State environmental agency
in the study region (DREAL), the authority in charge of examining
EIAs. They represent all the available EIAs that have caused irre-
versible impacts on terrestrial natural habitats in the study region.
2.2. A data base to examine biodiversity inclusion in EIA
We conducted a systematic examination of the extent to which
biodiversity is included in each of the 42 EIAs. To do so we analysed
six criteria, or questions, that reflect the organisation of the
different chapters of an EIA (Table 1). The first criterion concerns a
“baseline”description of the impacted zone in terms of species and
habitats present, ecological networks, ecological equilibria and
ecological interactions. The second involves how “data”are
collected and their pertinence. The third concerns a description of
the “impacts”which may be positive or negative, direct or indirect,
temporary or permanent and can be cumulative with those in other
development projects. The fourth requires an assessment of alter-
native (“substitution”) solutions and a test of the compatibility with
existing planning documents. The fifth involves descriptions of the
necessary “measures”that are proposed for implementation within
the mitigation hierarchy. The sixth criterion relates to propositions
for “monitoring and evaluation”. To provide quantitative and
qualitative response data in relation to these questions, 32 in-
dicators concerning how biodiversity is included in an EIA were
developed (Table 1). These indicators were developed in order to
encompass what the French policy reform and the national doc-
trine require in terms of biodiversity inclusion in EIAs.
C. Bigard et al. / Journal of Environmental Management 200 (2017) 35e4536

Each indicator is noted with a score of 1 or 0, depending on
whether the response is positive (inclusion of biodiversity) or
negative respectively. The sum of the scores for each indicator was
determined in order to examine how EIAs integrate biodiversity. To
do so, an “Index of Biodiversity Inclusion”(IBI), adapted from
Atkinson et al. (2000), was calculated. IBI calculation is based on the
number of positive answers (P) relative to the total number (N) of
questions (32 for EIAs involving offset measures, 30 for the others):
i.e. IBI ¼P/N.
We tested whether the adoption of new policy, aimed at
reforming the procedure for EIAs and the mitigation hierarchy in
France, has had an impact on the integration of criteria to more fully
assess impacts on biodiversity and measures for the mitigation
hierarchy. This policy came with the law n
#
2010-788 published on
the 12th July 2010 relative to national commitment for the envi-
ronment, with the application of the decree n
#
2011e2019 of
December 2011 and put into force in June 2012. The main changes
introduced by this reform concern the need to enlarge the scope of
EIA for all projects that may have a significant impact on the
environment, the requirement of propositions for measures for
implementation within the mitigation hierarchy, an evaluation of
cumulative impacts, and the necessity of a monitoring plan and
environmental compliance (Qu!
etier et al., 2014). To examine the
effect of this policy we tested if IBIs for the 21 EIAs that were made
after June 2012 were greater than the 21 EIAs made before June
2012. Simply by chance the number of pre and post-June 2012 EIAs
was the same. To do so, we did a one-tailed non-parametric Wil-
coxon test (alternative “greater”) due to the non-normal distribu-
tion of data (result of Shapiro-test not shown). We then focused on
each indicator separately, and tested for an increase of positive
responses after June 2012 by comparing changes in the ratio of
positive to negative responses with a Fisher exact test. All statistical
analysis were made using R statistical software (R development
core team, 2016).
We also analysed the relationship between responses of pre-
dictive variables and the IBI of each EIA, in order to understand
which variables are linked to biodiversity inclusion in EIAs
(Appendix 2). The predictive variables tested are composed of four
qualitative variables: the involvement of expert naturalists in the
EIA, the need for offset measures, the need for an authorisation to
destruct the habitat of protected species and the type of habitat
impacted (for six different habitat types - woodland, cultivated
land, post-cultural semi-open habitat undergoing secondary suc-
cession to scrubland and woodland, wetland, Mediterranean gar-
rigues, and heathland and scrubland), and two quantitative
variables (the surface area of the development project and the
number of pages of the EIA dedicated to natural environment
issue). Environment impacts were taken into account if more than
25% of the study area is concerned. To test for a relationship be-
tween IBI and the qualitative predictive variables, we used
nonparametric one-sided Wilcoxon tests, and compared the IBI
score of EIAs presenting either a negative or positive response to
each qualitative variable. A linear regression was used to test for a
significant relationship between the two quantitative predictive
variables and IBI scores.
To test for the cumulative impacts of the 42 development pro-
jects in the studied territory (Hawke, 2009) we examined the
spatial distribution of projects and the number of projects that have
Fig. 1. Spatial distribution of the 42 EIAs elaborated in and around the Montpellier Metropolitan territory in France from 2006 to 2016.
C. Bigard et al. / Journal of Environmental Management 200 (2017) 35e45 37

a moderate, high or very high impact on listed species using GIS
software. The level of impact is defined by expert judgment in the
EIA. We also quantified the number of projects that impact on each
of the listed species that incur impact in at least one EIA. The
identity of the listed species is provided in Appendix 1. They
correspond to priority species (from moderate to very high priority)
with a national protection status, listed in the European directives
and/or identified at a regional scale as “patrimonial”species.
Table 1
Criteria used to assess the inclusion of biodiversity in 42 EIAs in the Montpellier Metropolitan territoryfrom 20 06 to 2016. Criteria concern six questions that are assessed with
a total of 32 indicators. For each indicator, the number of positive responses (i.e., responses with a score of 1) before and after June 2012 are noted and significant differences
between EIAs done before and after 2012 tested with a Fisher exact test.
Criteria Question n
#
Indicator Number of positive responses
Before
June 2012
After
June 2012
Total number (%) Significance of increase
after June 2012
Baseline Is the baseline comprehensive
enough to provide a basis to
evaluate impacts?
1Definition of an area of effects
(different to study area)
3 12 15 (35.7%) **
2Study of a larger area than the
project boundaries
1 15 16 (38.1%) ***
3Expertise on all groups of species 4 20 24 (57.1%) ***
4Detailed inventory of flora in the
study area
7 18 25 (59.5%) **
5Detailed inventory of fauna in
study area
4 16 20 (47.6%) ***
6Description of natural habitats 10 17 27 (64.3%) ns
7Natural features totalised on a
map of the study area
5 12 17 (40.5%) ns
8Study of local ecological
connectivity
6 19 25 (59.5%) ***
9Study of regional ecological
connectivity
0 7 7 (16.7%) **
10 Study of ecosystems, species and
populations
0 7 7 (16.7%) **
11 Reference to/or study of
ecological interactions
0 0 0 (0%) 0
12 Reference to population
dynamics or studies
3 9 12 (28.6%) ns
13 Argumentation for inclusion of
common biodiversity
2 8 10 (23.8%) ns
Data Are data gathered in a reliable
way and correctly referenced?
