Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Biodiversity conservation in cities: Defining habitat analogues for plant species of conservation interest

  • M. Itani,

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Visualization, Writing – original draft, Writing – review & editing

    Affiliations Department of Landscape Design and Ecosystem Management, Faculty of Agricultural and Food Sciences, American University of Beirut, Beirut, Lebanon, Nature Conservation Center, American University of Beirut, Beirut, Lebanon

  • M. Al Zein,

    Roles Conceptualization

    Affiliation Department of Biology, Faculty of Arts and Sciences, American University of Beirut, Beirut, Lebanon

  • N. Nasralla,

    Roles Visualization, Writing – original draft, Writing – review & editing

    Affiliation Nature Conservation Center, American University of Beirut, Beirut, Lebanon

  • S. N. Talhouk

    Roles Conceptualization, Methodology, Project administration, Resources, Supervision, Visualization, Writing – original draft, Writing – review & editing

    ntsalma@aub.edu.lb

    Affiliations Department of Landscape Design and Ecosystem Management, Faculty of Agricultural and Food Sciences, American University of Beirut, Beirut, Lebanon, Nature Conservation Center, American University of Beirut, Beirut, Lebanon

Biodiversity conservation in cities: Defining habitat analogues for plant species of conservation interest

  • M. Itani, 
  • M. Al Zein, 
  • N. Nasralla, 
  • S. N. Talhouk
PLOS
x

Abstract

Urban green spaces, both unmanaged and managed, include novel ecosystems that may be suitable habitat analogues for native plant species of conservation interest. The objective of this study was to define habitat analogues in the Mediterranean city of Beirut for Matthiola crassifolia, a Lebanese steno-endemic only present in urban habitats. We adopted a stepwise method that integrates two vegetation assessments, floristics, and life form. We placed seventy-eight quadrats (1m x 1m) in 12 study sites following a deliberate biased approach to capture habitat diversity. In every quadrat, we performed taxonomic identification and recorded life forms of each species. We pooled species that shared the same life form and estimated area cover for each life form accordingly. We performed TWINSPAN analyses on both floristic and life form data, then combined these findings to generate a description of habitat analogues suitable for M. crassifolia. TWINSPAN analysis of floristic data clustered the 78 quadrats under 17 quadrat groups, while life form data assembled the quadrats under 11 quadrat groups. The integration of floristic and life-form classification results into one matrix generated 30 quadrat groups, 8, which were highly favorable to M. crassifolia, and 12, which excluded it. The stepwise method unveiled similarities between vegetation assemblages, which appeared distinct due to the high presence of ruderals. We found that habitat analogues favorable to M. crassifolia include green spaces dominated by palms, low-lying succulents, or shrubs with scale-like leaves. In contrast, areas dominated by turf grass, canopy trees, or vegetation that produces significant litter were not favorable to M. crassifolia’s persistence. Based on these findings, we generated a plant palette of life forms which guides designs of urban habitats favorable to M. crassifolia.

Synthesis and applications

The stepwise method was useful in producing informative plant lists and assemblages for planting designs and landscape management; it generated a plant selection palette that is not restrictive and does not enforce a native only policy. It also offered a wide range of potential habitat analogues for M. crassifolia.

Introduction

Novel ecosystems are human-modified ecosystems that have been irreversibly altered by intense impacts on abiotic conditions or biotic composition [1, 2, 3]. Novel ecosystems include urban green spaces that emerge mostly after built structures have replaced previously existing ecosystems. As such they include non native vegetation assemblages, consisting of native, spontaneous, naturalized, and invasive species [4]. Urban green spaces are sometimes abandoned after human disturbances or continue to experience disturbance regimes and consequently contain a range of both early- and late-succession vegetation.

Both unmanaged and managed green space can potentially contribute to urban biodiversity conservation. When unmanaged, urban green spaces are referred to as Informal Green Spaces (IGS) and can potentially contribute to urban biodiversity conservation [4]. IGS can provide valuable habitats [5, 6, 7, 8, 9, 10], and occasionally serve as a substitute for natural habitats [11, 12]. Certain cities are important for the conservation of threatened species [13]. Urban green spaces in Mediterranean cities, for instance, where plant diversity and endemism are high, offer a prospective of refuges to plant species regardless of whether the urban green spaces are semi-natural or anthropogenic [14]. However, despite the persistence of endangered species in cities [15], there are only few reported case studies of cities hosting viable populations of rare or endangered species, and thus directly contributing to conservation efforts [16].

Urban habitats tend to favor the persistence of plant species with particular trait combinations that appear well suited to the conditions [17]. Furthermore, certain plant functional traits tend to increase in response to urbanisation, while other traits have mixed responses [18]. For example, it has been shown that urbanized grid cells favor wind pollinated plants, plants with scleromorphic leaves, or plant seeds dispersed by animals, while other grid cells favor insect pollinated plants, plants with hygromorphic leaves, or plant seeds dispersed by wind [19]. Also, acidophiles may have disappeared in urban areas [20]. Identifying predictable relationships between plant traits and environmental conditions or disturbances is a promising approach for understanding how plant communities change in response to human land-use modification [21].

Many plant species can be found more or less regularly in various city habitats; the region of a habitat associated with a particular ecological community. For example, classification of urban habitat types inside the city of Berlin has revealed 19 habitats particularly worthy of protection and these were nominated as legally protected [22]. However, the classification of habitat types inside cities requires standard habitat classification systems which have not been developed in all countries, at least not in Lebanon [23]. Furthermore, the nomination of such habitat types becomes challenging when urban habitats are privately owned, as is the case of most informal green spaces in Beirut [24]. Another challenge is the protection of such habitats in cities like Beirut where law enforcement is weak and is unlikely to deter against infringement [25].

More relevant to cities like Beirut are urban biodiversity strategies that proposed to transform urban habitats into habitats suitable for native plant conservation [26]. One example of urban biodiversity strategy is the use of species-rich herbaceous communities to promote biodiversity in cities [27]. Another strategy, referred to as reconciliation ecology, proposes the conversion of spaces assigned to human activities into spaces that support the persistence of native species [28]. Identifying habitat analogues in this case is essential to guide reconciliation ecology strategy in cities [29]. If appropriate conservation targets are set, habitat analogues may dilute the distinction between disturbed and non-disturbed habitats as favorable sites for plant conservation [30, 31].

There are various methods that describe vegetation based on species identity and abundance, species functional traits, structural characteristics, or degree of naturalness. Floristic surveys are one of two main vegetation description methods used to collect data on native species of conservation interest, and to generate community classification schemes and structure patterns which vary predictably in response to external factors such as environmental stress and disturbance [32]. The floristic method uses taxonomic identification and species abundance to describe vegetation. From a floristics perspective, plant species found in an area are unique and capable of coexisting as distinct, recognizable units that are repeated regularly in response to biotic and environmental variations [33, 34, 35, 36, 37].

The other main vegetation description method, physiognomy, is frequently used to describe vegetation according to external morphology, life form, stratification, and size of each species [38, 39, 40, 41, 42, 43, 44, 45, 46, 47].

EcoVeg is a recent method that combines floristics and physiognomy, in addition to ecological descriptors, and that applies different rationales depending on whether the vegetation is natural or cultural [48]. Combining both approaches may be necessary to generate informative data from sites subjected to different disturbance conditions. The application of floristics in urban habitats may present a challenge when interpreting the data, since many studies reported an over-representation of ruderal species and high taxonomic diversity between relatively close sites [49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59]. In contrast, the application of physiognomic and structural vegetation description, may be more useful in urban habitats [18, 60] as the data informs about predominance of life strategies adopted by different life forms, and the method is applicable in highly modified sites, and at both, macro- and micro-climate conditions [61, 62].

