Content uploaded by Julien Andrieu
Author content
All content in this area was uploaded by Julien Andrieu on Mar 23, 2023
Content may be subject to copyright.
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
1
What is the plant biodiversity in a cultural landscape? A comparative, multi-scale and
interdisciplinary study in olive groves and vineyards (Mediterranean France).
Marianne Cohen, Clélia Bilodeau, Frédéric Alexandre, Michel Godron, Julien Andrieu,
Etienne Grésillon, Florence Garlatti, Aurélien Morganti
Corresponding author:
M. Cohen
Université Paris Sorbonne, Sorbonne-Universités, ENeC (UMR 8185)
191 rue St Jacques,
75005 Paris, France
E-mail : Marianne.Cohen@paris-sorbonne.fr
Phone : 00 33 1 44 32 14 44
C. Bilodeau, M. Godron, E. Grésillon, F. Garlatti, A. Morganti
Sorbonne Paris Cité, Université Paris Diderot, LADYSS (UMR 7533)
Case 7001,
75205 Paris Cedex 13, France
F. Alexandre
Sorbonne Paris Cité, Université Paris-Nord, Sorbonne Paris Cité, CRESC (EA 2356)
99, avenue Jean-Baptiste Clément
93430 Villetaneuse, France
J. Andrieu
Université de Nice Sophia Antipolis, ESPACE (UMR 7300)
98, boulevard Édouard Herriot - BP 3209 –
06204 Nice Cedex 03, France
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
2
Abstract:
In a context of agricultural intensification and increasing urbanization, the biodiversity of
farmed landscapes is a key to improving the sustainability of agro-ecosystems. We seek to
ascertain the plant biodiversity of farmed and abandoned vineyards and olive-groves and to
identify the factors underlying it: natural and cultural; on local, landscape, and regional scales.
To do so, we recorded and calculated the floral biodiversity of 106 georeferenced plots and
121 plot edges distributed across 6 French Mediterranean terroirs, surveyed the practices and
perception of 55 farmers, and mapped the landscapes of 20 communes in a GIS.
Statistical tests proved that richness and spatial diversity on plots are favored by local low
intensity management integrating heritage and landscape objectives. The presence of edges
augments the richness and diversity around vineyards. The highest value of spatial diversity
was found using the terroir variable. Maximum richness is found in olive groves which are
maintained by amateur gardeners and located in the middle of the urbanization gradient. The
diversity of biological traits is listed according to: a) an herbaceous diversity gradient
explained by management; b) a specialization gradient explained by landscapes and distance
to large urban areas.
Our results draw perspectives to improve existing models of the links between agriculture,
biodiversity, and landscape, considering cultural and geographical factors. They lead to
recommendations regarding the management of landscapes based on local knowledge and
good practices.
Key words: agroecosystem; biodiversity; management; perception; heritage; terroir;
urbanization; gradient
Graphical abstract
Highlights
Olive and vine cultural landscapes are scattered along the urbanization gradient
Richness increases in peri-urban landscape with weak disturbance, heritage concern
Specialization pattern increases with distance to city, natural landscape and slope
Spatial diversity is higher between vineyards and edges, and between home terroir
Geographical and cultural factors improve the existing models of agro-biodiversity
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
3
INTRODUCTION
In a context of agricultural intensification and increasing urbanization, the biodiversity of
farmed plots is a key to improving the sustainability of farmed landscapes and their provision
of ecosystem services (Plieninger et al., 2014; Termorshuizen and Opdam, 2009; Van Zanten
et al., 2014). This issue finds an echo in the hypothesis of this article, that there is a link
between food products, the manner they are produced in specific areas known in France as
‘terroirs’, and the ecological quality and cultural resonance of rural landscapes (‘terroir’
corresponds to a territory distinguished by a crop associated with practices and landscapes,
INAO-INRA, 2006). This argument appears in European landscape research, programs of
professional bodies and public policies (Buergi, 2002; Burlingame and Dernini, 2010; INAO-
INRA, 2006; Luginbühl, 2012; UNESCO-SCBD, 2014). The biodiversity farmed landscapes
can foster is also protected by specific public policies (Allag-Dhuisme et al., 2010;
Cambecèdes et al., 2012). In the Mediterranean region, olive and wine-growing landscapes
offer two kinds of diversity, natural and cultural, they are thus expressions of this kind of
agriculture. They are considered to be cultural landscapes, or "combined works of nature and
man" with deep historical roots (ICOMOS, 2005; Loumou and Giourga, 2003; Luengo, 2011;
Naveh, 1997; UNESCO-SCBD, 2014).
In rural landscapes shaped by agriculture, links between agriculture and biodiversity can be
viewed from several different angles (Antrop, 2003; Poudevigne and Baudry, 2003). First,
agriculture is a form of ecosystem disturbance, the intensity, regime, and spatial extension of
which depend on farming practices and crop types (Van der Maarel, 1993). Second, the
influence of economic variables on biodiversity is not easy to demonstrate (Zeichmeister et
al., 2003). Third, numerous studies have focused on how agricultural landscapes are perceived
(examples: Boillat et al., 2004; Buijs et al., 2006; Tatoni, 1991); but few have demonstrated
the role of cultural factors in shaping their biological traits (Cohen, 2003; Friedberg, 1997;
Grésillon, 2009). As cultural factors interact with farmers or gardeners’ practices, this role is
indirect and consequently complex. In addition, certain types of agricultural systems
considered to have “high nature value”, such as extensive grasslands or olive groves located
in natural areas, are associated with high biodiversity (Cohen et al., 2010; Hoogeveen et al.,
2001; Poudevigne and Baudry, 2003). Given that they are established for many years, the
functioning of groves should correspond somewhat to those of ecosystems which have
experienced average to little disturbance (Bruggisser et al., 2010; Camarsa et al., 2010;
Loumou and Giourga, 2003); this similarity is reflected in their biological traits (Pujadas-
Salvá, 1986; Spanou et al., 2013). More generally, in their model linking agriculture,
landscape and biodiversity, Le Roux et al. (2008) consider arboriculture to be an intermediary
case positioned between intensive and extensive agro-ecosystems.
Studies of farmed landscape biodiversity most frequently are undertaken on a local scale in a
specific region and involve a single crop type; they generally focus on functional groups
which may improve crop production (auxiliary species) and are sensitive to intensification
(Altieri, 1999; Hoffmann and Greef, 2003; Coutinot, 2012). Intensification is linked to
agricultural practices, depending on their type, frequency and intensity (Nascimbene et al.,
2013; Zeichmeister et al., 2003). Among these practices, weeding and ploughing harm
biodiversity more than mowing, sowing or grazing (Bolo et al., 2009; Bruggisser et al., 2010;
Cohen et al., 2010; Firbank et al., 2008; Gago et al., 2007; Hoffmann and Greef, 2003;
Jauzein, 2001; McLaughlin and Mineau, 1995; Pain in Durand et al., 2013; Sanguankeo and
León, 2011). Fertilization and irrigation have a direct positive effect on herbaceous biomass
and plant competition (Grime, 1979). The influence of the agriculture type, organic or
conventional, is hard to isolate from other factors (Bruggisser et al., 2010). In addition,
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
4
natural landscapes and ecological corridors surrounding agricultural plots, as well as small
size of plots, are favorable to biodiversity (Belo et al., 2009; Chateil and Porcher, 2015;
Fahrig et al., 2015; Fried et al., 2008; Roux et al., 2008). By contrast, the abandonment of
olive groves and vineyards has a favorable short term effect on biodiversity, but it becomes
unfavorable in the medium term due to competition mechanisms (Bonet and Pausas, 2004;
Houssard et al., 1980, Potts et al., 2006). On the scale of a century, the post-farming
succession on abandoned terraced plots is influenced by fires, the destruction of walls, and
later land use patterns (Boillat et al., 2004, Tatoni, 1991).
The role of abiotic factors has been studied in agro-ecosystems which have experienced less
human-driven transformation (grasslands: Bennie et al., 2006; Cohen, 2003; Grime, 1979).
Nascimbene et al. (2013) demonstrated the favorable effect of steep slopes with manual
mowing; due to a lack of data on management, Allen et al. (2006) could not verify possible
interactions between this parameter and soil types.
Comparative studies covering several regions or crop types are less frequent. Spanou et al.