14 Field trip 16 21 37 (88.1%) *
15 More than one season of
prospection
5 20 25 (59.5%) ***
16 Clear reference to database
employed
12 20 32 (76.2%) **
17 Consultation of the relevant
scientific literature
5 15 20 (47.6%) **
Impacts Are all the impacts explained
and properly evaluated?
18 Evaluation of the significance of
each impact
3 3 6 (14.3%) ns
19 Identification of direct and
indirect impacts
12 17 29 (69.0%) ns
20 Identification of temporary and
permanent impacts
15 17 32 (76.2%) ns
21 Description of possible
cumulative impacts
0 17 17 (40.5%) ***
22 Explanation of the method used
to evaluate impacts
4 14 18 (42.9%) **
Substitution Is there an attempt to avoid
impacts on natural
environments at the beginning
of the EIA?
23 Study of alternative solutions 0 4 4 (9.5%) ns
24 Study of the alternative “without
project”
1 0 1 (2.4%) ns
Measures Are the measures explained and
detailed enough to potentially
balance impacts?
25 Detailed description of mitigation
measures
6 19 25 (59.5%) ***
26 Distinction between each type of
measure
5 15 20 (47.6%) **
27 Use of a method to propose
offsets based on equivalence
0 0 0 (0%) N.A.
28 Reference to a time lag between
losses and future offsets
0 0 0 (0%) N.A.
Monitoring
and evaluation
Are ways to ensure success and
sustainability of measures
proposed?
29 Mention of success probability of
mitigation measures
0 2 2 (4.8%) ns
30 Scheduling of a monitoring-
evaluation programme
1 12 13 (31.0%) ***
31 Definition of indicators for
monitoring and evaluation
0 5 5 (11.9%) *
32 Mention of sustainability of the
measures
1 1 2 (4.8%) ns
ns, not significant; *p <0.05, **p <0.01, ***p <0.001.
C. Bigard et al. / Journal of Environmental Management 200 (2017) 35e4538

2.3. Attribution of measures to the mitigation hierarchy
Finally, we examined how the 42 EIAs propose measures within
the mitigation hierarchy relative to definitions for each level of the
hierarchy in recent national doctrine (MEDDE, 2012) and guidelines
(MEDDE, 2013) proposed by the French Ministry of Ecology. First,
avoidance measure are those that supress any impacts ahead of the
project development by the abandonment of the project, changes
in its perimeter or surface area, or the choice of a new site or use of
technical solutions. Second, reduction measures involve the
implementation of technical solutions to alleviate impacts during
construction and exploitation. Third, offset measures aim to
maintain or enhance biodiversity features that are impacted by a
project. These include ecological restoration and the recreation and
management of natural habitats, species communities, and
ecological networks and can thus include the reinforcement of
natural populations or their reintroduction. Finally, supporting
measures can be proposed to improve the efficiency or to ensure
the possible success of biodiversity offset initiatives and include
knowledge improvement, methodological development, etc.
We quantified the number of measures proposed within each
EIA for each of these four types of measure and re-evaluated and
recompiled the number of measures for each of these measures in
relation to definitions in the ministry doctrine. Then, we compared
the number of measures in each level as proposed by the EIA, with
the number of measures reclassified according to ministry defini-
tions and quantified the number of transitions among levels along
the hierarchy.
3. Results
3.1. Index of biodiversity inclusion and its temporal evolution
IBI ranged between 0.07 and 0.75 with a mean value of 0.38
(Appendix 2), i.e. on average a positive response was observed for
38% of the indicators per EIA. Fourteen EIAs (33%) had a very low IBI
(<0.2) and 15 EIAs (36%) had an index between 0.4 and 0.6. Only six
EIAs had indexes between 0.2 and 0.4, seven between 0.6 and 0.8
and none had an IBI >0.8. As a result, more than two thirds of the
EIAs showed either very little effort to integrate biodiversity
(IBI <0.2) or a higher than average IBI of 0.4e0.6.
These different groups showed a clear temporal difference in
their occurrence (Fig. 2;Appendix 2). We observed a significant
increase in values for the IBI after 2010 (Wilcoxon test: w ¼303,
p¼1.22e-5) and after 2012 (w ¼411.5, p ¼7.72e-7). EIAs conducted
before June 2012 had an average IBI of 0.21 and those conducted
post-June 2012 had an average IBI of 0.55. The former test should be
viewed with caution given the small number (n ¼10) of EIAs prior
to 2010. The criteria that contribute to this increase are mentioned
below.
3.2. Inclusion of biodiversity criteria and indicators
In terms of naturalist expertise on biodiversity, we found that
the inclusion of expert naturalist advice in a specific section of the
EIA represents a major contribution to the IBI (W ¼329, p<0.001).
Naturalist expert advice on the fauna, flora and habitats impacted
by a project was present in roughly 50% of the EIAs. In 25 EIAs (61%)
the spatial area of impacts due to the development project was not
clearly defined. When the studied area was described as going
beyond the current perimeter of the project, the limits were usually
defined on the basis of land-use borders (roads, field boundary, etc.)
or with a buffer zone with an arbitrary width. In none of the EIAs
was there evidence of an attempt to assess this area on the basis of
the knowledge of species present in the zone or the functional
characteristics of the local ecosystem. Field studies to provide up-
to-date information were made in 37 EIAs (90%) but in only 25
(60%) of these fauna and flora were prospected in more than one
season (Table 1), even though two seasons are a minimum under
the Mediterranean climate due to the marked seasonal contrast
that impacts on biodiversity in this region (Thompson, 2005).
Nineteen out of 21 EIAs (90%) produced after June 2012 involved
surveying in more than one season and provide references of da-
tabases employed. Most of the studies provided clear information
on the databases used to make the EIA and 20 out of 42 (48%)
contain scientific references. In all 21 of the EIAs conducted after
June 2012 a large range of taxonomic groups were analysed,
whereas only four of the 21 EIAs conducted before June 2012 had
such information. Only seven EIAs (~20%) took into account impacts
on all three scales of biodiversity (ecosystem function, species di-
versity and genetic variation). Ecological interactions were not
included in any of the EIAs. In 12 of the 42 EIAs (~25%) there were
either references to population dynamics or there was a presen-
tation of impacts on “ordinary nature”(species or habitats without
a protection status or a specific regional stake).