A widely used functional type or physiognomic vegetation description system is the life form classification of Raunkiær [39]. Although physiognomic and structural vegetation description methodologies were developed to describe vegetation over large areas, these methods have been considered as potentially more useful tools than floristics in highly modified sites because they better reflect life strategies, for example, the ruderal strategy, encompassed by certain life forms [61]. Down (1973) resorted to life-form in studying reclamation of spoil heaps [62].

There is consensus that physiognomic and physiological characteristics of plants, including species life-history strategies and population biology, are also important descriptors of vegetation communities [63, 64, 65, 66, 67, 68, 69, 70]. Plant communities were shown to be important indicators to determine suitable habitats for rare species [71], and for ecologically and economically important species [72]. Some of these studies, however, deliberately exclude disturbed areas from sampling [72].

One aspect of urban vegetation that might challenge classification is the abundance of ruderal plant species which, benefit from the absence of interspecific competition that normally occurs in later successional stages, and colonize bare and disturbed land [61]. By spreading from nearby semi-natural vegetation, ruderals contribute to high variability in urban plant diversity, even between close sites, limiting the value of vegetation classification using floristic methods [32]. Some of these ruderal species may be distantly related to agricultural weeds and others to plant species found across transportation networks [73]. Ruderals are also populating green walls in cities [74]. The overrepresentation of ruderals and the haphazard management of green spaces in Beirut make vegetation classification difficult due to the small scale of associated biotopes and abundance of structured biotope complexes.

The success of plant conservation strategies is highly influenced by perceptions and social preferences which should be taken into consideration in addition to field assessment challenges in cities. For example, studies have shown that spontaneous ‘unmanaged’ vegetation may not appeal to residents as aesthetically pleasing nor is it perceived as acceptable ‘urban nature’ by decision-makers [75, 76, 77]. This is further complicated by the fact that plant selection and management, is driven by landscape architects and landscape contractors who have limited experience with native species, and do not have clear guidelines on how to contribute to biodiversity conservation in cities [78, 79].

While several studies show how vegetation description using floristic assessments in urban areas is limited by over-representation of ruderal species, abundance of exotic plants, and a high taxonomic diversity between relatively close sites [49, 50, 51, 52, 73, 54, 55, 58, 80, 81], other studies suggest that descriptions of functional types, such as life form, may permit ecological comparisons among areas of similar ecology on a more general scale than would be possible when using a taxonomic approach [82, 83, 84]. For instance, structural and adaptation characteristics of beach and dune vegetation were found similar, even if their taxonomic spectrum differed [85]. Furthermore, life-form, among other descriptions of functional types, were associated with plant responses to environmental change, to plant competitive strength, and to plant effects on biogeochemical cycles and disturbance regimes [86]. Recently, life form and life history were found to be stronger predictors of underlying population processes than native status [87]. The first meta-analysis on intra-urban biodiversity variation worldwide showed that patch area and corridors have the strongest positive effects on biodiversity, and that vegetation structure, local scale, biotic factors, and management habitat variables, are significantly more important than landscape scale, abiotic factors, or design related variables [88].

The objective of this study was to define urban habitat analogues for a plant species of conservation interest, Matthiola crassifolia, which has persisted in varying abundance in the Mediterranean city of Beirut. M. crassifolia is a rare Lebanese steno-endemic, it is only present in urban habitats, and its largest population in Beirut is decreasing. As the natural habitat of M. crassifolia is described as coastal area rocks [89], we hypothesize that the expected habitat analogues for the target species will include urban green space typologies that include significant percent of bareground, minimal presence of plant litter, and vegetation that does not significantly produce shade.

Materials and methods

Study location

Located along the Eastern shores of the Mediterranean, Lebanon is a predominantly mountainous country consisting of five geomorphological regions namely, a narrow coast along the length of the country, two mountain ranges that run parallel to the sea, and a fertile high plain that separates the two mountain chains. Lebanon possesses botanical elements from temperate, arid and subtropical biomes.

Species of conservation interest and its distribution

There are four Matthiola species recorded in Lebanon, two of which are either national or regional endemics. The Species-Group Ovatifolia is represented by the regional endemic Matthiola damascena Boiss. The Species-Group Longipetala is represented by Matthiola tricuspidata and Matthiola longipetala. Species-Group Iincana is represented by the national endemic Matthiola crassifolia Boiss. & Gaill which is restricted to a few locations along the highly urbanized Lebanese coast and is the subject of this study. M. crassifolia is a taxon of conservation interest as the species is recognized as an endemic of Lebanon. However, Gowler (1998) has questioned the taxonomic status of the species proposing that it should be considered subspecies of Matthiola sinuata [90]. Even if future molecular analyses support this preference, the taxon will remain an endemic of Lebanon yet at the intra-specific level.

The most comprehensive record of the distribution and status of M. crassifolia prior to this study was by Rteil (2002) who performed a systematic survey of the Lebanese coast and recorded the presence of the species in three out of five previously reported sites, Beirut, Ras Beirut and Byblos [91]. In this study, Ras Beirut and Beirut were considered a single locality. Subsequent field investigations by added Sidon, Khaldeh and Amchit as localities for M. crassifolia [89]. Our field survey to all reported localities confirmed the extinction of M. crassifolia in Sidon and its continued presence in Khaldeh, Beirut, Amchit and Byblos [92].

Study area

Beirut (33.8869° N, 35.5131° E), the capital of the Republic of Lebanon, is located on the eastern coast of the Mediterranean. Archeological evidence shows that humans have continuously occupied Beirut for the last 5000 years [93, 94]. Today, the city of Beirut has one of the highest urban densities in the Middle East with an area roughly over 20 km2, population density is estimated at 21,000 people per sq. km [95, 96]. The topography of the city includes two hills, Achrafieh (100 m asl) and Mousseitbeh (80 m asl) [97]. Paul Mouterde, who conducted floristic studies in Beirut in the 20th century, reported 1200 floral species including native and non-native species [98].

Our study site, Beirut, is defined by a 6 km long and 2 km wide cape [99]. Today, this area consists of densely populated neighborhoods interspersed with managed landscapes and zones with spontaneous naturalized vegetation occurring within geographically adjacent lots. Recent floristic studies of semi natural areas of Beirut revealed low community similarity, patchy species distribution, and predominance of habitat non-specific species [81]. Green spaces in the southern part of the promontory of Beirut fall under two broad categories; managed landscapes, dominated by exotic ornamental species planted in raised beds with reconstructed soil, and spontaneous landscapes where spontaneous floral communities survive along with casual non-native species, in coastal cliffs, along the rocky water front, and in un-built/abandoned lots [100].

The study location, particularly the southern part of the promontory of Beirut, can be considered a type III city that is likely to be carrying an extinction debt because extensive landscape transformations occurred after initial floristic surveys [80]. Although the expansion of the city started in 1840, the city passed through five stages of transformation when the southern part of the promontory consisted mostly of semi-natural areas until 1943 [101]. The earliest botanical studies in the region took place in the mid-1800 [102] and continued giving considerable focus to Beirut and its environments till the 1930s [103; 104].