(2013) measured a similar level of diversity in both Greek vineyards and olive groves. Le
Roux et al. (2008, for France) and Firbank et al. (2008, for Great Britain) propose a general
model of agriculture, landscape and biodiversity relations, complementing the approach
developed by McLaughlin and Mineau (1995, for Canada) and Hoogeveen et al. (2001, for
Europe). Jauzein (2001) and Fried et al. (2008) demonstrated that agricultural intensification
leads to an increase in generalist plants replacing messicole species. According to Mayfield et
al. (2010), species loss related to intensification is not necessarily paired with a loss of
functional diversity. The environmental filter effect has been studied along urbanization
gradients, and is due to the reduction and fragmentation of habitats and selection pressure
linked to management practices, which favor generalist species at the expense of specialist
species (McKinney, 2006).
Our work focus on the environmental quality of olive-groves and vineyards by analyzing how
plant biodiversity contributes to the services provided by these cultural landscapes. We
compare six French Mediterranean terroirs presenting sufficiently diverse configurations to
enable the first lessons to be drawn on a regional scale. In addition, heavy urbanization
pressure on olive groves and vineyards presents an opportunity to verify the hypothesis
concerning the effect of the environmental filter on the biodiversity associated with farmed
landscapes (Angles, 2014). Our work has two original features: 1) we test factors at local,
landscape and regional scales; and 2) in addition to agronomic, environmental and landscape
factors, we also test cultural and socio-economic factors which generally are neglected in, or
are disconnected from, studies of biodiversity (Allen et al., 2006; Durand et al., 2013). In fine,
we assess the similarity of these groves and vineyards to ecosystems with average levels of
disturbance, and we propose a model of vineyard and olive grove plant biodiversity which
takes into account the actions and perceptions of society. These objectives are pursued
through a holistic and interdisciplinary approach (Antrop, 2003).
This approach consists of studying the quantitative, spatial, and functional dimensions of
plant biodiversity at nested scales: the farm plot and its internal heterogeneity, the plot and its
surrounding edges, the crop type and the home terroir. We further test the influence of factors
at different spatial scales. On the local level, we hypothesize that plant biodiversity is
influenced: a) directly by disturbances engendered by farming practices according to their
intensity; b) indirectly by cultural variables, c) weakly or indirectly by environmental
variables. On the plot (including the surrounding edges) and landscape scales, we hypothesize
that a high proportion of natural landscapes around groves and vineyards favors agricultural
plant biodiversity. On large geographical scales, we have two hypotheses: a) land use,
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
5
economic specialization and development projects of a terroir have a direct and/or indirect
influence on plant biodiversity; b) the environmental filter effect increases with proximity to
large urban areas.
By contributing to a better understanding of Mediterranean arboriculture biodiversity and its
determinants, our study should help to identify the farming and landscape practices which
best favor plant biodiversity, and, consequently, the multi-functionality and sustainability of
vine and olive cultural landscapes. Finally, the inclusion of cultural, socio-economic, and
geographic factors in our analysis renders it possible to offer realistic recommendations for
plant biodiversity management.
1. MATERIALS AND METHODS
2.1. Presentation of the case study
The study was carried out in 20 French communes situated on 6 terroirs in the Mediterranean
region (Figure 1; the commune is the lowest administrative unit in France). A terroir usually
is linked with agricultural products that bear protected designation of origin (PDO) labels, are
marketed through value chains, direct sales, or are (partially) home consumed. Specifically,
wine-growing terroirs are structured in particular around value chains. The economic viability
of the farms relies on the research of a high-quality niche product (ex. Olive oil in all terroirs,
Pic-St-Loup and Bandol wine) or a renewing vision of cooperatives (ex. new PDO label for
Terrasses-du-Larzac terroir. Olive-growing plays a much smaller role in Southern France
than vineyards, but it has been revived creatively over the past few decades. After a
devastating frost in 1956, olive groves have been rehabilitated by preserving or coppicing
local varieties of trees which are centuries old, or by planting new varieties of young trees,
and by restoring the heritage value of low walls. These renovations were carried out by
individuals working alone or were instigated by local non-profit projects. In addition, organic
or biodynamic agriculture is practiced on 22% of the farms studied, with little difference
between the crop types, or across the terroirs studied.
The landscapes of wine and olive-growing terroirs differ in their morphology and topography
(Table 1). The composition of the landscapes varies according to the proportion of space
dedicated to vines and olive trees, and the urbanization pressure, assessed by the percentage
of built-up plots and the distance to a large urban area. In general, natural landscapes cover
large areas on the hillsides, and are detached from the agricultural landscapes and buildings
located further down.
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
6
Table 1: Land-use in the communes studied and average characteristics of plots in the 6
terroirs studied
Scale of the
study
Land-use of IGN
plots
Terroirs
Wine-growing
Olive-growing
PDO Bandol
PDO Pic-St-Loup
PDO Terrasses du Larzac
Provence olive
oil
Nice
olive
oil
St-Guilhem-du-Désert
Pays d’Aubagne
Nice Hinterland
% of surface
in the
studied
communes
Olive-groves
2
2
2
<1
5
4
Vineyards
planted with
olive-trees
7
<1
<1
0
<1
0
Vineyards
9
21
19
0
<1
0
Other crops
8
0
<1
<1
5
8
Built surfaces
[built olive-
groves] and
infrastructures
20 [2]
10 [<1]
3 [<0.5]
<1 [0]
39 [4]
18 [2]
Wasteland,
hedges
3
11
9
2
10
5
Garrigue,
woodland
51
56
66
97
40
65
Feature* or
mean**
value of the
plots
Terraces*
…….Under-represented…...
...Over-represented……
Slope (°)
7.1
3
3.7
16.6
12.8
15.4
Altitude (m)
134
129
155
120
271
401
Size of plots (m²)
3951
10151
4008
1682
229
1459
Distance from
the large city
centroid (km)
16.3
20.9
29.2
30.9
12
15.1
Large city
Toulon
.......….Montpellier……….
Marseille
Nice
Sources: ORTHO IGN, IGN Plot map, IGN DTM, IGN Topographic map1/50000, Photo-interpretation: M. Cohen, F. Garlatti, C. Bilodeau,
J. Andrieu (see §2.2.2.4). * Chi squared test (p<0.0001), ** Kruskal-Wallis test (p<0.0001).
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
7
2.2. Sampling
The plant biodiversity of the vineyards and olive groves was studied in six terroirs, chosen to
be meaningful of the diversity of olive and wine growing systems by the Patermed scientific
team (Angles, Dir., 2014). Among the farmers listed in these six terroirs, we selected farms
located along an urbanization gradient in 20 communes (Figure 1) and depending from their
management intensity. The advantages of a sampling method along a landscape gradient were
noted by Poudevigne and Baudry (2003). Finally, fifty-five people among this list agreed to
receive the authors on their farms, plots were chosen to represent the management intensity
sampling.
Figure 1: Location map
They shared information about their farm types, farming practices and views of their activity,
the history of their groves or vineyards, and their perceptions of agricultural landscapes. The
farmers also pointed out abandoned grove and vineyard plots where the vegetation could be
inventoried despite the absence of their owners. GPS geo-referenced botanical surveys were
conducted on 106 cultivated and abandoned vineyard and olive grove plots; 69 plots were
surrounded by edges of contrasting vegetation described by 121 edge surveys; a total of 227
surveys were thus made (Table 2). Another set of 7 surveys, collected in vineyards with
another protocol, was used only to calculate the proportion of vineyards surrounded with
contrasting edges.
Table 2: Number and location of the studied plots and borders
Terroir
Olive groves
Abandoned
Vineyards
Edges
Total
Country of Aubagne
10
2
0
33
45
PDO Bandol
6
4
13
12
35
St-Guilhem-le-Désert
11
5
0
16
32
Hinterland of Nice
16
3
0
19
38
PDO Pic-St-Loup
3
2
13
29
47
PDO Terrasses-du-Larzac
9
3
6
12
30
Total
55
19
32
121
227
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
8
2.3. Material
2.3.1. Botanical survey protocol
The surveys were made in the spring between 2010 and 2013. On the plots, ten 1m² quadrats
associated with two needle points were set down at regular intervals along the longest
diagonal of the plot. The first quadrat was set down between point 0 and 1 meter, with a
wooden frame across a measuring tape and the needle points at the intersections, the second
quadrat at a distance equal to one tenth of the length of the diagonal of the plot and so on. The
area studied was identical for all of the plots regardless of their size, and the frequency of
species takes into account their frequency in the quadrats and their coverage rates to add
realism (Jauzein, 2001). On the contrasting edges of 69 plots, an inventory was made on an
area 2 meters wide and as long as the edge. This allowed us to approach plant biodiversity on
interlocking scales, from the patch (quadrat) and the plant community (plot), up to the scale of
the plot and its edges. The list of plant species was updated (www.tela-botanica.org) and their
biological traits were specified (Julve, 1998).