Statistical analyses showed that the size of a development
project does not lead to a higher IBI and we detected no difference
in the IBI for projects impacting different types of habitats. The
larger the part of the EIA dedicated to the natural environment
(based on the number of pages in the EIA), the higher the IBI ob-
tained (Linear regression: F¼2.6948, p<0.001). The significance of
impacts was only defined and evaluated in six EIAs (14%). In 27 EIAs
(64%) the nature of the impacts was detailed as being either direct
or indirect and either temporary or permanent. The necessity of an
authorisation to destroy protected species and their habitats or the
necessity of offsetting significant residual impacts of the project,
also contributed significantly to the IBI (W¼178, p<0.001 and
W¼264, p<0.001 respectively).
We observed a significantly higher number of positive responses
concerning the inclusion of impacts on the local ecological network
after June 2012 (Table 1), i.e. in 21 EIAs, 19 of which were conducted
post-June 2012. Impacts at the scale of the regional ecological
network were only assessed in seven EIAs (20%), all of which were
Fig. 2. Index of biodiversity inclusion (IBI) as a function of EIA submission date and
temporal benchmarks for the EIA reform law n#2010-788 of July 2010 (dotted line)
and the EIA reform decree n#2011e2019 of December 2011 implemented from June
2012 (dashed line).
C. Bigard et al. / Journal of Environmental Management 200 (2017) 35e45 39

conducted after June 2012.
Cumulative impacts were assessed in 17 of the 21 EIAs published
after 2O12 and absent from all EIAs published prior to 2012 (Fig. 3).
In 2012, the three EIAs that identified a cumulative impact all occur
after June 2012. Basically, cumulative impact assessment involved
taking into account known projects spatially close to the project
under EIA, with an identification of whether individual species are
impacted in the neighbouring projects. In EIAs, this assessment can
range from a simple expert judgment to a further analysis of
impacted species, based on field ecological valuation. Among the 17
EIAs which mentioned cumulative impacts, 12 included an
assessment of their impacts on listed species, nine detected cu-
mulative impacts due to the project, but only two of these explicitly
proposed to take them into account in the mitigation measures. As
a result, although projects assessed after 2012 more rigorously
described cumulative impacts, there was a lack of proposed action
to precisely quantify such impacts and propose adequate measures
within the mitigation hierarchy.
To examine cumulative impacts on listed species, the two
infrastructure projects that cross the territory from East to West
were added to the study of cumulative impacts. We found that 19
(20%) species are impacted by a single project, 37 species (38%) are
impacted by two to three projects and 41 species (42%) are
impacted by more than three projects, with a maximum of 20
projects impacting one species (Fig. 4). It should also be noted that
the number of projects that impact the study species is clearly
underestimated; most EIAs (60%) do not refer to cumulative im-
pacts and the older EIAs do not propose a complete study of species,
natural habitats and ecological functions.
Finally, only four EIAs (10%) studied alternative solutions with
criteria on the natural environment and none of the 42 EIAs studied
alternative solutions “without the project”(Table 1). Socio-
economic arguments relating to the need for accommodation or
employment and coherence with urban planning documents were
the primary reasons used to justify choices made for the project.
Mitigation measures were described and distinguished from one
another in roughly 50% of the EIAs. EIAs conducted post-June 2012
were significantly clearer about propositions than pre-June 2012
EIAs an important point for the results presented below. For the
nine EIAs that proposed offset measures, methods based on
equivalence between losses and gains were never used and none of
the EIAs referred to a time lag between destruction and offset
measures. Proposals for monitoring and evaluation were provided
in only 13 EIAs (30%) and in these EIAs only one out of three pro-
vided ecological indicators with which to evaluate and monitor the
benefits of mitigation measures. There is almost no mention about
how to assess the success and sustainability of the mitigation
measures.
3.3. Attribution of measures to the mitigation hierarchy
For the 42 EIAs analysed in this study, a total of 358 measures
were proposed for the different elements of the mitigation hier-
archy. The number of measures proposed in EIAs published after
June 2012 (n ¼243) was twice as high as in those published prior to
June 2012 (n ¼115). When we compared proposed measures with
ministry definitions for these different elements of the mitigation
hierarchy, we found that only 39% of the proposed measures fit the
definitions of the national guidelines (Fig. 5). Most of the proposed
measures for avoidance were in fact measures to reduce impacts
(42 out of 50 proposed avoidance measures). The five measures
that were truly avoidance measures involved a reduction of the
boundary of the project and landscaping. Almost all measures that
proposed a reduction in impacts fit the ministry definitions for a
reduction in the impact. For the 30 proposed offset measures, 11
were in fact measures that reduce impacts, and 17 were correctly
defined as offset measures. Lastly, after reclassification, two of the
proposed 37 supporting measures fit ministry definitions, 32 are
Fig. 3. The number of EIAs that include (dark part of histogram) or do not include
(white part of histogram) a reference to cumulative impacts on biodiveristy in relation
to publication of the EIA reform law (July 2010 - dotted arrow) and decree (december
2011 implemented in June 2012 - dashed arrow).
Fig. 4. Number of listed species impacted as a function of the number of EIAs in which
an impact is detected.
C. Bigard et al. / Journal of Environmental Management 200 (2017) 35e4540

reduction measures, one is an offset measure and the two
remaining measures involve monitoring. The majority of EIAs did
not propose “real”offset measures (only 9 out of 42 EIAs propose
offset measures). When present, offset measures showed little di-
versity. The most common measures are linked to a management
project to maintain open habitats that favour the presence of
Mediterranean listed species. It should also be noted that 35% of all
measures proposed in the 42 EIAs, were made in absence of a clear
statement about what type of measures were being proposed
(what we refer to as “not qualified”). When we analysed these
measures, we found that all such measures except one concerned a
reduction in the impact of the project (Fig. 5).
There was thus a major bias towards a weakening of the miti-
gation hierarchy when measures are compared with ministry
guidelines, primarily because avoidance is a less-used measure
than what propositions would suggest. A common feature of all
EIAs is that measures to reduce impacts were by far the dominant
type of proposition (Table 2,Fig. 5). In fact, only 5 of the 42 EIAs
provide for avoidance, prior to reduction measures (Table 2).
4. Discussion
For the Montpellier Metropolitan territory we have shown an
improvement in the inclusion of biodiversity indicators in the
framework of EIA that is correlated with the elaboration of a new
law (2010) and its application decree (2012). This policy reform
provides a legal framework to elaborate a more complete identifi-
cation of the biodiversity concerns (baseline approach), impacts on
species, habitats and ecological networks, and the cumulative im-
pacts on biodiversity. However, several important weaknesses
persist and there is semantic confusion concerning proposed
measures for the different elements of the mitigation hierarchy.