Concurrent with early botanical studies of semi natural areas in Beirut, since 1840, Beirut has passed through five phases of transformation which extensively altered its landscape (Fig 1) [101]. Today, Beirut still harbors significant remnant native vegetation, especially the southern side of the promotary where urban expansion took place after 1970.

thumbnail
Fig 1. Beirut phases of urbanisation (From “Beirut,” by N. Yassin, 2010, Cities, 29, p. 64–73. Reprinted with permission.).

https://doi.org/10.1371/journal.pone.0220355.g001

Field data collection

Baseline data collection was initiated in 2012, three years before the start of the study, to ensure comprehensive coverage of all informal green space locations. During 2012, all informal green spaces in the study location were surveyed to locate all spaces where M. crassifolia was present. In subsequent years (2013 and 2014), during flowering season of the target species, annual visits were made to all identified green spaces, regardless of whether the species was present or not. In 2015, sites were selected according to management intensity (high or low management intensity) of anthropogenic sites, and to levels of diversity of semi-natural habitats. Fig 2 presents the distribution of M. crassifolia in Beirut as well as the location of the selected sites for the study. All visited sites were either accessible public spaces or abandoned private lots. No permits were needed as no plant material was collected from study sites: only plant voucher specimens were collected from the field for taxonomic identification.

thumbnail
Fig 2. The distribution of M. crassifolia in Beirut and selected sites for the study.

https://doi.org/10.1371/journal.pone.0220355.g002

As our objective is to capture habitat diversity of a rare species in small plots,we used a deliberate biased method to select study locations and to lay out sampling quadrats [71]. In their attempt to compare the effects of random to non-random sampling on patterns of species abundance, species richness and vegetation-environment relationships, Diekmann et al. (2007) concluded that random sampling resulted in a larger number of common species, and a smaller number of rare species when compared to non-random sampling. They also found that for small plots, the number of species in the non-randomly placed plots was higher than in the randomly placed plots, and that in random sampling, there was considerable redundancy [105].

We set a total of 78 quadrats in 12 sites. In vegetation patches with clearly visible boundaries, one to two 1 m × 1 m quadrats were placed [32]. We placed larger quadrats, 2 m × 2 m, in locations where shrubs are present [32]. As in Dinsdale, deliberate bias method consists of placing quadrats in areas judged representative of the selected location and for capturing the maximum observed variation [71,106]. We made three modifications to the sampling technique to address site-specific challenges. First, when the boundary of a plant assemblage was not clearly defined due to site disturbance, we placed quadrats in locations where the target species was highly represented and at least one quadrat where the target species was minimally represented assuming that this location constitutes the boundary of the sampled plant assemblage. Second, when species had an ‘individualistic’ distribution pattern, we increased the number of quadrats, up to six, to capture the observed variation and compensate for the difficulty to define the boundaries of the plant assemblage [107]. Third, when the target species was consistently not present in a given vegetation assemblage, we placed one quadrat.

We divided each quadrat into a grid of 100 subunits to ensure speed of measurement and relative accuracy [108, 109, 110, 111]. In every quadrat, we determined percent cover using the 11-point Domin cover scale by visually assessing subunits as: fully covered, empty, and partially covered for each species and each life form [32]. Data obtained from all subunits within a quadrat was then added to determine Domin cover per quadrat.

As the analysis of non-randomly placed plots such as phytosociological quadrats may be biased, especially regarding estimates of species abundance and species richness patterns [105] ordination was not attempted to analyse vegetation-environment relationships in this study.

Taxonomic and life form identification

We identified each plant specimen by consulting published floras, voucher specimens at the American University of Beirut Herbarium (Post Herbarium), and photographic floras [104, 98, 89]. All identified species were described by their life form according to Ellenberg and Mueller-Dombois, with amendment to include bunched shoot arrangement in reptant hemicryptophytes which forms a partially decomposed thick mat and causes peat accumulation [40]. We then pooled species that shared the same life form under the one category and estimated area cover for each life form accordingly.

Analysis

Based on the 11-point Domin cover scale, we analyzed floristic data, species and percent species cover, using TWINSPAN [112]. Also called dichotomized ordination analysis, the Two Way INdicator SPecies ANalysis is a method for classifying communities according to hierarchical divisions based on progressive refinement of a single ordination axis of a (sites × species) data matrix [113]. Using the same tool, TWINSPAN, we analyzed the life form data, life-form categories and percent cover (as relative abundance of each life form within each quadrat). In the TWINSPAN, the cut levels 0-3-4-5-6-8 were applied. The TWINSPAN groups were characterized by constancy-percentage, average cover, and representation of target species. A matrix, integrating floristic and physiognomic TWINSPAN findings, was then created to find intersections between quadrat groups defined by classifying life form and floristic data sets. This process led to the identification of new classification that consisted of quadrat groups that share similar life form and species composition. The full dataset can be found in [92]. A conceptual extrapolation of these findings allowed us to define landscape plant typologies with vegetation assemblages similar to quadrat groups in which the target species is highly represented and we considered these typologies as suitable locations for the introduction of M. crassifolia.

Results

M. crassifolia is most widely distributed in Beirut; based on our field surveys its presence was confirmed in 73 sites of which only one site, Pigeon Rock, is protected by law, and another site, the limestone cliff facing Pigeon Rock, is almost inaccessible and may be considered de facto protected. The remaining 71 sites offer highly diverse habitats and are not protected [92]. In remnant semi-natural sites, M. crassifolia is found in, spiny Mediterranean heaths, screes, sea cliffs and rocky offshore islands, growing on both sandstone and limestone formations and on (stabilized) coastal sand dunes. In anthropogenic sites, it grows near open sewers, in abandoned dump sites, through cracks in concrete walls and asphalt, on heaps of gravel, in street medians and on two occasions, almost epiphytically, out of the trunks of date and fan palms. The species’ tendency to utilize modified habitats reflects its partial behavior as a ruderal [61]. During the course of this study, M. crassifolia was lost in 20 sites to urban development including one site which harbored the largest clump count, and only four of these sites were recolonized. As a result, the plant species’ range in Beirut was reduced by 17% between 2012 and 2015 [92].

We recorded the presence of 124 plant species belonging to 107 genera and 40 families in the 78 sampled quadrats [92]. Plant species co-occurring with M. crassifolia included 16% non-native species. Analysis of floristic data by TWINSPAN clustered the 78 quadrats under 17 quadrat groups labeled af to qf. M. crassifolia had the highest constancy and abundance in three groups, df, gf and if. In contrast, the species was not present in eight groups, cf, ff, kf, mf, nf, of, pf and qf. The low community similarity, patchy species distribution, and predominance of habitat non-specific species reported by Talhouk et al. (2005) in their study of the floristics of the Lebanese coast was confirmed in this study [81]. High floristic variability between and within different sites resulted in more than half the quadrat groups (58.8%) consisting of no more than two quadrats. Only one group (ef) consisted of a large number of quadrats and represented a perceptible community of sparse vegetation on sandstone outcrops. Other groups were not site specific, but included quadrats exposed to similar disturbance; for example, in group gf the nine quadrats were sampled from street medians and side walks and consisted of a combination of evergreen exotic ornamental species such as Agave americana, Agave attenuata, and Lampranthus multiradiatus. Similarly, tf included quadrats characterized by a high representation of graminoids, Cyperus rotundus and Cynodon dactylon, which often grow in gardens and street medians under and around evergreen ornamentals such as the shrub Pittosporum tobira, and the creeping herbaceous forb Sphagneticola trilobata.

One problem we encountered with floristics based TWINSPAN analysis is that many groups did not represent actual communities i.e. plant species in an area that are unique and capable of coexisting as distinct, recognizable units that are repeated regularly in response to biotic and environmental variations [33, 34, 35, 36]. For example, group ef, which included about 28% of sampled quadrats, consisted of several distinct vegetation assemblages that occur in different habitats, both semi-natural and anthropogenic, and the target species, a stress-tolerant ruderal, was the only common indicator species between these assemblages.