2.3.2. Plot management on local scale
Data on plot management were obtained through observation and farmer surveys. Four crop
types were identified: olive groves, vineyards, abandoned, burnt. The surveys and field
observations of the burnt plots were verified with old aerial photos (www.geoportail.fr).
Disturbances due to agricultural practices were classed into three levels – nil, low and high –
based on their magnitude according to the literature, regularity and spatial extension (Van der
Maarel, 1993). For example, ploughing or use of herbicides are considered as intense
practices, while grazing or mowing are low intensity ones. Inputs like irrigation and
fertilization were classed into three levels based on their frequency. Pruning and
phytosanitary treatment practices were not taken into account in this study because their
influence on the biodiversity of the herbaceous strata on plots is fairly indiscriminant.
Farmers’ views of their activity are cultural variables which interact with agricultural
practices and can influence floral diversity. Two types were considered here, a professional
view and a heritage view. A professional view of the activity is defined by a quest for
efficiency and conformance with professional standards, while a heritage view integrates
landscape objectives into farm management, along with a desire to respect the life and
diversity of cultivated plants and animals (for example, bees for their own or neighbors’
honey production). In this case, agricultural practices are implemented more cautiously:
undesirable species are periodically and manually uprooted or weeded in such a way as to
protect old stumps of local varieties. These practices aim to conserve or reinforce the esthetic
quality of the farmed plot according to farmers' cultural models (well-maintained walls, trees
artfully pruned, trellising with wood stakes, and flowery lawn; Angles, 2014). Heritage,
esthetic, and landscape objectives are most often associated with low disturbance and olive-
growing, and professional strategy with high disturbance and vine-growing (Chi-squared test,
p<0.0001).
2.3.3. Plot environment at different scales
On the local scale, the central GPS points of the 106 plots and their edges were integrated into
a GIS (Arcgis 10.2®). A DTM (IGN®) was used to calculate the slope, altitude and
orientation of each plot. The geological substratum type was extracted from BRGM’s
geological map (www.infoterre.brgm.fr).
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
9
In 20 communes, the land use of IGN® plots was determined through photo-interpretation of
ORTHO IGN® aerial photography (2010) and was classed into 4 types: artificial-urban,
farmed, semi-natural and natural. This classification is based on both the morphology and
function of landscapes. Building density varies in urban plots, but the primary function of an
artificial-urban classed IGN plot is human habitation. The results of this landscape mapping
indicate that our survey sample was coherent with land use (Tables 1 and 2).
The percentage of land use types were calculated along the length of plot edges, within a 500
meter radius around the plots, and on the commune where the plots were located. The slope
was taken into consideration to calculate the distance of 500 meters. The Shannon index
describes the diversity of land use on plot edges (Table 3, Appendix A).
By calculating the distance to the centroid of the closest large urban area, plots were situated
along an urbanization gradient (Table 1, source: IGN 1/50000).
2.4. Calculations of indexes at the different scales
Several indexes were calculated with Dyalog® software package, Xlstat® and R®
software, to grasp the quantitative, spatial and functional diversity (interactions with fauna
and with the soil, climate adaptation, succession stage) on the plots and their landscapes
(Lavorel and Garnier 2002) (Table 3, Appendix A). Floral richness is significantly correlated
with structural and Shannon’s diversity indexes (r²=0.860, p<0.0001 and r²=0.927, p<0.0001).
Table 3: Calculated indexes at the different scales
Scale
Richness
Biological traits*
Similarity
Land-use
diversity
Spatial
diversity
Intra-plot
Species
frequency index
ISf
(Godron et al., 1983)
Structural
diversity
IStd
(Godron, 2012)
Plot
Floral
richness Rf
(Godron et al.,
1983)
Biological trait
index IBt
(Kent 2012)
Steinhaus
Index ISs
(Kent, 2012)
Spatial
diversity
ISpd
(Godron, 2012)
Edge
Edge
diversity IEd
(Shannon Index, in
Kent 2012)
Crop type
Terroir
*Biological type, chorology, pollination mode, seed dispersal mode, succession stage of the optimal habitat
2.5. Data processing
Our multi-scale data sets were statistical processed with Xlstat® software and Dyalog®
software package to analyze and determine the relationships between the floral, agronomic,
landscape, cultural, and geographic data. Our step by step analysis relies on the use of
complementary treatments. A first set of treatments are based on rather simple methods
belonging to inferential statistics. Statistical tests were chosen according to the nature and
distribution of variables: relationships between qualitative variables (Chi squared test);
between qualitative and quantitative variables (Mann-Whitney and Kruskal-Wallis tests);
between quantitative variables (Pearson’s correlation coefficient). Principal component
analysis (PCA, Spearman type) was finally applied in a synthetic purpose to give an image
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
10
of the relations between the biological traits indexes with, successively, three sets of variables
at local, landscape and regional scales, introduced as supplementary variables. Previously, we
checked the links between quantitative (biological traits indices, explanatory variables, r>0.8)
and qualitative variables (Chi squared test) and create composite variables to avoid
redundancy. Another treatment was based on non-inferential methods to measure the links
between species groups, the crop type and the home terroir. We obtained convergent results
by both methods, and we report below the more significant ones.
3. RESULTS
3.1. Diversified, ecologically interesting flora
Among the 646 species inventoried, ten were observed in over one third of the surveys and
are presented here in decreasing order: Crepis sancta, Geranium molle, Bromus madritensis,
Asparagus acutifolius, Crepis vesicaria subsp. taraxacifolia, Hordeum murinum, Daucus
carota, Lolium rigidum, Sanguisorba minor and Rubia peregrina.
The chorology and the biological traits of these plants are similar to those of the ensemble of
flora found on the plots: 43.5% were Mediterranean and sub-Mediterranean species, less than
expected in this region; 40% were annual species, a form of adaptation to drought and a
response to agricultural practices that periodically destroy the herbaceous cover (Jauzein
2001). We found two exotic species out of the 34 listed by Jauzein (2001) in French
Mediterranean crops. While neither olive trees nor grape vines require insects to pollinate
their flowers, about half of the species depend on fauna for reproduction: 54% were insect-
pollinated species and 43% zoochore species. The plots studied contain species of different
dynamic stages: annual (40%) and perennial herbaceous (38%), shrubby (18%) and forest
(4%) vegetation.
Nearly one out of six species in olive and vineyard landscapes are sufficiently rare to be given
national or regional protection (115 out of 646 species; MNHN-FCBN-UICN 2012;
Cambecèdes et al., 2012). Nonetheless, nearly 3/4 of these 115 species are on the limits of
their geographic area or messicole species which are endangered outside the Mediterranean
region and are a legacy of protohistoric farming activities (Jauzein, 2001).
Richness, biological traits and protection of plant species give a comprehensive picture of the
biodiversity of these agricultural lands.
3.2 On the plot scale
3.2.1. Management intensity influences structural diversity
The spatial distribution of plant species within a plot, measured by the structural diversity
index IStd, is influenced by crop type, farmers’ views of the activity (Kruskal-Wallis test,
p<0.0001), disturbance level and secondarily by the exposure of the plot (Mann-Whitney test,
p<0.0001; r=-0.226, p=-0.023, respectively). The cases most conducive to the
homogenization of herbaceous cover are burnt plots, plots farmed professionally, intensively
farmed plots, and plots located on slopes with little exposure to the sun.
3.2.2. Management intensity and crop type influence floral richness
Floral richness Rf on plots ranges from 14 to 86 species. Floral richness and number of
protected species vary as a function of crop type (Mann-Whitney test, p=0.001). On average,
olive groves and abandoned plots present a considerably higher floral richness and a greater
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
11
number of protected species than vineyards and burnt plots (51 and 7 species, 29 and 4
species, respectively). Floral richness also vary as a function of intensity of practices and, in
a very significant manner, farmers' view of the activity (Kruskal-Wallis test, p=0.039 and
<0.0001, respectively). Little disturbed plots and those owned by individuals pursuing
heritage objectives present a higher mean number of species than highly disturbed plots and
plots farmed professionally (55 and 54 species, 32 and 31 species, respectively).