4.1. Improved but incomplete integration of biodiversity
The policy reform proposed that alternative solutions to projects
and their impacts be carefully examined prior to project
development. However, our analysis of 42 EIAs reveals that such
alternative solutions are rarely explored; only 5 EIAs out of the 42
propose true avoidance measures. As a result, avoidance, which is
supposedly the first element in the mitigation hierarchy, is rarely
employed. This is a critical result because it illustrates one reason
why “no net loss”is almost impossible to achieve. Basically,
biodiversity conservation occurs in a world where there is a back-
ground of generalised “net loss”(Maron et al., 2016; Moreno-
Mateos et al., 2015). This absence of a search for alternative op-
tions in the early phases of development projects is a clear indi-
cation of the priority for a systematic conservation planning
approach to the question of avoidance (and offset proposition) that
ensure more efficient biodiversity conservation at a territorial scale,
i.e. beyond the scale of individual projects (Kujala et al., 2015).
Although the identification of baseline information on the key
environmental issues in a site where a project occurs has been
more completely assessed in EIAs since the policy reform, several
important issues are open for improvement. For instance, the
definition of the study area (beyond the area directly impacted) is
in most cases made on an arbitrary basis instead of being made on
the basis of species, habitats and functional characteristics of the
local ecosystem. The policy reform produced a three-fold increase
in the inclusion of local ecological continuities in EIAs and stimu-
lated a small number of studies on regional ecological networks
(absent from all EIAs prior to June 2012). However, the global
ecological network and ecosystem levels are still rarely considered
and analysed in EIAs. Moreover, as illustrated elsewhere (Atkinson
et al., 2000; Gontier et al., 2006; Regnery et al., 2013a), we found
that the main focus is on listed species and habitats with less in-
terest in common species and habitats. In addition, we have shown
that the presence of listed species has a significant positive effect on
IBI, i.e., their presence partly conditions the quality of the EIA. The
problem here is that common habitats and species play a major role
as a part of the habitat or landscape used by listed species (Elliott
and Whitfield, 2011; Gaston and Fuller, 2008) and in terms of
provision of ecological services (Tardieu et al., 2015).
As recommended by the policy reform, cumulative impacts have
Fig. 5. Total number of proposed measures (in the 42 EIAs) for each category of the mitigation hierarchy (in left) and how these numbers change (in right) once the proposed
measures are reclassified in relation to definitions of the national guidelines following the EIA reform law of 2010 (A- Avoid, R- Reduce, O-Offset, S-Support, M-Monitor, NQ- Not
qualified in the proposition).
C. Bigard et al. / Journal of Environmental Management 200 (2017) 35e45 41

been increasingly identified in EIAs. However, their analysis and
quantification remain superficial; the majority of EIAs that refer to
cumulative impacts do not propose adequate measures to reduce
them within the mitigation hierarchy. This is despite our finding
that cumulative impacts are a common feature for listed species in
the study area. The accumulation of impacts by numerous, small
and isolated projects, that individually may have minor impacts on
biodiversity in comparison with large individual projects, leads to
important cumulative impacts on listed species. This represents a
second major reason why the objective of no net loss remains
practically impossible to currently achieve within the mitigation
hierarchy. The issue of how to correctly assess cumulative impacts
raises several questions for the scientific community working on
the efficiency of no net loss policy and the mitigation hierarchy
(Halpern and Fujita, 2013; Kiesecker et al., 2010; Tallis et al., 2015).
Hence, there is a need for a methodology to assess such impacts in
order to go beyond the “first come, first served”logic that unfor-
tunately persists (Qu!
etier et al., 2014). One methodology to include
cumulative impacts within land use planning could be a form of
strategic environmental assessment (Whitehead et al., 2016).
In addition, the capacity of EIA to take into account environ-
mental issues is directly linked to the issue of determining what
significant impacts actually represent (George, 1999). In our study
we found that the true significance of impacts is rarely defined or
explicitly addressed (i.e. in only six EIAs). This is probably because
the identification of “significant”impacts remains difficult due to
the lack of clarity about how to define such impacts (Geneletti,
2006) and a lack of ecological details in the baseline study, poor
understanding of ecological process and a lack of monitoring and
feedback (Briggs and Hudson, 2013).
Our study illustrates the paucity of monitoring and evaluation
measures, despite the fact that half of the post-June 2012 EIAs
include a schedule for their implementation. This lack of feedback
on the true nature and extent of impacts and the efficiency of
mitigation measures is a major concern (Briggs and Hudson, 2013;
Curran et al., 2014). The development of such feedback, including
negative results, could allow environmental managers to propose
more feasible and efficient measures. Reviews of wetland restora-
tion experiments (Benayas et al., 2009; Curran et al., 2014; Maron
et al., 2012; Moreno-Mateos et al., 2012) have shown that biodi-
versity equivalency between restored areas and reference areas is
rarely, if ever, reached, and that there are major limits to the
effectiveness of restoration action as a result of time-lags, uncer-
tainty and the measurability of success. Maron et al. (2012) also
argue that restoration action will attain no net loss only when
impacted ecosystem values can be measured, when results about
restoration trials already exist to evaluate their feasibility, and
when time-lag and uncertainty (ecological risk) are assessed and
clarified in the “loss compared to gain”equation. Results associated
with ecological restoration should thus be examined with caution
(Benayas et al., 2009; Curran et al., 2014; Maron et al., 2012;
Moreno-Mateos et al., 2012; Palmer and Filoso, 2009). All in all, a
wider use of avoidance and reduction measures is a necessity that
can no longer be brushed under the carpet.
Finally, according to the EIA proportionality principle
1
promul-
gated in the policy reform, one would expect a larger project in
terms of surface area to have a higher index in terms of biodiversity
inclusion because the impact is higher. However, this is not the case
in our study. Also, we found no evidence that projects impacting
Table 2
Numbers of measures proposed and reclassified within the mitigation hierarchy (A- Avoid, R- Reduce, O-Offset, S-Support, M-Monitor, NQ- Not qualified) for each EIA
published pre-June 2012 (21 projects with a total of 115 measures) and post- June 2012 (21 projects with a total of 243 measures).
1
Principle that establishes a link between the size or the level of impact, as a
justification for the intensity, and the requirement level of precision needed to
assess the environmental impact (European Directive 2014/52/UE).
C. Bigard et al. / Journal of Environmental Management 200 (2017) 35e4542

semi-natural and typical Mediterranean habitats have a higher IBI
than those subject to human activities.
4.2. Blurred semantics
The reform policy is associated with a clarification of the nature
of the different measures for the mitigation hierarchy (avoid,
reduce or offset) and the modalities for monitoring and evaluation
of these measures. Our study reveals critical gaps in terms of both
incorporating a functional and wider-scale approach for biodiver-
sity integration in land-use planning and how actors understand
the true meaning of the different elements of the mitigation hier-
archy. Bull et al. (2016) previously identified this kind of ambiguity
and the lack of clarity concerning the concept of biodiversity offsets
in relation to no net loss objectives. Our study confirms this point
and provides a quantification of the types of confusion concerning
the different elements of the mitigation hierarchy.