Life form description of plant species yielded 55 different life forms. Results revealed that more than half of all recorded species were therophytes with a total of 64 autotrophic therophyte and two heterotrophic annual vascular parasites. The high representation of therophytes reflects high disturbance of study sites [61].

Fig 3 presents the life-form spectrum of all species recorded in the 78 plots. Chamaephytes constituted the most prominent perennial life form and included 24 species. Over half of all chamaephytes were either regional or national endemics and only three were not native. Phanerophytes were represented by 14 species, 10 of which were not native. Perennials characterized by a periodic shoot reduction were represented by 15 hemicryptophytes and six geophytes.

thumbnail
Fig 3. Raunkiaer life-form spectrum of plant species recorded in 78 quadrats in 12 sites in Ras Beirut.

https://doi.org/10.1371/journal.pone.0220355.g003

Analysis of life form data by TWINSPAN clustered the 78 quadrats under 11 quadrat groups labeled

Al to Kl (Table 1). M. crassifolia was highly represented in three of these groups (Cl, Dl, and El) with a percent cover ranging between 11% and 25% in almost all quadrats within these groups. Examples of life forms in these three groups include, unbranched dwarf palm like trees (Phanerophyte08), typical and tall evergreen dwarf-shrubs (Chamaephyte03 & Chamaephyte04), low reptant evergreen succulents (Chamaephyte14), tall drought-deciduous hemicryptophytes (Hemicryptophyte01) and small reptant evergreen hemicryptophytes (Hemicryptophyte03) were common. Ornamental examples of these life forms include Agave and Yucca species (Phanerophyte08), cultivated Sea Lavender species (Chamaephyte03 and Chamaephyte04), and Lampranthus multiradiatus (Chamaephyte13).

thumbnail
Table 1. TWINSPAN analysis of life form data set collected in Ras Beirut.

(Alphabetical naming of quadrat groups by floristic and life form classification are not related.).

https://doi.org/10.1371/journal.pone.0220355.t001

Five groups (Al, Hl, Il, Jl, Kl) excluded the target species and the dominant life form in these groups was mostly phanerophytes. These groups include mesophyllous large evergreen trees with spherical crown restricted to their upper half (Phanerophyte01), mesophyllous normal-sized evergreen shrubs with spherical crown extending to near their base (Phanerophyte04), microphyllous normal-sized evergreen shrubs with spherical crown extending to near their base (Phanerophyte03), and mesophyllous tall deciduous shrub with spherical crown extending to near the base of the shrub (Phanerophyte07). Ornamental examples of these life forms include various shade trees (Phanerophyte01), and shrubs used as hedges such as Pittosporum tobira (Phanerophyte04 and Phanerophyte03). They also include typical evergreen reptant herbaceous chamaephytes (Chamaephyte12) and ornamental plant species belonging to this life form and similar life forms such as turfgrass species and the Singapore Daisy, Sphagneticola trilobata.

The integration of floristic and life-form classification results into one matrix to identify quadrats at the intersection of both classifications generated a new set of quadrat groups that shared similar life form and species composition, and where M. crassifolia presented similar constancy and abundance (Table 2). This stepwise approach generated 30 quadrat groups, 8 which were highly favorable to M. crassifolia, and 12 which excluded it. We then proceeded to describe life form and species prevalent in these groups.

thumbnail
Table 2. Matrix of floristic and life-form classifications of quadrats from plant data set collected in Ras Beirut and southern part of the promontory of Beirut.

Intersections show M. crassifolia represented by constancy and abundance and help define favorable and unfavorable vegetation assemblages.

https://doi.org/10.1371/journal.pone.0220355.t002

The intersections that resulted in quadrat groups with the highest representation of M. crassifolia belonged to 4 out of 11 quadrat groups that were derived from the classification of the life form data set (Cl, Dl, El and Fl) and 4 out of 17 quadrat groups that were derived from the classification of the floristic data set (af, df, gf and if) (Table 3).

thumbnail
Table 3. Description of urban plant habitat analogues (habitat condition, life forms, plant habitat, and species) for M. crassifolia in Beirut following a stepwise approach that intersects floristic and life form data classifications.

https://doi.org/10.1371/journal.pone.0220355.t003

The intersections that resulted in quadrat groups with the lowest representation of the target species belonged to 8 out of 11 quadrat groups that were derived from the classification of the life form data set and 11 out of 17 quadrat groups that were derived from the classification of the floristic data set (Table 4).

thumbnail
Table 4. Description of urban plant habitats (habitat condition, life forms, plant habitat, and species) unsuitable for M. crassifolia in Beirut following a stepwise approach that intersects floristic and life form data classifications.

https://doi.org/10.1371/journal.pone.0220355.t004

Discussion

The similarity in the infrastructure of a city may explain homogeneity of urban ruderal species, which out-compete sown species [114]. For example, a 30-year green roof study concluded that spontaneous colonization should be accepted and considered as a design factor; and regional plant communities could serve as a model for seed recruitment and installations [114]. However, preventing a rapid loss of area-sensitive species necessitates large sites greater than 50 ha [88]. Such areas are absent in Beirut, and the loss of Matthiola crassifolia in the city is highly likely. Utilizing habitat analogues to increase the area of habitat patches and create a network of corridors is the most plausible strategy to ensure the persistence of this narrow endemic.

Urban environments share many features in common because they are designed to perform standard functions to meet human needs [115]. Such an environment, common to cities around the world, might be expected to select for species with a similar suite of traits favouring persistence in highly disturbed and human-modified habitats. Indeed the process of urbanization has been conceptualized as a series of filters acting on an existing species pool and selectively removing those species with traits unfavourable for persistence in this new environment [59].

The peculiarity of our study is that, not only is classification influenced by ruderals, but the species of conservation interest M. crassifolia also behaves as a ruderal. Considering the diversity of habitats occupied by M. crassifolia, it was not possible to resolve this lack of location specificity with floristic assessments, which in turn did not allow us to develop an understanding of urban habitat analogues. Instead, the number of quadrat groups generated by the floristic analysis was large, and some of these clusters did not represent actual plant community assemblages. Although the natural habitat of the target species is described as coastal area rocks [89], the behavior of the target species as a ruderal led to TWISPAN quadrat groups of highly variable species constituency and quadrat locations ranging from highly managed street medians to semi-natural coastal cliffs and including commercially introduced species and native ones.

Classifying life form data by including percent cover for each category helped specify which life forms and their respective abundance were positively or negatively associated with M. crassifolia. Our findings are in line with Kent [32], who emphasized that physiognomy might be more useful as a tool than floristics in highly modified habitats at different scales due to the responses of plant species to macro- and micro-climate conditions. Life history and life form are stronger predictors of underlying population processes than native status [87,116].

By using a stepwise approach which combines the two methods, floristics and physiognomy, we were able to minimize the masking effect of ruderal species and to identify life form similarities within distinct vegetation assemblages. In the last decade, researchers have combined life form and floristic vegetation description methods to overcome difficulties in analyzing data in disturbed habitats. For example, Vestergaard [117] generated quadrat groups based on floristic data through TWINSPAN and then described the life-form spectra in each to investigate the relationship between plant diversity and artificial dune development processes. Although similar to our methodology, Vestergaard did not use this combined methodology to define habitat analogues for target plant species. In 2014, a new vegetation classification approach that relies on both physiognomy and floristics over large areas was published under the name EcoVeg [48]. Our approach, however, differs from EcoVeg in that we first mathematically classify physiognomic data and later sort the classifications according to a specific floristic trend. In addition, we base our study on field data collected from small urban habitat sites while EcoVeg uses map data and is meant to classify vegetation over large natural landscapes.