3.2.3. Crop type has little influence on assemblages of species
The frequency analysis indicates that crop type has little influence on floristic composition
(1.64 binons, binon is the standard unit on
information
), due to the somewhat random
assemblages of species in agricultural environments (many opportunistic and annual species).
Nonetheless, a group of widely distributed species is significantly associated with vineyards:
Cynodon dactylon, [Vitis vinifera], Lolium rigidum, Lactuca serriola, Lamium amplexicaule,
Diplotaxis erucoides, Convolvulus arvensis, and Sonchus oleraceus. These species are highly
resistant to intense disturbance engendered by repeated ploughing and chemical treatments,
thanks to their vegetative vigor or long-lasting flowering period.
3.2.4. Two biodiversity gradients depending on biological traits
The first two axes of the PCA of the biological trait indexes IBt of the 106 plots describe
59.95% of the variability and indicate two gradients (Figure 2). The herbaceous biodiversity
gradient, described by axis 1, is strongly correlated with structural and the Shannon’s
diversity indices (r=0.910, 0.832, p<0.0001), themselves correlated with floral richness (§
2.4). It orders surveys according to the value of indexes of certain biological traits and the
extent of plant cover.
The traits most highly correlated with this axis are an annual lifetime, correlated with mixed-
pollination, wind-borne seed dispersal and European distribution (0.8<r<0.9). The ensemble
of traits correlated with axis 1 indicates herbaceous species with diverse ecological functions.
Annuals, (0.8<r<0.9) perennials and bulbs (0.5<r<0.4) are associated with dynamic stages
ranging from annual sparse and crop commensal grasslands, herbaceous ecotones and
prairies, and including wastelands (0.5<r<0.8). All geographical distribution areas (0.5<r<0.9)
and pollination and seed dispersal modes (0.6<r<0.9) are represented, with the exception of
endozoochory. This gradient is mainly explained by management intensity. Low disturbed
olive-groves with heritage objective and important inputs are positively correlated (r=0.532
and 0.294 respectively), while strongly disturbed vineyards with professional objectives and
other strongly disturbed plots are negatively correlated (r=-0.396 and -0.281 respectively).
This is also the case for burnt plots (r=-0.222).
The biodiversity specialization gradient sets specialist species in the advanced stages of plant
succession against generalist species in disturbed environments (axis 2, Figure 2). The traits
of specialist species are the following: endozoochory, correlated with woody phanerophytes
and chamaephytes, advanced dynamic stages (woods, garrigues, grasslands, shrubs)
(0.7<r<0.8), Mediterranean origin, correlated with insect pollination (0.6<r<0.7), perennial
herbs, (0.5<r<0.6). By contrast, generalist species have the following traits: cosmopolitan
origin (-0.5>r>-0.6); less dependence on fauna: mixed insect-pollinated, short lived, early
dynamic stages: annual wastelands and crop commensal grasslands (-0.3>r>-0.5). Steep
slopes, incidental disturbance by fire and abandoned plots are associated with the specialist
pattern (r=0.570, 0.512 and 0.320, respectively). The average slope varies from 14.6° on olive
groves, abandoned and burnt plots, characteristic of the specialist pattern, to 6.3° on vineyards
highly disturbed with a professional purpose (r=-0.336) characteristic of the generalist pattern
(Kruskal-Wallis test, p=0.0002).
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
12
Figure 2: Axis 1-2 Plan of Biological traits PCA Analysis.
3.3. On plot and surrounding edges scale and on landscape scale
3.3.1. Less floral richness on the edges
The mean floral richness is significantly lower on edges than on plots (27 and 43 species
respectively, Mann-Whitney test, p<0.0001). The mean number of protected species is lower
on edges and vineyards than on olive groves and abandoned plots (4 and 7 species
respectively; Mann-Whitney test, p<0.0001). This may be related to the edaphic constraint
(low walls), disturbance (paths, use of herbicides) or to competition for light on tree-covered
edges. However, among the 646 species inventoried, 130 (20%) were only found on edges,
which thus contribute to increasing plant biodiversity on the plot and surrounding edges scale.
3.3.2. Plot-edge similarity depends on crop type
The floral similarity between plots and their edges ISs depends in part on crop type (Kruskal-
Wallis test, p=0.012). The highest values of similarity between edge and plot are observed on
burnt wasteland (mean=0.65), due to the homogenization and selection effect of fire already
demonstrated in section 3.1.1. By contrast, the similarity index is lowest in the vineyards
(mean=0.23). Their edges contain specific species that double on average the floral richness at
the plot and edges scale (k=2.08±1.21); the species number vary from 15 to 62 species to 35
to 105 species, respectively without and with edges. The biological traits of edge species are
significantly different from those of vineyard plots (Chi squared test, p<0.0001); perennial
herbaceous species with Euro-Asian and Mediterranean distribution ranges characteristic of
prairie and wasteland stages are over represented in vineyard edges. Nonetheless, the mean
number of protected species is identical on the edges and on vineyard plots (4), and 9 out 39
of vineyards studied (23%) are not surrounded by contrasting edges.
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
13
3.3.3. A regional, contagion or selection effect homogenizes the flora around the plots
Between the plots and the edges, the spatial diversity index ISpd is 0.442. In 31% of the
surveys among which the Hamming distance Dh was the lowest (<60), the similarity factors
are, in order of frequency: home terroir (77.1%), contagion effect due to proximity (55.7%),
selection effect due to the management mode (35.7%), crop type (28.6%) or fire (11.4%), and
exposure (8.5%).
3.3.4. The diversity of edges is linked to crop type and to the plant biodiversity pattern
of plots
The diversity of edge land use IEd is almost twice less on burnt plots (mean Shannon index
=0.421) than on plots that have not been burned (0.814, Kruskal-Wallis test, p=0.003). This
result reflects the homogenization effect of fire, already observed in sections 3.1.1 and 3.2.2.
Urbanized edges and edge diversity are negatively correlated with the specialist biodiversity
pattern described by Axis 2 of PCA (r=-0.339, -0.290 r respectively). We observe the reverse
for natural edges (r=0.266).
3.3.5. The landscape has a different influence on richness and biodiversity pattern
A farmed landscape within a radius of 500 meters around plots and on the edge of plots is
unfavorable to Shannon’s diversity index (r=-0.251, p=0.009, r=-0.216, p=0.026
respectively), itself correlated with floral richness, while the reverse is the case for edges with
urban features or terraces (r=0.266 with p=0.006 and r=0.219 with p=0.024 respectively). A
farmed landscape is also negatively correlated with a specialist biodiversity pattern while
natural and semi-natural landscapes are positively correlated (r=-0.380, r=0.291 and 0.249
respectively).
3.4. On terroir and regional scales
3.4.1. Floral richness and structural diversity partially depend on home terroir
Belonging to a terroir is linked significantly with structural diversity (Kruskal-Wallis test,
p=0,008); it is also linked with floral richness Rf when the cultivated vineyards and olive
groves in the terroirs of Bandol and Terrasses du Larzac are differentiated (Kruskal-Wallis
test, p<0.0001, Table 4). Bandol vineyards, maintained with frequent chemical and
mechanical treatments in accordance with the standards of the Bandol PDO wine sector, show
the lowest values, while olive-groves maintained for leisure purposes or managed organically
the highest.
Table 4: Average floral richness by terroir and by type of cultivated crop
Terroir
Vineyards
Olive-Groves
All
Pays d’Aubagne
41.9
PDO Bandol
26.5
67.5
St-Guilhem-le-Désert
50.9
Nice hinterland
45.3
PDO Pic-St-Loup
34.4
PDO Terrasses-du-Larzac
32.3
49.3
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
14
3.4.2. Floristic composition depends on home terroir
The spatial diversity ISpd linked to terroir is equal to 0.634, which means that floristic
composition differs according to the terroir. The spatial diversity linked to edges is weaker
(0.442) and the minimums were observed in section 3.2.3 when a group of locations – plots
and edges – belonged to the same terroir (77.1% of cases).