For the 42 projects we studied, 61% of the proposed measures are
not correctly described in terms of their place in the mitigation hi-
erarchy and, after reclassification by comparison with the national
guidelines in the reform doctrine, it turns out that almost 90%
concern a reduction of impacts. Moreover, very few true offset
measures are suggested. For supporting measures, there was a major
confusion of what such measures represent. This kind of measure
should concern action to improve knowledge (research, experi-
mental project) and methods, to implement a larger-scale conser-
vation strategy, and/or to delimit protected areas, all of which
should contribute to improve the effectiveness of offset measures.
But almost all supporting measures were, according to ministry
definitions, measures that directly relate to impact reduction. So, in
practice, a reduction in impact is by far the most common measure
and after reclassification all EIAs have at least one reduction mea-
sure. Indeed, the 42 reduction measures proposed in EIAs as
avoidance measures do not supress impacts on natural environment
features, they simply minimize them. For instance, propositions (for
avoidance in some EIAs) based on the adaptation of the construction
schedule for species or the maintenance of ecological network fea-
tures in the impacted site do not avoid or supress impacts, they
reduce them. Only a change in the project perimeter or its reduction
so as not to impact the identified species or features of the ecological
network would represent true avoidance. Likewise, “nesting box
installation in the development project site”or “plantation of native
flora for green areas”are proposed in EIA as offset measures, but in
fact, they attenuate (i.e. reduce) impacts. The second “step”in the
mitigation hierarchy is thus more common than alternative solu-
tions, avoidance is rarely proposed.
The semantic confusion in the definitions of avoidance, reduc-
tion, offset and supporting measures may stem from a lack of un-
derstanding of ministry guidelines, or from the technical and
economical facility of implementing reduction measures rather
than searching to avoid impacts or implement offset measures that
are often more expensive and more constraining for the developer.
This issue could be resolved by the formation of consulting agencies
that elaborate EIA and clearer explanation of what different mea-
sures actually are in terms of the mitigation hierarchy. Enhance-
ment of regulatory agency control, through standardisation of
methods, could also limit this problem. Such options could ho-
mogenize and reduce misunderstanding among stakeholders,
optimize decision-making in terms of biodiversity conservation,
and improve the IBI scores of EIAs. For true avoidance measures to
be proposed, developers and experts should be in contact from the
very beginning of the project conception to make changes in the
project boundaries and its global form that avoid impacts. Hence, a
real anticipation of where and what to avoid remains a critical step
towards making mitigation measures more efficient in terms of
biodiversity conservation (Kareksela et al., 2013; Kujala et al., 2015;
McKenney and Kiesecker, 2010; Regnery et al., 2013b).
Finally, the proposition of mitigation measures does not ensure
their practical implementation and success in the field, especially
during the construction and operation phases. Hence, there is a dire
need for monitoring measures that assess any positive effects of
reduction and offset measures on biodiversity conservation. Offsets
should be seen as a last resort solution, with more emphasis and an
accurate focus on avoidance and reduction measures. Otherwise no
net loss will remain a lost cause.
5. Conclusion: towards a territorial-scale analysis
Progress on EIA mitigation propositions reflects a shift in
approach to biodiversity conservation. Gontier et al. (2006) pro-
posed three scales of approach to characterize biodiversity inclu-
sion in EIAs: (i) an approach focused on single sites or single
biodiversity element with no general overview, (ii) a functional and
dynamic ecosystem approach and (iii) a habitat suitability
approach based on processes. The policy reform in France recom-
mend an approach focused on the “natural habitats, animal and
vegetal species,ecological continuities,biological balance, ecological
functions, physical and biological features that are the support of
former elements and services provided by ecosystems”(MEDDE,
2012). This has stimulated a move towards a patchwork and
habitat suitability approach in which ecological connectivity and
cumulative impacts are targeted. However, despite the evolution of
such guidelines and their ambitions, the gap between EIA
commitment and practice thus persists. This result highlights the
dilemma discussed by Calvet et al. (2015) in which the higher the
ecological complexity, the more difficult it is to achieve ecological
equivalency and no net loss.
Our study also illustrates the pertinence of a territorial-scale
assessment of impacts on biodiversity in order to more efficiently
contribute to no net loss. In our study area, a correct assessment of
cumulative impacts is absent from land-use planning and impacts
on ecological networks are only partially addressed. Hence, the
development of a territorial strategy that shifts from an approach
based on treating “symptoms”at the scale of individual projects to
a more preventive approach focused on the avoidance of biodi-
versity loss and mitigation of cumulative impacts is now necessary.
In this context, Strategic Environmental Assessment, a tool that
assesses the impacts of policies, plans and programs (Wood and
Djeddour, 1989), could be used as an instrument to help formu-
late a proactive and more strategic approach in the early stages of
the decision-making processes (Bina, 2007; Partidario, 2015). For
such reasons, Strategic Environmental Assessment represents an
ideal tool to anticipate for avoidance in order to render mitigation
measures a true hierarchy based on priorities.
Acknowledgements
We thank the DREAL Occitanie for access to their archives and
staff at the Montpellier Mediterran!
ee M!
etropole for providing advice
and for their confidence. We thank Perrine Gauthier for comments
on a preliminary version of the manuscript and Guillaume Papuga
for his help and encouragement. This work was carried out with a
PhD grant awarded to Charlotte Bigard from the National Associa-
tion for Research and Technology (ANRT) and funded by the Min-
istry for Higher Education and Research and Montpellier
Mediterran!
ee M!
etropole.
C. Bigard et al. / Journal of Environmental Management 200 (2017) 35e45 43

Appendix 2. Details of EIAs published pre-June 2012 (n ¼21) and post- June 2012 (n ¼21) with their overall IBI based on the
relative number of positive (P) and negative (N) responses to the different criteria (n ¼30 or 32 depending on whether offset
measures are necessary). Project types concern economic development or housing (EDH), infrastructures (INF) and one
photovoltaic solar power plant (PSP). The types of environment impacted are cultivated land (C), post-cultural fields (P), garrigue
(G), heathland and scrubland (HS), woodland (Wo), wetland (We) and urban zone (U).