More recently, several studies have sought to explore the potential of light detection and ranging (LiDAR) to inform landscape biodiversity assessments. In fact, the utilization of this technology has developed from quantification of gaps (above bare ground, low vegetation and medium vegetation), canopy cover and its vertical density in open landscapes [118] to mapping tree cover and vegetation spatial and vertical structure in cities and estimating above ground biomass despite particular challenges posed by urban areas [119,120].

Furthermore, accurate mapping of vegetation communities within highly disturbed urban landscapes was recently achieved through incorporating a hierarchical object-based image analysis (OBIA) approach with high-spatial resolution imagery and canopy height surfaces derived from LiDAR data [121]. Provided the range of outputs these recent methods are producing, LiDAR technology may serve for rapid indentification of potential locations for habitat analogues and the exclusion of areas that are known not to be favorable to the target species, for example canopy cover in the case of M. crassifolia.

Improving the quality of existing green spaces throughout the entire urban matrix has been suggested as an effective approach to enhancing biodiversity experience [122]. The information we generated using a stepwise approach integrating floristics and physiognomy, may serve as blueprints for planting designs; it offers a plant selection palette that is not restrictive and does not enforce a native only policy. The habitat conditions in quadrat groups of high representation of the target species were not the same and reflected a wide range of potential habitat analogues for M. crassifolia. These varied from abandoned buildings to highly managed street medians.The urban habitat analogues that we identified include green spaces dominated by palms, low-lying succulents, or shrubs with scale-like leaves. In contrast, the species does not seem to persist in green spaces dominated by turf grass, canopy trees, or vegetation that produces a significant litter. Furthermore, since knowledge of a target species’ preferred physiognomies includes an understanding of its position in the vertical stratification of its ecological community [32], we were able to identify additional habitats suitable for the introduction of M. crassifolia. Streetscapes, such as street medians, sidewalks and street tree planters, that lack both peat accumulating ground cover and canopy species, ubiquitous throughout the city, provide patches of optimal vegetation composition that could potentially accommodate M. crassifolia. Such streetscapes that can function as habitat analogues for M. crassifolia are illustrated in Fig 4 ana Fig 5. A change in landscape management strategy, however, needs to preceed design and development of habitat analogues. At the end of the four-year study, M. crassifolia was no longer seen in 16 out of 73 sites. Our field observations, revealed that management strategies such as the intentional uprooting of M. crassifolia considered by gardeners as a weed led to the disappearance of the species from these locations. On the other hand, there are locations where the species persists probably due to the fact that in these sites gardeners remove plants during their dieback stage, which includes seed-bearing silique fruits, but they keep seedlings and flowering plants.

thumbnail
Fig 4. Illustrated scene for a planted street median functioning as a habitat analogue for M. crassifolia.

Rosulate phanerophytes and reptant succulent chamaephytes, often used as ornamentals in green spaces in Beirut, dominate the street median without excluding the target species.

https://doi.org/10.1371/journal.pone.0220355.g004

thumbnail
Fig 5. Illlustrated scene for sidewalk functioning as a habitat analogue for M. crassifolia.

The cracks in the concrete of the sidewalk due to poor management and the adjacent sandstone wall resemble coastal cliffs occupied by the species. Small and medium-sized therophytes like Plantago coronopus L. abd Polycarpon tetraphyllum are often observed occupying such spaces.

https://doi.org/10.1371/journal.pone.0220355.g005

In the case where more than one species is a conservation target, then a conservation strategy conducive to the persistence of both species. In our study, we found that M. crassifolia persisted as part of the low shrub layer under taller nanophyllous shrubs like the Shaggy sparrow-wort, Thymalea hirsuta, another species of conservation interest in Lebanon. M. crassifolia also thrived in the understory of tuft-trees like the fan palm, Washingtonia robusta, an introduced species, and within groves of the giant reed, Arundo donax, a spreading native. Species belonging to these life forms, or similar ones, dominate many sites in Beirut including street medians and could serve as favorable habitats for M. crassifolia. Our findings also show that some exotic invasive species impacted M. crassifolia positively. M. crassifolia grew in sites dominated by Carpobrotus edulis, a potentially invasive in Lebanon, planted at the edge of pedestrian paths. Pedestrians avoided stepping onto these areas, maybe due to their appreciation of C. edulis as an evergreen ground cover [123]. As a result, this plant assemblage protected M. crassifolia and allowed C. edulis to spread constrained by water availability. Removal of invasive plant species should be determined based on its impact on endemic and rare vegetation present in a given region, and eradication should focus on those invasive species that compete with endemic species in general and those of conservation interest especially [124]. Huenneke and Thomson [125] suggest criteria for determining whether such species pose problems for specific rare native taxa and indicated the possibility that some species may be beneficial to endemics.

Equipped with the findings above, landscape designers, architects, and managers can better reconcile between desired conservation targets and, socio-behavioral, and aesthetic outcomes by including M. crassifolia in an aesthetically pleasing setting. They can design urban habitat analogues that promote the persistence of M. crassifolia by excluding from the plant palette native or non-native species belonging to life forms associated with its low representation as reported in this study. Alternatively, they can design an urban habitat analogue using a vegetation architecture conducive to the persistence of M. crassifolia. In the case of established green spaces, they can manage the space to become suitable for M. crassifolia by selectively removing species with a life form that is incompatible or that restricts its abundance. In some situations, horticultural techniques, such as pruning, can modify the micro environment without changing species existing on site, to create suitable urban habitat analogues; for example, improving light conditions in cases where species of conservation interest is shade intolerant.

Identifying predictable relationships between plant traits and environmental conditions provides a promising framework for understanding how vegetation responds to environmental change in a variety of ecosystems [21].

Conclusion

Given the rate of expansion of urban landscapes [126, 127, 128, 129], increasing a target species’ site area in a city is highly desired [28]. Our findings may serve as guidance on how to create or modify, through landscape planting designs, suitable habitats for species of conservation interest. By understanding the physiognomy and structure, and environmental conditions in which a species occurs, green areas may be designed to suit the requirements of a target species while established areas may be surveyed for candidate sites suited for the introduction of a target species. Our stepwise approach offers a detailed field assessment tool for urban plant habitat analogue characterization.

Acknowledgments

This paper is derived from the dissertation submitted by M. Itani in partial fulfillment of the requirements

for the MSc. degree at the American University of Beirut. We thank Drs. R. Zurayk, N. Farajalla, and K. Knio for their inputs throughout the study. We thank K. Mohamed, A. Jammool, S. El Masri, O. El Tal, R. Atallah and N. Halabi for their in field data collection. We thank M. Bou Kanaan for producing the illustrations.