3.4.3. Biological traits depends on home terroir
The home terroir has greater influence on the floristic composition of plots and edges (2.48
binons) than crop type (1.64 binons, see section 3.1.3). Among the species very significantly
linked to terroir, the proportion of Mediterranean species varies in the range 3:1 and annual
species in the range of 2:1 (Figure 3). The highest proportion of Mediterranean species in the
Pays d’Aubagne is due to the influence of fires, lower latitude and sunny-oriented terraces
while the lowest are in the hinterlands of Nice and Pic-Saint-Loup terroir, due to higher
latitude (Figure 1). In the hinterlands of Nice, the modified Mediterranean climate, due to
higher latitude and altitude, and the frequency of irrigation, on half of the plots, favor
perennial herbaceous species (Figure 1 and Table 1). The proportion of cosmopolitan species
and species of early succession stages increases with the intensity of management (maximum:
PDO Bandol); thicket, garrigue and forest species characterize the natural landscape and
extensive heritage management mode of Saint-Guilhem-le-Désert.
Figure 3: Geographical origins (A) and succession stages of the optimal habitat (B) of species
highly related with the terroirs
3.4.4. Place of the terroirs in the plant biodiversity gradients
Terroirs are significantly linked with agricultural practices studied on local level (Chi squared
test, p<0.0001). For example, in Bandol, highly disturbed vineyards for professional
objectives are over-represented while in Saint-Guilhem-Le-Désert this is the case for lowly
disturbed olive-groves with heritage objectives. Their floristic traits are also partly explained
by the rate of urbanization (28.6 and <3%, respectively).
3.4.5. Edges are more or less diversified and contrasting depending on the terroir
Edges are more diversified on the Bandol and Pic-Saint-Loup wine-growing terroirs than on
the rest (mean IEd=1.01 and 0.69 respectively; Kruskal-Wallis test, p=0.0002). The increase of
floral richness due to contrasting edges around vineyards is higher in Pic-St-Loup than in
Bandol and Terrasses du Larzac (k=2.6 and 1.7, respectively, Mann Whitney test, p=0.023); it
is the same for the proportion of plots with contrasting edges and the complexity of their
vegetation (100% and 65% of the plots, forested and herbaceous edges, respectively).
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
15
3.4.6. Landscape of the commune and distance to city influence plant biodiversity
pattern
The rate of farmed landscape in the home commune of the plot is negatively and weakly
correlated with structural and Shannon’s diversity indexes, themselves correlated with floral
richness (r=-0.211, p= 0.0.030, and r=-0.227, p=0.022). The specialization gradient is
positively correlated with the rate of natural landscape in the home commune, itself
negatively correlated with the rate of built surfaces, and with the distance to the closest large
city centroid (r=0.450 and 0.265 respectively). The rate of farmed landscape in the commune
is highly correlated (r=0.819) with the rate of farmed landscape in a radius of 500 meters
around the plots, for this reason it was not included in this analysis; this correlation suggests a
very homogeneous landscape, unfavorable to specialist species. These results illustrate the
role of environmental filter in peri-urban and in farmed landscapes.
3.5. A synthetic image of vineyard and olive-grove plant biodiversity
We synthesize the ensemble of results to give in an image of the relations between the two
gradients of biodiversity and factors on local and regional scales corresponding to different
services rendered by these multi-functional landscapes: ecological, economic and cultural
(Table 5). The herbaceous diversity gradient is mainly linked with local factors and the
specialization gradient with regional factors; landscape and edges composition are variables
of minor importance. The map allows one to visualize the diversity of situations between and
within the terroirs, and as a function of their location within a gradient going from natural to
peri-urban landscapes (Figure 4).
Table 5: Gradients of plant biodiversity and multi-scale variables (correlations >0.4 are in
bold characters)
Multi-scale variables in
supplementary position in
3 successive PCA
Herbaceous richness and
diversity (PCA-Axis1)
Specialization of biological
traits (PCA-Axis2)
Low
High
High
Low
Local
Management
Highly
disturbed
vineyards, for
professional
Lowly
disturbed
olive-groves,
including
heritage
objectives
Nil or burnt
Highly
disturbed
vineyards, for
professional
Environment
al conditions
Steep slope
Flat land
Landscape
Landscape
composition
and edges
Natural and
semi-natural
landscape,
natural edges
Farmed
landscape,
urbanized and
diversified
edges
Regional
Landscape in
the home
commune
Natural,
lowly built
Distance to
cities
Far
Close
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
16
Figure 4: Patterns of plant biodiversity of 106 studied plots in 6 terroirs
4. DISCUSSION
We shall successively address two points in this discussion: the validation of our hypotheses
concerning the role of factors explaining plant biodiversity of agro-ecosystems, and the
comparison of our results with the agro-landscape models. For this purpose, we project our
results in the models proposed by Hoogeven et al. (2001), Le Roux et al. (2008) and Firbank
et al. (2008).
4.1. How may vineyard and olive grove plant biodiversity be explained?
This study chose to explore several hypotheses based on the conviction that the plant
biodiversity of cultural landscapes like Mediterranean olive groves and vineyards depends on
a complex gamut of factors, even more so today as they are undergoing fast-paced change due
to their proximity to very dynamic, large urban centers. Four type of variables were tested.
First, we examined the role of the intensity of agricultural practices on floral richness (Fried et
al., 2008). The comparison of our results with the model proposed by Hoogeven et al. (2001)
enables the crop types to be positioned along a gradient of intensity, in keeping with our
hypothesis and with the literature (olive groves: Camarsa et al., 2010; Loumou and Giourga,
2003; vineyards: Bruggisser et al., 2010; abandoned and burnt plots: Bonet and Pausas, 2004;
Houssard et al., 1980; Potts et al., 2006, Tatoni, 1990; Figure 5). In contrast with the results
obtained in Greece by Spanou et al. (2013), the floral richness of studied olive groves is well
above that of vineyards (k=1.76 compared to k=0.81). This is coherent with the higher
intensity of agricultural practices on vineyards, implemented to control the competition effect
of the herb layer on vine stocks. In addition, mean crop area is larger in vine terroir (Table 1),
and the negative influence of this factor on biodiversity has been shown by Fahrig et al.
(2015).
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
17
Figure 5: Species richness and agriculture intensity
Second, we examined the role of cultural and socio-economic factors. On the plot scale,
agricultural practice is implemented with more or less care according to farmers’ views of the
activity, meaning the manner by which farmers perceive their activity and relationship with
the environment, orienting themselves towards professional objectives or integrating a
concern for heritage and landscape. This cultural factor is well-related with species richness
and herbaceous diversity in our sample. On the terroir scale, farmer associations play a role,
guided by a vision of the territory and a dominant economic model, and implementing
practices oriented by organized and informal professional structures. However, the link
between terroir and plant biodiversity is more complex than expected. It is fairly weak with
richness and structural diversity, due to the diversity of crop types in some terroirs; it is
significant with spatial diversity and the plant biodiversity pattern, due to interactions
between climate, practices and landscapes differences, which is in line with the definition of
terroir given by INAO-INRA (2006).
Third, environmental factors have, as expected, little or no influence on richness and
structural diversity, except slope which is linked with the specialization gradient and
orientation, negatively correlated with structural diversity.
The feature shared by farmers’ views, terroir and slope is that they are all indirect factors.
Cultural and socio-economic factors interact with practices (Friedberg, 1997). Slope would
favor the abandonment of plots (it is hard to work on them) and the spread of fire (ICONA,
1990). In contrast with the studies of Allen et al. (2006) and Nascimbene et al. (2013), slope
is not an environmental factor with a direct influence on plant biodiversity. Similarly, the
positive influence of the small size of plots and the age of groves on biodiversity, due to
ecological processes according to Fahrig et al. (2015) and Spanou et al. (2013), is probably
linked to the more careful practices on old and small groves in our study case, according to
farmers’ survey.
Finally, landscape and urbanization have a complex influence on plant biodiversity, due to
their role on environmental filtering (McKinney, 2006), and on the mix of species pool
(Mayfield et al., 2010), illustrating the individualization of two gradients of plant biodiversity,
discussed below. This influence is more noticeable at regional scale than on landscape scale
(edges and radius of 500 meters around the plots).
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
18
4.2. Vineyards and olive groves: agro-landscape models?
Firbank et al. (2008) proposed a conceptual model to characterize the three dimensions of
agricultural intensification and their effects on biodiversity: land use, landscape structure, and
agricultural practices, corresponding to three interlocking scales – region, landscape and plot.