EIAs IBI Number of responses Project
type
Submission
year
Surface
area (ha)
EIA number
of pages
Need for
offset
measures
Involvment of
expert naturalists
Environment impacted Need for authorisation
to destruct protected species
or habitats
PN C P G HS Wo We U
Pre_1 0.20 6 24 EDH 2006 15.8 150 no yes X no
Pre_2 0.23 7 23 EDH 2011 25 52 no yes X X X no
Pre_3 0.23 7 23 EDH 2011 26 167 no no X X no
Pre_4 0.50 15 15 EDH 2011 19 86 no yes X no
Pre_5 0.27 8 22 EDH 2011 5108 no yes X no
Pre_6 0.40 12 18 EDH 2011 30 117 no yes X X no
Pre_7 0.63 19 11 EDH 2011 582 no yes X X no
Pre_8 0.13 4 26 EDH 2011 985 no no X no
Pre_9 0.10 3 27 EDH 2010 13 185 no no X no
Pre_10 0.07 2 28 INF 2008 369 no no X X no
Pre_11 0.47 14 16 EDH 2010 16.7 129 no yes X no
Pre_12 0.07 2 28 EDH 2009 11 116 no yes X no
Pre_13 0.07 2 28 EDH 2010 21 89 no no X no
Pre_14 0.13 4 26 EDH 2009 9268 no no X X no
Pre_15 0.07 2 28 EDH 2010 7.8 153 no no X X no
Pre_16 0.07 2 28 EDH 2010 481 no no X X no
Pre_17 0.13 4 26 EDH 2009 35.6 63 no yes X X no
Pre_18 0.07 2 28 EDH 2009 8112 no no X no
Pre_19 0.17 5 25 INF 2008 10 88 no no X no
Pre_20 0.17 5 25 EDH 2008 6100 no no X no
Pre_21 0.20 6 24 EDH 2011 39 71 no yes X no
Mean 0.21 6.24 23.76 15.19 112.9
Post_1 0.75 24 8 EDH 2014 12.5 219 yes yes X X yes
Post_2 0.57 17 13 EDH 2013 60 2014 no yes X X no
Post_3 0.37 11 19 EDH 2014 10 89 no yes X no
Post_4 0.63 20 12 EDH 2014 13.5 194 yes yes X yes
Post_5 0.75 24 8 EDH 2016 16 yes yes X yes
Post_6 0.60 18 12 EDH 2013 112 323 no yes X X no
Post_7 0.73 22 8 EDH 2013 15 286 no no X X no
Post_8 0.56 18 14 EDH 2013 39 392 yes yes X no
Post_9 0.50 15 15 EDH 2013 13.7 120 no yes X X no
Post_10 0.56 18 14 EDH 2013 23.5 197 yes yes X X no
Post_11 0.43 13 17 EDH 2013 12.8 169 no yes X X no
Post_12 0.59 19 13 EDH 2013 17 132 yes yes X X yes
Post_13 0.44 14 18 EDH 2013 14.3 130 yes yes X no
Post_14 0.50 15 15 EDH 2012 24.5 176 no yes X X no
Post_15 0.69 22 10 EDH 2013 25 117 yes yes X X yes
Post_16 0.33 10 20 EDH 2012 29 241 no no X no
Post_17 0.53 17 15 PSP 2012 66 277 yes yes X no
Post_18 0.50 15 15 EDH 2012 5.5 101 no yes X X no
Post_19 0.47 14 16 EDH 2012 8.3 161 no yes X no
Post_20 0.50 15 15 EDH 2013 5.2 246 no yes X X no
Post_21 0.63 19 11 EDH 2012 4.24 172 no yes X no
Mean 0.55 17.14 13.71 25.10 274
Appendix 1. Listed species that are impacted in the studied EIAs.
Taxonomic group Species
Avifauna Upupa epops,Otus scops, Clamator glandarius, Circaetus gallicus, Burhinus oedicnemus, Lullula arborea, Coracias garrulus, Emberiza calandra, Milvus
migrans, Anthus campestris, Falco naumanni, Carduelis cannabina, Merops apiaster, Athene noctua, Phoenicurus phoenicurus, Lanius senator, Caprimulgus
europaeus, Tetrax tetrax, Burhinus oedicnemus, Gelochelidon nilotica, Galerida cristata, Muscicapa striata, Anthus pratensis, Sylvia undata, Strix aluco,
Passer montanus, Sylvia melanocephala, Saxicola rubicola, Lanius meridionalis, Bubulcus ibis, Egretta garzetta, Tyto alba, Circus pygargus, Saxicola
rubetra, Sylvia cantillans, Emberiza hortulana, Tachybaptus ruficollis, Oenanthe oenanthe, Sylvia hortensis
Reptile Malpolon monspessulanus, Chalcides striatus, Timon lepidus, Lacerta bilineata, Rhinechis scalaris, Psammodromus hispanicus, Psammodromus algirus,
Podarcis muralis, Tarentola mauritanica, Natrix maura, Anguis fragilis, Emys orbicularis
Mammal Miniopterus schreibersii, Pipistrellus pygmaeus, Pipistrellus kuhlii, Nyctalus leisleri, Rhinolophus ferrumequinum, Pipistrellus nathusii, Myotis blythii,
Erinaceus europaeus, Pipistrellus pipistrellus, Sciurus vulgaris, Hypsugo savii, Plecotus austriacus, Tadarida teniotis, Myotis capaccinii, Rhinolophus
hipposideros, Myotis myotis, Myotis emarginatus, Castor fiber
Amphibien Hyla meridionalis, Pelodytes punctatus, Pelophylax perezi, Lissotriton helveticus
Insect Saga pedo, Zerynthia polyxena, Zerynthia rumina, Cerambyx cerdo, Coenagrion mercuriale, Oxygastra curtisii, Roeseliana azami, Arcyptera brevipennis
vicheti, Macromia splendens, Ischnura pumilio, Euphydryas aurinia, Lycosa tarantula, Uroctea durandi, Zygaena rhadamanthus, Satyrium w-album,
Gomphus graslinii
Plant Allium chamaemoly, Anemone coronaria, Gagea granatelli, Leucojum aestivum, Mentha cervina, Tulipa sylvestris, Astragalus glaux, Isoetes duriei
C. Bigard et al. / Journal of Environmental Management 200 (2017) 35e4544

References
Apostolopoulou, Evangelia, Adams, William M., 2015. Biodiversity offsetting and
conservation: reframing nature to save it. Oryx 51 (01), 23e31.
Aronson, J., Moreno-Mateos, D., 2015. !
Etat des lieux sur les actions de restauration
!
ecologique. In: Restaurer La Nature Pour Att!
enuer Les Impacts Du
D!
eveloppement. Quae, pp. 162e170.
Atkinson, Samuel F., Bhatia, Sarika, Schoolmaster, F. Andrew, Waller, William T.,
2000. Treatment of biodiversity impacts in a sample of US environmental
impact statements. Impact Assess. Proj. Apprais. 18 (4), 271e282.