References

  1. 1. Planchuelo G, von Der Lippe M, Kowarik I. Untangling the role of urban ecosystems as habitats for endangered plant species. Landscape and Urban Planning. 2019 Sep 1;189:320–34.
  2. 2. Hobbs RJ, Higgs E, Harris JA. Novel ecosystems: implications for conservation and restoration. Trends in Ecology & Evolution. 2009 Nov 1;24(11):599–605.
  3. 3. Kowarik I, von der Lippe M. Plant population success across urban ecosystems: A framework to inform biodiversity conservation in cities. Journal of Applied Ecology. 2018 Sep;55(5):2354–61.
  4. 4. Rupprecht CD, Byrne JA, Garden JG, Hero JM. Informal urban green space: A trilingual systematic review of its role for biodiversity and trends in the literature. Urban Forestry & Urban Greening. 2015 Jan 1;14(4):883–908.
  5. 5. Brandes D. Flora und vegetation von Stadtmauern. Tuexenia. 1992;12:315–39.
  6. 6. Colla SR, Willis E, Packer L. Can green roofs provide habitat for urban bees (Hymenoptera: Apidae)?. Cities and the Environment (CATE). 2009;2(1):4.
  7. 7. Brown RN, Sawyer CD. Plant species diversity of highway roadsides in southern New England. Northeastern Naturalist. 2012 Mar;19(1):25–42.
  8. 8. Dallimer M, Rouquette JR, Skinner AM, Armsworth PR, Maltby LM, Warren PH, Gaston KJ. Contrasting patterns in species richness of birds, butterflies and plants along riparian corridors in an urban landscape. Diversity and Distributions. 2012 Aug;18(8):742–53.
  9. 9. Rebele F. Results of floral surveys of industrial areas in West Berlin. Landschaft und Stadt 20, 49–66
  10. 10. Winter M. Changes in the urban flora of the commercial harbor in Bremen over the past 18 years (Master thesis, University of Bremen, Bremen).
  11. 11. Joger HG. Untersuchungen über die Tierwelt einer Stadtmauer. Zoologische Jahrbücher. Abteilung für Systematik, Ökologie und Geographie der Tiere. 1988;115(1):69–91.
  12. 12. Kaupp A, Brenneisen S, Klausnitzer B, Nagel P. Eco-faunistic characteristics of the beetle fauna of vegetated roofs (Insecta: Coleoptera). Entomol Blätter. 2004; 100: 47–83
  13. 13. Ives CD, Lentini PE, Threlfall CG, Ikin K, Shanahan DF, Garrard GE, Bekessy SA, Fuller RA, Mumaw L, Rayner L, Rowe R. Cities are hotspots for threatened species. Global Ecology and Biogeography. 2016 Jan;25(1):117–26.
  14. 14. Kantsa A, Tscheulin T, Junker RR, Petanidou T, Kokkini S. Urban biodiversity hotspots wait to get discovered: The example of the city of Ioannina, NW Greece. Landscape and Urban Planning. 2013 Dec 1;120:129–37.
  15. 15. Schmidt KJ, Poppendieck HH, Jensen K. Effects of urban structure on plant species richness in a large European city. Urban Ecosystems. 2014 Jun 1;17(2):427–44.
  16. 16. Shwartz A, Turbé A, Julliard R, Simon L, Prévot AC. Outstanding challenges for urban conservation research and action. Global Environmental Change. 2014 Sep 1;28:39–49.
  17. 17. Duncan RP, Clemants SE, Corlett RT, Hahs AK, McCarthy MA, McDonnell MJ, Schwartz MW, Thompson K, Vesk PA, Williams NS. Plant traits and extinction in urban areas: a meta‐analysis of 11 cities. Global Ecology and Biogeography. 2011 Jul;20(4):509–19.
  18. 18. Williams NS, Hahs AK, Vesk PA. Urbanisation, plant traits and the composition of urban floras. Perspectives in Plant Ecology, Evolution and Systematics. 2015 Feb 1;17(1):78–86.
  19. 19. Knapp S., Kühn I., Wittig R., Ozinga W.A., Poschlod P. & Klotz S. (2008) Urbanization causes shifts in species' trait state frequencies. Preslia 80, 375–388
  20. 20. Knapp S, Kühn I, Bakker JP, Kleyer M, Klotz S, Ozinga WA, Poschlod P, Thompson K, Thuiller W, Römermann C. How species traits and affinity to urban land use control large‐scale species frequency. Diversity and Distributions. 2009 May;15(3):533–46.
  21. 21. Westoby M, Wright IJ. Land-plant ecology on the basis of functional traits. Trends in ecology & evolution. 2006 May 1;21(5):261–8.
  22. 22. Köstler et al. Biotoptypen Berlins [Biotope types of Berlin], prepared for the Senate Department for Urban Development Berlin. 2003.
  23. 23. Lebanese Ministry of Environment. Sixth National Report of Lebanon to the Convention on Biological Diversity. 2019.
  24. 24. Nazzal M, Chinder S. Lebanon Cities’ Public Spaces. The Journal of Public Space. 2018 Apr 30;3(1):119–52.
  25. 25. Masri R. Environmental challenges in Lebanon. J Dev Soc 1997;13:73–115.
  26. 26. Breuste JH. Investigations of the urban street tree forest of Mendoza, Argentina. Urban Ecosystems. 2013 Dec 1;16(4):801–18.
  27. 27. Bretzel F, Vannucchi F, Romano D, Malorgio F, Benvenuti S, Pezzarossa B. Wildflowers: From conserving biodiversity to urban greening—A review. Urban Forestry & Urban Greening. 2016 Dec 1;20:428–36.
  28. 28. Rosenzweig ML. Win-win ecology: how the earth's species can survive in the midst of human enterprise. Oxford University Press on Demand; 2003
  29. 29. Lundholm JT, Richardson PJ. MINI‐REVIEW: Habitat analogues for reconciliation ecology in urban and industrial environments. Journal of Applied Ecology. 2010 Oct;47(5):966–75
  30. 30. Miller JR. Biodiversity Conservation and the Extinction of Experience. TRENDS in Ecologyand Evolution. 2005: 20, 430–434.
  31. 31. Gaston KJ. Biodiversity and extinction: species and people. Progress in Physical Geography. 2005 Jun;29(2):239–47.
  32. 32. Kent M. Vegetation description and data analysis: a practical approach. John Wiley & Sons; 2011 Nov 14.
  33. 33. Clements FE. Plant succession: an analysis of the development of vegetation. Carnegie Institution of Washington; 1916.
  34. 34. Clements FE. Plant succession and indicators. 1928.
  35. 35. Gleason HA. The structure and development of the plant association. Bulletin of the Torrey Botanical Club. 1917 Oct 1;44(10):463–81.
  36. 36. Gleason HA. The individualistic concept of the plant association. Bulletin of the Torrey botanical club. 1926 Jan 1:7–26.
  37. 37. Gleason HA. The individualistic concept of the plant association. American Midland Naturalist. 1939 Jan 1;21(1):92–110.
  38. 38. Raunkiaer C. The life forms of plants and statistical plant geography; being the collected papers of C. Raunkiaer. The life forms of plants and statistical plant geography; being the collected papers of C. Raunkiaer. 1934.
  39. 39. Raunkiær C. Plant life forms. The Clarendon Press; 1937
  40. 40. Ellenberg H, Mueller-Dombois D. A key to Raunkiaer plant life forms with revised subdivision. Berlin Geobotanical Institute ETH, Stiftung. 1967;37:56–73.
  41. 41. Ellenberg H. Indicator values of vascular plants in central Europe. Indicator values of vascular plants in central Europe. 1974;9.
  42. 42. Dansereau P. Description and recording of vegetation upon a structural basis. Ecology. 1951 Apr;32(2):172–229.
  43. 43. Dansereau P. Biogeography. An ecological perspective. Biogeography. An Ecological Perspective. 1957.