Nonetheless, they consider that such a model is not realistic if the richness of just one
taxonomic group is considered as an indicator of biodiversity. On the regional scale, the most
appropriate indicators would be those which consider changes in the biological traits of
species rather than the richness of a taxonomic group. While we used only one taxonomic
group, namely plants, our results confirm this theory because floral richness is determined by
local plot management, while biological traits are related to land use and slight differences in
climate observed at the regional scale. As Mayfield et al. (2010) observed, species and
functional diversity losses related to land use intensification are not always associated. The
individualization of the two plant biodiversity gradients, the first for diversity and herbaceous
cover, itself correlated with species richness, the second for trait specialization, would
confirm this decoupling process, one related to local scale factors on the one hand and
regional factors on the other.
Considering the mechanisms underlying the links between agriculture, landscape and plant
biodiversity, competition likely acts on the local scale, through the direct disturbance effect of
farmers’ herbaceous cover control practices, and spontaneously on abandoned plots. By
contrast, environmental filter and species pool change mechanisms likely act on the regional
scale due to urbanization. However, on our sample of cultural landscapes, the role of
urbanization in biotic homogenization remains limited to increasing the proportion of
cosmopolitan species, contrary to what was observed in urban areas or other agricultural
environments, where invasive species modify the species pool (Jauzein, 2001; McKinney,
2006; Williams et al., 2008). Moreover, a spatial homogenization is observed in 31% of plots
and edges due to home terroir, contagion and selection effects noticed by different authors
(Fried et al., 2008; Gounot, 1969; INA0-INRA, 2006; Jauzein, 2001). Subtle climate
differences between the terroirs further accentuate this spatial effect on regional scale.
However, the role of climate cannot be isolated from that of landscapes shaped by agricultural
practices which are over-represented on these terroirs (INAO-INRA, 2006).
Considering each type of crop, our results show that the studied vineyards are associated with
low species richness due to intensive ploughing and chemical inputs, as found by Fried et al.,
2008 and numerous other authors. The most significant species on vineyards are able to
withstand the high level of disturbance caused by agricultural practices thanks to their
biological traits as observed by Fried et al. (2008, 2012). The standards imposed by
cooperatives and professional bodies (“keep vineyards tidy”) contribute to shaping and
homogenizing the individual practices of wine growers, and to justifying their intensity and
repetition (Angles, 2014). On the terroirs where wine growers are more autonomous, practices
are less intense and more variable (chemical weeding is not systematic), the flora is more
diverse and less generalist. The presence of hedges, garrigue and forest around vineyards
allows floral richness and traits diversity to be increased on the plot and edges scale (Figure
6); this edge effect is higher in Pic-St-Loup terroir where edges are covered with unmanaged
forest. The low trait diversity is linked to an agricultural landscape located near urban centers,
with well-developed infrastructure and buildings around the plots, illustrating the role of the
environment filter (McKinney, 2006). Finally the influence of natural and semi-natural
elements in the landscape on richness and specialization gradient is less favorable than in the
model proposed by Le Roux et al. (2008) (Figure 6). Our sample of vineyards imperfectly
represents the agro-landscape gradient modeled by these authors.
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
19
Figure 6: Biodiversity, landscape and crop type
Abandoned and burnt plots that we studied are dominated by specialist species due to their
removed position on the urbanization gradient, their insertion in a landscape dominated by
natural surface areas, and the cessation of agricultural activities, linked to steep slopes. Plots
which have been burned have a floral richness and a structural and spatial diversity much
lower than other abandoned plots due to selection and homogenization fire processes. These
results are in line with previous studies of plant biodiversity in abandoned groves (Bonet and
Pausas, 2004; Houssard et al., 1980; Potts et al., 2006, Tatoni, 1990), but they do not
correspond well to models which do not integrate the specific features of the Mediterranean
hinterlands (Figures 5 and 6).
With regard to olive groves, those we studied are rich in herbaceous and protected species
whose biological traits are diversified. The co-existence of two types of diversity, specific and
functional, is linked to the low intensity of agricultural practices, which are in turn linked to
cultural and socio-economic factors. However, we cannot exclude the role of co-evolution
between olive trees and herbaceous communities due to the age of the groves, which
influence as a direct or indirect factor should be further investigated (Spanou et al., 2013). It
is the same for plot size (Fahrig et al., 2015).
When olive groves are located in landscapes with a low rate of natural surfaces (25 to 50 %),
one third of plots are built up (on average) and the floral richness is higher (mean 53 species,
Figure 6). In this peri-urban situation, mean floral richness practically doubles, from 36 to 69
species, between olive groves managed intensively and those maintained carefully by
gardeners who are interested in the cultural and esthetic value of olive groves (Angles, 2014;
Loumou and Giourga, 2003). Here, the “esthetic experience” is not a conventional one of
scenic beauty, but rather one of “perceived care, attachment, and identity” (Gobster et al.,
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
20
2007), linked with gardeners’ cultural models. The high floral richness observed at this level
of the urbanization gradient thus is not only explained by the mix of the species pool
(Mayfield et al., 2010), but also by socio-cultural factors. By contrast, in landscapes
dominated by natural surfaces far removed from large urban areas, the mean species richness
of olive groves diminishes (50 species), but the functional diversity increases with the
presence of specialist species (Figure 6). These specialist species are preserved by the
mitigation of the environmental filter effect due to the distance to large cities (McKinney,
2006) and by the very careful, heritage management of plots rehabilitated through the efforts
of individuals or groups (ex: Saint-Guilhem-le-Désert).
These two particular features, the floral richness of peri-urban olive groves and the specialist
pattern of olive groves plant biodiversity in natural landscapes, explain the weak match
between our results and the agro-landscape model proposed by Le Roux et al. (2008), in
contrast to what was observed in Andalusia (Cohen et al., 2010). Our sample of olive groves
thus is a counterexample of the loss of functional redundancy linked to the intensification of
agriculture observed by Mayfield et al. (2010), because they generally are managed by low
intensity practices implemented with care. Our hypothesis of proximity between the
functioning of plots and the functioning of natural ecosystems with average disturbance is
thus validated in the case of the olive groves studied in France, and is consistent with the
results obtained by Pujadas-Salvá in past times in Spain (1986) and by Spanou et al. (2013)
recently in Greece. This proximity between olive groves and natural ecosystems is indirectly
linked to the perception of olive groves as a cultural landscape and to the symbolic and
esthetic role of the olive tree (Angles, 2014).
5. CONCLUSION
In conclusion, our holistic and multi-scale approach enables considerable progress in our
understanding of the complex relations between cultural landscape and plant biodiversity
(Antrop, 2003; Naveh, 1997), complementing the existing models (Firbank et al., 2008;
Hoogeven et al., 2001; Le Roux et al., 2008). Our main finding is that the multi-functionality
of olive groves and vineyards and the potential ecological service they render to territories are
realized in different ways along the urbanization gradient, when they apply the set of cultural
landscape attributes: "combined works of nature and man" with deep historical roots, cultural
resonance and careful management of plots and edges. Plant biodiversity is not incompatible
with the economic feasibility of farms. We observed medium to high plant diversity on plots
and/or on edges in farms where economic valuation relies not only on productivity, but also
on heritage and landscape values, promoting niche products.
Our results renders it possible to offer realistic recommendations for landscape and plots
management. These are based on local best practices and knowledge, and consider the
economic constraints of agricultural systems and landscape dynamics observed on a regional
scale (Angles, 2014; Friedberg, 2007; Roué, 2014). One recommendation is to disseminate
the knowledge regarding ecological management of plots and edges and rehabilitation held by
peri-urban gardeners and by farmers in rural areas. With the exception of the art of pruning,
which is passed down locally, this intangible heritage is little valued (Angles, 2014). This
would lead to a greater proportion of plots contributing to the presence and circulation of
biodiversity in the agricultural matrix along the entire urbanization gradient, including
auxiliary species (vine: Hoffmann and Greef, 2003; olive-tree: Coutinot, 2012; alternate plant
host for auxiliary species are frequent in our sample, Warlop, 2006). Another is to rehabilitate
abandoned olive groves, which will not lessen their ecological value and will restore their
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
21
economic and cultural functions; this rehabilitation will furthermore contribute to preventing
fires, whose negative effects on plant biodiversity we have demonstrated. Lastly, the
ecological and landscape transition of wine-making territories should be promoted by
encouraging mechanical practices over chemical weeding and the sowing of non-
cosmopolitan and non-competitive herbaceous species. A more extensive management of plot
edges also should be encouraged so that these vegetal structures, which represent a landscape
heritage with high esthetic and cultural value, also can fulfill an ecological corridor function
within the agricultural space.
Appendix A: Indexes formula
Index
Eq.