Benayas, J.M.R., Newton, A.C., Diaz, A., Bullock, J.M., 2009. Enhancement of biodi-
versity and ecosystem services by ecological restoration: a meta-analysis. Sci-
ence 325 (5944), 1121e1124. http://dx.doi.org/10.1126/science.1172460.
Bina,Olivia,2007.A critical reviewofthedominantlinesofargumentationonthe needfor
strategic environmental assessment. Environ.Impact Assess.Rev. 27 (7), 585e606.
Bonthoux, D., Branger, M., Dias, P., Kessler, M., Thoyer, S., Pariente, A., 2015. Note de
synth"
ese bibliographique - Notions «;d’int!
er^
et public majeur»et «d’absence de
solution alternative de moindre impact »dans le cadre de la s!
equence Eviter -
R!
eduire - Compenser. SupAgro - DREAL LR.
Briggs, Sam, Hudson, Malcolm D., 2013. Determination of significance in Ecological
Impact Assessment: past change, current practice and future improvements.
Environ. Impact Assess. Rev. 38, 16e25.
Bull, J.W., Gordon, A., Law, E.A., Suttle, K.B., Milner-Gulland, E., 2014. Importance of
baseline specification in evaluating conservation interventions and achieving
no net loss of biodiversity. Conserv. Biol. 28, 799e809.
Bull, Joseph W., Gordon, Ascelin, Watson, James E.M., Maron, Martine, 2016. Seeking
convergence on the key concepts in ‘no net loss’policy. J. Appl. Ecol. 53 (6),
1686e1693. http://dx.doi.org/10.1111/1365-2664.12726.
Byron, Helen J., Treweek, Joanna R., Sheate, William R., Thompson, Stewart, 2000.
Road developments in the UK: an analysis of ecological assessment in envi-
ronmental impact statements produced between 1993 and 1997. J. Environ.
Plan. Manag. 43 (1), 71e97.
Calvet, Coralie, Napol!
eone, Claude, Salles, Jean-Michel, 2015. The biodiversity off-
setting dilemma: between economic rationales and ecological dynamics. Sus-
tainability 7 (6), 7357e7378.
Curran, Michael, Hellweg, Stefanie, Beck, Jan, 2014. Is there any empirical support
for biodiversity offset policy? Ecol. Appl. 24 (4), 617e632. http://dx.doi.org/
10.1890/13-0243.1.
Drayson, K., Wood, G., Thompson, S., 2015. Assessing the quality of the ecological
component of english environmental statements.J. Environ.Manage.160, 241e253.
Elliott, M., Whitfield, A.K., 2011. Challenging paradigms in estuarine ecology and
management. Estuar. Coast. Shelf Sci. 94 (4), 306e314. http://dx.doi.org/
10.1016/j.ecss.2011.06.016.
Fahrig, Lenore, 2003. Effects of habitat fragmentation on biodiversity. Annu. Rev.
Ecol. Evol. Syst. 34 (1), 487e515.
Gardner, T.A., VON Hase, A., Brownlie, S., Ekstrom, J.M., Pilgrim, J.D., Savy, C.E.,
Stephens, R.T., Treweek, J., Ussher, G.T., Ward, G., Ten Kate, K., 2013. Biodiversity
offsets andthe challenge of achieving no net loss.Conserv. Biol.27 (6), 1254e1264.
Gaston, K.J., Fuller, R.A., 2008. Commonness, population depletion and conservation
biology. Trends Ecol. Evol. 23, 14e19.
Geneletti, Davide, 2006. Some common shortcomings in the treatment of impacts of
linearinfrastructureson naturalhabitat.Environ.ImpactAssess.Rev.26(3), 257e267.
George, Clive, 1999. Testing for sustainable development through environmental
assessment. Environ. Impact Assess. Rev. 19 (2), 175e200.
Gobert, J., 2015. Mesures compensatoires socio-environnementales et acceptation
sociale. In: Restaurer La Nature Pour Att!
enuer Les Impacts Du D!
eveloppement.
Quae, pp. 34e45.
Gontier, Mikael, Balfors, Berit, M€
ortberg, Ulla, 2006. Biodiversity in environmental
assessmentdcurrent practice and tools for prediction. Environ. Impact Assess.
Rev. 26 (3), 268e286.
Gordon, Ascelin, Bull, Joseph W., Wilcox, Chris, Maron, Martine, 2015. FORUM:
perverse incentives risk undermining biodiversity offset policies. J. Appl. Ecol.
52 (2), 532e537.
Habib, T.J., Farr, D.R., Schneider, R.R., Boutin, S., 2013. Economic and ecological
outcomes of flexible biodiversity offset systems. Conserv. Biol. 27, 1313e1323.
Halpern, B.S., Fujita, R., 2013. Assumptions, challenges, and future directions in
cumulative impact analysis. Ecosphere 4, 1e11.
Hassan, F., Levrel, H., Scemama, P., Vaissi"
ere, A.-C., 2015. Le cadre de gouvernance
am!
ericain des mesures compensatoires pour les zones humides. In: Restaurer
La Nature Pour Att!
enuer Les Impacts Du D!
eveloppement. Quae, pp. 45e58.
Hawke, A., 2009. Report of the Independent Review of the Environment Protection
and Biodiversity Conservation Act 1999.
Jacob, C!
eline, Qu!
etier, Fabien, Aronson, James, Pioch, Sylvain, Levrel, Harold, 2015.
Vers une politique française de compensation des impacts sur la biodiversit!
e
plus efficace: d!
efis et perspectives, vertigo, vol. 14. Num!
ero 3.
Kareksela, S., Moilanen, A., Tuominen, S., Kotiaho, J.S., 2013. Use of inverse spatial
conservation prioritization to avoid biological diversity loss outside protected
areas. Conserv. Biol. 27, 1294e1303. http://dx.doi.org/10.1111/cobi.12146.
Khera, Neeraj, Kumar, Ajay, 2010. Inclusion of biodiversity in environmental impact
assessments (EIA): a case study of selected EIA reports in India. Impact Assess.
Proj. Apprais. 28 (3), 189e200.
Kiesecker, Joseph M., Copeland, Holly, Pocewicz, Amy, McKenney, Bruce, 2010.
Development by design: blending landscape-level planning with the mitigation
hierarchy. Front. Ecol. Environ. 8 (5), 261e266.
Kujala, H., Whitehead, A.L., Morris, W.K., Wintle, B.A., 2015. Towards strategic off-
setting of biodiversity loss using spatial prioritization concepts and tools: a case
study on mining impacts in Australia. Biol. Conserv. 192, 513e521.