  44. 44. Kuchler AW. Vegetation mapping. Vegetation mapping. 1967.
  45. 45. Fosberg FR. A classification of vegetation for general purpose. Trop. Ecol. 1961;2:1–28.
  46. 46. Elton CS, Miller RS. The ecological survey of animal communities: with a practical system of classifying habitats by structural characters. Journal of Ecology. 1954 Jul 1;42(2):460–96.
  47. 47. Elton CS. The pattern of animal communities, Methuen and Co. Ltd.: London, UK. 1966.
  48. 48. Faber-Langendoen D, Keeler-Wolf T, Meidinger D, Tart D, Hoagland B, Josse C, Navarro G, Ponomarenko S, Saucier JP, Weakley A, Comer P. EcoVeg: a new approach to vegetation description and classification. Ecological Monographs. 2014 Nov;84(4):533–61.
  49. 49. Shenstone J. C. (1912). The flora of London building sites. Journal of Botany, 50, 117–124.
  50. 50. Salisbury EJ. The flora of bombed areas.1943.
  51. 51. Davis BN. Wildlife, urbanisation and industry. Biological Conservation. 1976 Dec 1;10(4):249–91.
  52. 52. Nature Conservancy Council. Nature Conservation in Urban Areas: Challenge and Opportunity: a Discussion Paper. The Council; 1979.
  53. 53. Whitney GG, Adams SD. Man as a maker of new plant communities. Journal of Applied Ecology. 1980 Aug 1:431–48.
  54. 54. Gilbert OL. Urban Commons. InThe Ecology of Urban Habitats 1989 (pp. 68–109). Springer, Dordrecht.
  55. 55. Kent M, Stevens RA, Zhang L. Urban plant ecology patterns and processes: a case study of the flora of the City of Plymouth, Devon, UK. Journal of Biogeography. 1999 Nov;26(6):1281–98.
  56. 56. Pyšek P, Chocholousková Z, † Pyšek A, Jarošík V, Chytrý M, Tichý L. Trends in species diversity and composition of urban vegetation over three decades. Journal of Vegetation Science. 2004 Dec;15(6):781–8.
  57. 57. Thompson K, Hodgson JG, Smith RM, Warren PH, Gaston KJ. Urban domestic gardens (III): composition and diversity of lawn floras. Journal of Vegetation Science. 2004 Jun;15(3):373–8.
  58. 58. Loram A, Thompson K, Warren PH, Gaston KJ. Urban domestic gardens (XII): the richness and composition of the flora in five UK cities. Journal of Vegetation Science. 2008 Jun;19(3):321–30.
  59. 59. Williams NS, Schwartz MW, Vesk PA, McCarthy MA, Hahs AK, Clemants SE, Corlett RT, Duncan RP, Norton BA, Thompson K, McDonnell MJ. A conceptual framework for predicting the effects of urban environments on floras. Journal of Ecology. 2009 Jan;97(1):4–9.
  60. 60. Lavorel S, McIntyre S, Landsberg J, Forbes TD. Plant functional classifications: from general groups to specific groups based on response to disturbance. Trends in Ecology & Evolution. 1997 Dec 1;12(12):474–8.
  61. 61. Grime JP. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. The American Naturalist. 1977 Nov 1;111(982):1169–94.
  62. 62. Down CG. Life form succession in plant communities on colliery waste tips. Environmental Pollution (1970). 1973 Jul 1;5(1):19–22.
  63. 63. Grime JP. Vegetation classification by reference to strategies. Nature. 1974 Jul;250(5461):26.
  64. 64. Grime JP. Ecology sans frontieres. Oikos. 1993 Dec 1:385–92.
  65. 65. Grime JP. Plant strategies, vegetation processes, and ecosystem properties. John Wiley & Sons; 2006 Aug 11.
  66. 66. Harper JL. Population biology of plants. Population biology of plants. 1977.
  67. 67. Tilman D. Resource competition and community structure. Princeton university press; 1982.
  68. 68. Tilman D. Plant strategies and the dynamics and structure of plant communities. Princeton University Press; 1988 Mar 21.
  69. 69. van der Maarel E. Vegetation ecology–an overview. Vegetation Ecology. 2005:1–51.
  70. 70. Silvertown J, Charlesworth D. Introduction to plant population biology.,(Blackwell Science, Oxford). 2001.
  71. 71. Dinsdale J, Dale P, Kent M. The biogeography and historical ecology of Lobelia urens L.(r)(the heath lobelia) in southern England. Journal of Biogeography. 1997;24(2):153–75.
  72. 72. Siddiqui MF, Ahmed M, Wahab M, Khan N, Khan MU, Nazim K, Hussain SS. Phytosociology of Pinus roxburghii Sargent (chir pine) in lesser Himalayan and Hindu Kush range of Pakistan. Pak. J. Bot. 2009 Oct 1;41(5):2357–69.
  73. 73. Haigh MJ. Ruderal communities in English cities. Urban Ecology. 1980 Mar 1;4(4):329–38.
  74. 74. Jim CY, Chen WY. Habitat effect on vegetation ecology and occurrence on urban masonry walls. Urban Forestry & Urban Greening. 2010 Jan 1;9(3):169–78.
  75. 75. Cilliers SS, Van Wyk E, Bredenkamp GJ. Urban nature conservation: vegetation of natural areas in the Potchefstroom municipal area, North West Province, South Africa. Koedoe. 1999 Jul 31;42(1):1–30.
  76. 76. Cilliers SS, Müller N, Drewes E. Overview on urban nature conservation: situation in the western-grassland biome of South Africa. Urban Forestry & Urban Greening. 2004 Jul 13;3(1):49–62.
  77. 77. Nassauer JI. Messy ecosystems, orderly frames. Landscape journal. 1995 Sep 21;14(2):161–70.
  78. 78. Jim CY. Sustainable urban greening strategies for compact cities in developing and developed economies. Urban Ecosystems. 2013 Dec 1;16(4):741–61.
  79. 79. Threlfall CG, Mata L, Mackie JA, Hahs AK, Stork NE, Williams NS, Livesley SJ. Increasing biodiversity in urban green spaces through simple vegetation interventions. Journal of Applied Ecology. 2017 Dec;54(6):1874–83.
  80. 80. Hahs AK, McDonnell MJ, McCarthy MA, Vesk PA, Corlett RT, Norton BA, Clemants SE, Duncan RP, Thompson K, Schwartz MW, Williams NS. A global synthesis of plant extinction rates in urban areas. Ecology Letters. 2009 Nov;12(11):1165–73. pmid:19723284
  81. 81. Talhouk SN, Dardas M, Dagher M, Clubbe C, Jury S, Zurayk R, Maunder M. Patterns of floristic diversity in semi-natural coastal vegetation of Lebanon and implications for conservation. Biodiversity & Conservation. 2005 Apr 1;14(4):903–15.
  82. 82. García‐Mora MR, Gallego‐Fernández JB, García‐Novo F. Plant functional types in coastal foredunes in relation to environmental stress and disturbance. Journal of Vegetation Science. 1999 Feb;10(1):27–34.
  83. 83. Box EO. Plant functional types and climate at the global scale. Journal of Vegetation Science. 1996 Jun;7(3):309–20.
  84. 84. Díaz S, Briske D, McIntyre S, Grice A, Hodgkinson K. Range management and plant functional types. AC Grice and KC Hodgkinson. Global rangelands: Progress and prospects. CABI Publishing. Wallingford, UK: CAB International. 2002 Jul 29:81–100.
  85. 85. Doing H. Coastal fore-dune zonation and succession in various parts of the world. InEcology of coastal vegetation 1985 (pp. 65–75). Springer, Dordrecht.
  86. 86. Cornelissen JH, Lavorel S, Garnier E, Diaz S, Buchmann N, Gurvich DE, Reich PB, Ter Steege H, Morgan HD, Van Der Heijden MG, Pausas JG. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian Journal of Botany. 2003;51(4):335–80.