Definition And Formula
Floral
richness
Rf
A.1
Number of species in a plot or an edge
Rf=
Species
frequency Isf
A.2
ISf= [0 ; 3]
Absence (i=0), presence (i=1), contact between a species and one
(i=2) or two needle-points (i=3), computed for each quadrat.
Biological
traits IBt (%)
A.3
IBt
n=number of species with a modality of biological trait, E=species,
Q=quadrat, Ifs= Species frequency index
Steinhaus
similarity
ISs
A.4
Comparison between floristic composition in the plot and in the
edge ISs = 2a/(2a+b+c)=2a/(n1+n2)
a=number of species present in the plot and in the edge, b= number
of species present only in the plot, c= number of species present
only in the edge, n1=total number of species in the plot, n2= total
number of species in the edge.
Edge land
use diversity
(Shannon)
IEd
A.5
IEd
pi ln pi
pi= proportion of edge length belonging to a land use type.
Structural
diversity IStd
A.6
Description of the spatial pattern of species within the plot.
IStd = (log2C10Si - log2CASi-2 /E
Si=sum of species present in the quadrats, E=number of species in
the plot, A=number of quadrats where the species is present and
Cnp=number of combinations between n objects p to p.
Spatial
diversity
ISpd
A.7
Spatial diversity within a group of locations is the mean value ISpd
of Hamming distance Dh between all the locations compared by
pair.
Dh = (1- c)/(a+b-c)
c=number of species present in the two locations, a=total number of
species in the first location, b= total number of species in the second
location
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
22
Acknowledgements:
This study was carried out thanks to the financial support for field work given by ANR
(French National Agency for Research, Systerra Program) to the Patermed project (Paysages
et Terroirs Méditerranéens). We thank in particular its scientific head, Stéphane Angles,
Senior Lecturer at Université Paris-Diderot-UMR 7533 Ladyss, the geography and biology
masters students at Université Paris-Diderot who participated in data collection (Rémi Cabel-
Simon, Romain Courault, Cédric Diamantino, Benjamin Fouillen-Marcel, Laurence Jézequel,
Maria Laraki, Christopher Mongeville, Marie-Isabelle Mutin, Richard Philippe, Myriam
Volut, Stefano Zanini), the Pole Image for its technical support, Grace Delobel for translating
the manuscript into English, the UMR Ladyss for the financial support for translation, and the
farmers and wine and olive sector professionals, without whose help this research could not
have been realized. We are grateful to the reviewers for their suggestions and their help to
improve our manuscript.
REFERENCES
Allen, H.D., Randall, R.E., Amable, G.S., Devereux, B.J., 2006. The Impact of Changing
Olive Cultivation Practices on the Ground Flora of Olive Groves in the Messara and Psiloritis
Regions, Crete, Greece. Land Degrad. and Development 17, 269-273.
Altieri, M.A., 1999. The ecological role of biodiversity in agroecosystems, Agriculture,
Ecosystems and Environment 74, Issues 1–3, 19–31.
Angles, S. (Ed.), 2014. Atlas des paysages de la vigne et de l’olivier en France
méditerranéennes, Quæ, Versailles.
Antrop, M., 2003. Expectations of Sciences towards Interdisciplinary/Transdisciplinary
Research. In: Tress, B., Tress, G., Fry, G. (Eds.) Interdisciplinary and Transdisciplinary
Landscape Studies: Potential and Limitations. Delta Series 2, Wageningen, pp. 44-54.
Belo, A.F., Simões, M.P., Pinto-Cruz, C., Castro, M.C., 2009. Efeitos da gestão do solo na
diversidade vegetal do olival. In de Sousa, E. et al. (Eds.) Herbologia e Biodiversidade numa
agricultura sustentável (XII Congresso da SEMh/XIX Congresso da ALAM/II Congresso da
IBCM, Lisboa, 10-13 de novembro de 2009), vol1, ISA Press, Lisbon, pp. 61-64.
Bennie, J., Hill, M., Baxter, R., Huntley, B., 2006. Influence of slope and aspect on long-term
vegetation change in British chalk grasslands. J. Ecology 94, 355-368.
Boillat, S., Burga, C.A., Gigon, A., Backhaus, N., 2004. La succession végétale sur les
cultures en terrasses de la Vallée de la Roya (Alpes-Maritimes, France) et sa perception par la
population locale. Geographica Helvetica 59(2), 154-167.
Bonet, A., Pausas, J., 2004. Species richness and cover along a 60-year chronosequence in
old-fields of southeastern Spain. Plant Ecology 174(2), 257-270.
Bruggisser, O.T., Schmidt-Entling, M.H., Bacher, S., 2010. Effects of vineyard management
on biodiversity at three trophic levels. Biological Conservation 143(6), 1521–1528.
Buergi, E. (Ed.), 2002. La convention européenne du paysage. Naturopa, vol 98, Conseil de
l’Europe, Strasbourg.
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
23
Buijs, A.E., Pedroli, B., Luginbühl, Y., 2006. From hiking farmland to farming in a leisure
landscape: changing social perceptions of the European landscape. Landscape Ecology 21,
375-389.
Burlingame, B., Dernini, S. (Eds.) 2010. Sustainable Diets and Biodiversity. Directions and
solutions for policy, research and action (Proceedings of the International Scientific
Symposium: Biodiversity and Sustainable Diets united against Hunger), FAO, Roma.
Camarsa, G., Gardner, S., Jones, W., Eldridge, J., Hudson, T., Thorpe, E., O’Hara, E., 2010.
Life among the olives: Good practice in improving environmental performance in the olive oil
sector. Official Publications of the European Union, Luxembourg.
Cambecèdes, J., Largier, G., Lombard, A., 2012. Plan national d’actions en faveur des plantes
messicoles. Conservatoire botanique national des Pyrénées et de Midi-Pyrénées / Fédération
des Conservatoires botaniques nationaux / Ministère de l’Écologie, du Développement
durable et de l’Énergie, Paris.
Chateil, C., Porcher, E., 2015. Landscape features are a better correlate of wild plant
pollination than agricultural practices in an intensive cropping system, Agriculture,
Ecosystems and Environment 201, 51–57.
Cohen, M. (Ed.), 2003. La Brousse et le Berger. Une approche interdisciplinaire de
l’embroussaillement des parcours. CNRS Editions (coll Espace et Milieux), Paris.
Cohen, M., Sol, S. et al., 2010. Paysage oléicole et diversité biologique. Résultats
préliminaires dans la Sierra Mágina. In: Araque E (Ed.), Seminario hispano-francés. El olivar:
Paisaje, Patrimonio y Desarrollo Sostenible, Universidad de Jaén/Asociación para el
Desarrollo rural de la Sierra Magina, Jaén, pp. 23-55.
Coutinot, D., 2012. Un bio-agresseur de l’olivier : la mouche de l’olive. In: Bervillé A, Breton
C (eds), Histoire de l'olivier. Quæ, Versailles, pp. 155-161.
Durand, L., Cipière, M., Carpentier, A.S., Baudry, J. (Eds), 2013. Concilier agricultures et
gestion de la biodiversité. Quæ (coll. Matière à débattre et décider), Versailles.
Fahrig, L. , Girard, J., Duro, D., Pasher, J., Smith, A., Javorek, S., King; D., Freemark Lindsay,
K., Mitchell, S., Tischendorf, L. , 2015. Farmlands with smaller crop fields have higher within-
field biodiversity, Agriculture, Ecosystems & Environment 200, 219–234.
Firbank, L.G., Petit, S., Smart, S., Blain, A., Fuller, R.J., 2008. Assessing the impacts of
agricultural intensification on biodiversity: a British perspective. Philosophical Transactions
of the Royal Soc. B: Biological Sciences 363(1492), 777-787.
Fried, G., Chauvel, B., Reboud, X., 2008. Évolution de la flore adventice des champs cultivés
au cours des dernières décennies : vers la sélection de groupes d’espèces répondant aux
systèmes de cultures. Innovations Agronomiques 2008(3), 15-26.
Fried, G., Kazakou, E., Gaba, S., 2012. Trajectories of weed communities explained by traits
associated with species’ response to management practices. Agriculture, Ecosystems and
Environment 158, 147-155.
Friedberg, C., 1997. Diversité, ordre et unité du vivant dans les savoirs populaires. Nature,
Sciences, Sociétés, 5(1), 5-17.
Gago, P., Cabaleiro, C., Garcia, J., 2007. Preliminary study of the effect of soil management
systems on the adventitious flora of a vineyard in northwestern Spain. Crop Protection 26(4),
584–591.