Levrel, H., Frascaria-Lacoste, N., Hay, J., Martin, G., Pioch, S., 2015. Restaurer la nature
pour att!
enuer les impacts du d!
eveloppement: Analyse des mesures compen-
satoires pour la biodiversit!
e. Editions Quae.
Lucas, Marthe, 2009. La compensation environnementale, un m!
ecanisme inefficace
"
a am!
eliorer. rjenv 34 (1), 59e68.
Mandelik, Y., Dayan, T., Feitelson, E., 2005. Planning for biodiversity: the role of
ecological impact assessment. Conserv. Biol. 19, 1254e1261.
Maron, Martine, Bull, Joseph W., Evans, Megan C., Gordon, Ascelin, 2015. Locking in
loss: baselines of decline in Australian biodiversity offset policies. Biol. Conserv.
192, 504e512.
Maron, Martine, Hobbs, Richard J., Moilanen, Atte, Matthews, Jeffrey W.,
Christie, Kimberly, Gardner, Toby A., Keith, David A., Lindenmayer, David B.,
McAlpine, Clive A., 2012. Faustian bargains? Restoration realities in the context
of biodiversity offset policies. Biol. Conserv. 155, 141e148.
Maron, M., Ives, C.D., Kujala, H., Bull, J.W., Maseyk, F.J., Bekessy, S., Gordon, A.,
Watson, J.E., Lentini, P.E., Gibbons, P., others, 2016. Taming a wicked problem:
resolving controversies in biodiversity offsetting. BioScience. http://dx.doi.org/
10.1093/biosci/biw038.
Martin, G.J., 2015. Fonctions du droit et mesures compensatoires françaises. In:
Restaurer La Nature Pour Att!
enuer Les Impacts Du D!
eveloppement. Quae,
pp. 16e25.
McKenney, Bruce A., Kiesecker, Joseph M., 2010. Policy development for biodiversity
offsets: a review of offset frameworks. Environ. Manag. 45 (1), 165e176.
McKinney, Michael L., 2008. Effects of urbanization on species richness: a review of
plants and animals. Urban Ecosyst. 11 (2), 161e176. http://dx.doi.org/10.1007/
s11252-007-0045-4.
MEA [Millenium Environmental Assessment, 2005. Current State and Trends.
Global Assessment Reports, vol. 1. Island Press, Washington, USA.
MEDDE [Minist"
ere de l'Environnement, du D!
eveloppement Durable et de l'Energie],
2013. Lignes directrices nationales sur la s!
equence !
eviter, r!
eduire et compenser
les impacts sur les milieux naturels (R!
ef!
erenceS). Minist"
ere de l’!
ecologie, du
d!
eveloppement durable, des transports et du logement, Paris, France.
MEDDE [Minist"
ere de l'Environnement, du D!
eveloppement Durable et de l'Energie],
2012. Doctrine relative "
a la s!
equence !
eviter, r!
eduire et compenser les impacts
sur le milieu naturel. Minist"
ere de l’!
ecologie, du d!
eveloppement durable, des
transports et du logement, Paris, France.
Moreno-Mateos, David, Maris, Virginie, B!
echet, Arnaud, Curran, Michael, 2015. The
true loss caused by biodiversity offsets. Biol. Conserv. 192, 552e559.
Moreno-Mateos, David, Power, Mary E., Comín, Francisco A., Yockteng, Roxana,
2012. Structural and functional loss in restored wetland ecosystems. PLoS Biol.
10 (1), e1001247. http://dx.doi.org/10.1371/journal.pbio.1001247.
Morgan, Richard K., 2012. Environmental impact assessment: the state of the art.
Impact Assess. Proj. Apprais. 30 (1), 5e14.
Palmer, M.A., Filoso, S., 2009. Restoration of ecosystem services for environmental
markets. Science 325 (5940), 575e576.
Partidario, M.R., 2015. A strategic advocacy role in SEA for sustainability. J. Environ.
Assess. Policy Manag. 17, 1550015.
Qu!
etier, Fabien, Regnery, Baptiste, Levrel, Harold, 2014. No net loss of biodiversity or
paper offsets? A critical review of the French no net loss policy. Environ. Sci.
Policy 38, 120e131.
R development core team, 2016. R: a Language and Environment for Statistical
Computing. R Foundation for Statistical Computing, Vienna, p. 2014.
Regnery, B., Couvet, D., Kerbiriou, C., 2013a. Offsets and conservation of the species
of the EU habitats and birds directives. Conserv. Biol. 27, 1335e1343.
Regnery, B., Qu!
etier, F., Cozannet, N., Gaucherand, S., Laroche, A., Burylo, M.,
Couvet, D., Kerbiriou, C., 2013b. Mesures compensatoires pour la biodiversit!
e:
comment am!
eliorer les dossiers environnementaux et la gouvernance? Sci.
Eaux Territ. Hors-s!
erie num!
ero, vol. 12, pp. 1e8.
Samarakoon, Miriya, Rowan, John S., 2008. A critical review of environmental
impact statements in Sri Lanka with particular reference to ecological impact
assessment. Environ. Manag. 41 (3), 441e460.
Tallis, Heather, Kennedy, Christina M., Ruckelshaus, Mary, Goldstein, Joshua,
Kiesecker, Joseph M., 2015. Mitigation for one &all: an integrated framework
for mitigation of development impacts on biodiversity and ecosystem services.
Environ. Impact Assess. Rev. 55, 21e34.
Tardieu, L., Roussel, S., Thompson, J.D., Labarraque, D., Salles, J.-M., 2015. Combining
direct and indirect impacts to assess ecosystem service loss due to infrastruc-
ture construction. J. Environ. Manag. 152, 145e157.
Thompson, J.D., 2005. Plant Evolution in the Mediterranean. Oxford University Press
on Demand.
Thompson, S., Treweek, J.R., Thurling, D.J., 1997. The ecological component of
environmental impact assessment: a critical review of british environmental
statements. J. Environ. Plan. Manag. 40 (2), 157e172 .
Treweek, Jo, Thompson, Stewart, 1997. A review of ecological mitigation measures
in UK environmental statements with respect to sustainable development. Int. J.
Sustain. Dev. World Ecol. 4 (1), 40e50.
Whitehead, A.L., Kujala, H., Wintle, B.A., 2016. Dealing with cumulative biodiversity
impacts in strategic environmental assessment: a new frontier for conservation
planning. Conserv. Lett. 1e10. http://dx.doi.org/10.1111/conl.12260.
Wood, C., Djeddour, M., 1989. Environmental Assessment of Policies, Plans and
Programmes. Interim Report to the Commission of European Communities. EIA
Centre. University of Manchester.
C. Bigard et al. / Journal of Environmental Management 200 (2017) 35e45 45