  87. 87. Nelis LC. Life form and life history explain variation in population processes in a grassland community invaded by exotic plants and mammals. PloS one. 2012 Aug 20;7(8):e42906. pmid:22916178
  88. 88. Beninde J, Veith M, Hochkirch A. Biodiversity in cities needs space: a meta‐analysis of factors determining intra‐urban biodiversity variation. Ecology Letters. 2015 Jun;18(6):581–92. pmid:25865805
  89. 89. Tohmé G, Tohmé H. Illustrated flora of Lebanon. National Council for Scientific Research. 2014.
  90. 90. Gowler ZR. A taxonomic revision of the genus Matthiola R. Br.(Cruciferae) and related genera (Doctoral dissertation, University of Edinburgh). 1998.
  91. 91. Rteil HH. Evaluating methods for assigning conservation status of Lebanese plant species (Masters thesis); 2002.
  92. 92. Itani MA. Physiognomy as a basis for plant species conservation in urban areas: Beirut as a case-study (Master’s thesis). 2015
  93. 93. Mouterde R. Regards sur Beyrouth phénicienne: hellénistique et romaine. Impr. catholique; 1966
  94. 94. Hall LJ. Roman Berytus: Beirut in late antiquity. Routledge; 2004 Jun 1.
  95. 95. Population and Development Strategies Programme, 2014. Demographic Characteristics of Residents. http://www.pdslebanon.org/UserFiles/Chapter1_DemographicCharacteristicsOfResidents.pdf
  96. 96. MOE/LEDO/ECODIT. Lebanon State of the Environment Report, 2001. http://www.unep.org/dewa/westasia/Assessments/national_SOEs/west%20asia/Lebanon/Chap1Population.pdf
  97. 97. Kassir S. Histoire de Beyrouth. Fayard; 2003.
  98. 98. Mouterde P. Nouvelle flore du Liban et de la Syrie. Editions de l'Impr. catholique; 1966.
  99. 99. Dubertret L. Sur la structure de la plateforme de Beyrouth et sur ses gres Quaternaires. Compte-Rendu de la Société géologique de France. 1940;8:83–4.
  100. 100. Chmaitelly H, Talhouk S, Makhzoumi J. Landscape approach to the conservation of floral diversity in Mediterranean urban coastal landscapes: Beirut seafront. International Journal of Environmental Studies. 2009 Apr 1;66(2):167–77.
  101. 101. Yassin N. Beirut. Cities. 2012 Feb 1;29(1):64–73.
  102. 102. Boissier E. Flora Orientalis: sive enumeratio plantarum in Oriente a Graecia et Aegypto ad Indiae fines hucusque observatorum. apud H. Georg [etc.]; 1867.
  103. 103. Mouterde P. Petite flore des environs de Beyrouth. 1935.
  104. 104. Post GE. Flora of Syria, Palestine, and Sinai. Рипол Классик; 1932.
  105. 105. Diekmann M, Kühne A, Isermann M. Randomvs non-random sampling: Effects on patterns of species abundance, species richness and vegetation-environment relationships. Folia Geobotanica. 2007 Jun 1;42(2):179.
  106. 106. Dinsdale JM. The conservation and ecology of the heath lobelia, Lobelia urens L. 1996. (Doctoral dissertation).
  107. 107. Goldsmith FB. An assessment of the Fosberg and Ellenberg methods of classifying vegetation for conservation purposes. Biological Conservation. 1974 Jan 1;6(1):3–6.
  108. 108. Causton DR. An Introduction to Vegetation Analysis: Principles. Practice and Interpretation. Unwin Hyman, London. 1988.
  109. 109. Kent M, Coker P. Vegetation Description and Analysis: A Practical Approach. 1992.
  110. 110. Kennedy KA, Addison PA. Some considerations for the use of visual estimates of plant cover in biomonitoring. The Journal of Ecology. 1987 Mar 1:151–7.
  111. 111. Bergstedt J, Westerberg L, Milberg P. In the eye of the beholder: bias and stochastic variation in cover estimates. Plant Ecology. 2009 Oct 1;204(2):271.
  112. 112. Hill MO. TWINSPAN. A FORTRAN program for arranging multivariate data in an ordered two-way table by classification of the individuals and their attributes. Ecology and Systematics, Cornell University: Ithaca, New York. 1979.
  113. 113. Legendre P, Legendre LF. Numerical ecology (Vol. 24) Elsevier. 2012.
  114. 114. Catalano C, Marcenò C, Laudicina VA, Guarino R. Thirty years unmanaged green roofs: Ecological research and design implications. Landscape and Urban Planning. 2016 May 1;149:11–9.
  115. 115. Grimm NB, Faeth SH, Golubiewski NE, Redman CL, Wu J, Bai X, Briggs JM. Global change and the ecology of cities. Science. 2008 Feb 8;319(5864):756–60. pmid:18258902
  116. 116. Meiners SJ. Functional correlates of allelopathic potential in a successional plant community. Plant Ecology. 2014 Jun 1;215(6):661–72.
  117. 117. Vestergaard P. Temporal development of vegetation and geomorphology in a man‐made beach‐dune system by natural processes. Nordic Journal of Botany. 2004 Jul;24(3):309–26.
  118. 118. Miura N.The utility of LiDAR for landscape biodiversity assessment, Doctor of Philosophy (PhD), Mathematical and Geospatial Sciences, RMIT University. 2010.
  119. 119. Caynes RJ, Mitchell MG, Wu DS, Johansen K, Rhodes JR. Using high-resolution LiDAR data to quantify the three-dimensional structure of vegetation in urban green space. Urban Ecosystems. 2016 Dec 1;19(4):1749–65.
  120. 120. Wilkes P, Disney M, Vicari MB, Calders K, Burt A. Estimating urban above ground biomass with multi-scale LiDAR. Carbon balance and management. 2018 Dec;13(1):10. pmid:29943069
  121. 121. Snavely RA, Uyeda KA, Stow DA, O’Leary JF, Lambert J. Mapping vegetation community types in a highly disturbed landscape: integrating hierarchical object-based image analysis with lidar-derived canopy height data. International journal of remote sensing. 2019 Jun 3;40(11):4384–400.
  122. 122. Shwartz A, Turbé A, Julliard R, Simon L, Prévot AC. Outstanding challenges for urban conservation research and action. Global Environmental Change. 2014 Sep 1;28:39–49.
  123. 123. Misgav A. Visual preference of the public for vegetation groups in Israel. Landscape and Urban Planning. 2000 May 1;48(3–4):143–59.
  124. 124. Blackwood J, Hastings A, Costello C. Cost-effective management of invasive species using linear-quadratic control. Ecological Economics. 2010 Jan 15;69(3):519–27.
  125. 125. Huenneke LF, Thomson JK. Potential interference between a threatened endemic thistle and an invasive nonnative plant. Conservation Biology. 1995 Apr;9(2):416–25.
  126. 126. McDonnell MJ, Pickett ST. Ecosystem structure and function along urban-rural gradients: an unexploited opportunity for ecology. Ecology 71: 1232–1237. 1990.
  127. 127. Pickett ST, Burch WR, Dalton SE, Foresman TW, Grove JM, Rowntree R. A conceptual framework for the study of human ecosystems in urban areas. Urban Ecosystems. 1997 Dec 1;1(4):185–99.
  128. 128. Alberti M, Botsford E, Cohen A. Quantifying the urban gradient: linking urban planning and ecology. In Avian ecology and conservation in an urbanizing world 2001 (pp. 89–115). Springer, Boston, MA.
  129. 129. Marzluff JM, Bowman R, Donnelly R. A historical perspective on urban bird research: trends, terms, and approaches. InAvian ecology and conservation in an urbanizing world 2001 (pp. 1–17). Springer, Boston, MA.