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
24
Gobster, P.H., Nassauer, J.I., Daniel, T.C., Fry, G., 2007. The shared landscape: what does
aesthetics have to do with ecology? Landscape Ecology 22, 959-972.
Gounot, M., 1969. Méthodes d’étude quantitative de la végétation. Masson, Paris.
Godron, M., 2012. Écologie et évolution du monde vivant, Harmattan, Paris.
Grésillon, É., 2009. Une géographie de l’au-delà ? Les jardins de religieux catholiques, des
interfaces entre profane et sacré, PHD Thesis, Université Paris IV-Sorbonne, Paris.
Grime, J.P., 1979. Plant strategies and vegetation processes. John Wiley and Sons, Chichester.
Hoffmann, J., Greef, J.M., 2003. Mosaic indicators. Theoretical approach for the development
of indicators for species diversity in agricultural landscapes. Agriculture, Ecosystems and
Environment 98(1-3), 387-394.
Hoogeveen, Y.R., Petersen, J.E., Gabrielsen, P., 2001. Agriculture and biodiversity in Europe.
Background report to the High-Level European Conference on Agriculture and Biodiversity
(5–7 June 2001), STRA-CO/AGRI/Council of Europe/UNEP, Paris.
Houssard, C., Escarre, J., Romane, F., 1980. Development of species diversity in some
Mediterranean plant communities. Vegetatio 43, 59-72.
ICOMOS – International Council on Monuments and Sites, 2005. Monuments and sites in
their setting: conserving cultural heritage in changing townscapes and landscapes. ICOMOS
15th General Assembly and Scientific Symposium, ICOMOS, Charenton-le-Pont (France).
ICONA, 1990. Clave fotográfica para la identificación de modelos de combustible. Defensa
contra incendios forestales, MAPA, Madrid.
INAO-INRA, 2006. Appellations d’Origine Contrôlée et Paysages. INAO, Paris.
Jauzein, P., 2001. Biodiversité des champs cultivés: l’enrichissement floristique. Dossiers de
l’Environnement de l’INRA, 21, 43-64.
Julve, Ph., 1998 ff. - Baseflor. Index botanique, écologique et chorologique de la flore de
France. Consulted on 5/5/2013. http://perso.wanadoo.fr/philippe.julve/catminat.htm
Kent, M. (2012). Vegetation description and data analysis: A practical approach (2nd
ed.), Wiley-Blackwell, Chichester.
Lavorel, S., Garnier, E., 2002. Predicting changes in community composition and ecosystem
functioning from traits: revisiting the Holy Grail. Functional Ecology 16, 545-556.
Le Roux, X., Barbault, R., Baudry, J., Burel, F., Doussan, I., Garnier, E., Herzog, F., Lavorel,
S., Lifran, R., Roger-Estrade, J., Sarthou, J.P., Trommetter, M. (Eds.), 2008. Agriculture et
Biodiversité. Valoriser les synergies. Expertise scientifique collective. Synthèse du rapport,
INRA, Paris.
Loumou, A., Giourga, C., 2003. Olive groves: The life and identity of the Mediterranean.
Agriculture and Human Values, 20, 87-95.
Luengo, M., 2011. Looking ahead: the olive grove cultural landscapes. In: Heritage, a driver
of development. Proceedings of the 17th ICOMOS General Assembly Symposium, ICOMOS,
Charenton-le-Pont (France), pp. 616-623.
Luginbühl, Y., 2012. La mise en scène du monde. Construction du paysage européen. CNRS
éditions, Paris.
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
25
Mayfield, M.M., Bonser, S.P., Morgan, J.W., Aubin, I., McNamara; S., Vesk, P.A., 2010.
What does species richness tell us about functional trait diversity? Predictions and evidence
for response of species and functional trait diversity to land-use change. Glob. Ecology and
Biogeography 19(4), 423-431.
McKinney, M.L., 2006. Urbanization as a major cause of biotic homogenization. Biological
Conservation 127, 247–260.
McLaughlin, A., Mineau, P., 1995. The impact of agricultural practices on biodiversity.
Agriculture, Ecosystems and Environment 55(3), 201–212.
MNHN-FCBN-UICN, 2012. La liste rouge des espèces menacées en France. Flore vasculaire
de France métropolitaine : Premiers résultats pour 1000 espèces, sous-espèces et variétés.
[National list:
http://www.uicn.fr/IMG/pdf/Tableau_Liste_rouge_flore_vasculaire_de_metropole.pdf]
[Regional lists: www.inpn.mnhn.fr/collTerr/region/24/tab/especesmenacees]
Naveh, Z., 1997. Landscape ecology and sustainability. Landscape Ecology 22(10), 1437–
1440.
Nascimbene, J., Marini, L., Ivan, D., Zottini, M., 2013. Management Intensity and
Topography Determined Plant Diversity in Vineyards, PLoS ONE 8(10), e76167.
Plieninger, T., van der Horst, D., Schleyer, C., Bieling, C., 2014. Sustaining ecosystem
services in cultural landscapes. Ecology and Soc. 19(2), 59.
Potts, S., Petanidou, T., Roberts, S., O’Toole, C., Hulbert, A., Willmer, P.G., 2006. Plant-
pollinator biodiversity and pollination services in a complex Mediterranean landscape.
Biological Conservation 129(4), 519-529.
Poudevigne, I., Baudry, J., 2003. The implication of past and present landscapes patterns for
biodiversity research: introduction and overview. Landscape Ecology 18(3), 223-225.
Pujadas-Salvá, A., 1986. Flora arvense y ruderal de la provincial de Córdoba. Tesis doctoral,
Escuela Técnica Superior de Ingenieros Agrónomos de Córdoba.
Roué, M., 2014. La théorie anthropologique au secours de la complexité. Comment penser et
étudier les relations homme-nature. In: Chenorkian R and Robert S (eds), Les interactions
homme-milieux. Questions et pratiques de la recherche en environnement, Quæ Editions (coll
Indisciplines), Versailles, pp. 131-146.
Sanguankeo, P.P., León, R.G., 2011. Weed management practices determine plant and
arthropod diversity and seed predation in vineyards. Weed Res. 51, 404–412.
Spanou, S., Tiniakou, A., Georgiadis, T., 2013. Exploring agroecosystem floristic and
vegetation diversity in a Mediterranean landscape. J Bio and Env Sci 3(3), 11-25.
Tatoni, T., 1991. Perception de l'évolution post-culturale des paysages de terrasses. Écologie
Humaine IX(2), 39-53.
UNESCO–SCBD, 2014. Florence Declaration on the links between biological and cultural
diversity. http://www.cbd.int/portals/culturaldiversity/docs/florence-report-2014-en.pdf
Termorshuizen, J.W., Opdam, P., 2009. Landscape services as a bridge between landscape
ecology and sustainable development. Landscape Ecology 24, 1037-1052.
Van der Maarel, E., 1993. Some remarks on disturbance and its relations to diversity and
stability. J. Vegetation Sci. 4, 733-736.
Draft manuscript published in Agriculture, Ecosystems and Environment, 212, pp. 175–186,
ISSN: 0167-8809 n°AGEE5069, DOI:10.1016/j.agee.2015.06.023
26
Van Zanten, B., Verburg, P., Espinosa, M., Gomez-y-Paloma, S., Galimberti, G., Kantelhardt,
J., Kapfer, M., Lefebvre, M., Manrique, R., Piorr, A., Raggi, M., Schaller, L., Targetti, S.,
Zasada, I. and Viaggi, D., 2014. European agricultural landscapes, common agricultural
policy and ecosystem services: a review, Agron. Sustain. Dev. 34(2), 309-325.
Warlop F., 2006. Limitation des ravageurs de l’olivier par le recours à la lutte biologique par
conservation. Cahiers Agriculture, vol. 15(5), 449-455.
Williams, N.S.G., Schwartz, M.W., Vesk, P.A., McCarthy, M.A., Hahs, A.K., Clemants, S.E.,
Corlett, R.T., Duncan, R.P., Norton, B.A., Thompson, K., McDonnell, M.J., 2008, A
conceptual framework for predicting the effects of urban environments on floras. J. Ecology
97(1), 4-9.
Zechmeister, H.G., Schmitzberger, I., Steurer, B., Peterseil, J., Wrbka, T., 2003. The influence
of land-use practices and economics on plant species richness in meadows. Biological
Conservation, 114(2), 165–177.
