SEARCH

SEARCH BY CITATION

Keywords:

  • alien plants;
  • Bayesian analysis;
  • dispersal;
  • Ellenberg numbers;
  • extinction;
  • growth form;
  • plant traits;
  • soil pH;
  • SLA;
  • urban ecology

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Data and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    As urbanization accelerates, interest is growing in the traits of urban plants. We classified 822 UK vascular plants on the basis of their occurrence along an urban–rural gradient in 26 710 samples of vegetation from 2508 UK 1-km grid squares, including a wide range of rural habitats and Sheffield and Birmingham, two of the UK's largest cities.
  • 2
    Both alien and native species were classified with respect to mean proportion of urban land cover in the 1-km grid squares in which the species occurs (urbanity), and absolute frequency in highly urban grid squares (urban frequency). Bayesian regression models were then developed for both measures, with a wide range of plant traits as explanatory variables.
  • 3
    Results for aliens and natives and for both urban measures were remarkably similar: the single, coherent picture of ‘successful urban species’ that emerges from our analysis is of robust plants of relatively fertile, dry, unshaded, base-rich habitats. Only seed mass behaved very differently for natives and aliens; seed mass was related positively to success of urban natives, and negatively to success of urban aliens. Neither clonality, seed dispersal nor seed persistence in the soil were strongly linked to success in urban habitats.
  • 4
    Synthesis. Cities can provide opportunities for surprisingly rich floras, but the traits of species that can persist in cities are quite narrowly circumscribed. More generally, it is clear that analysis of traits reveals important patterns in floristic data that would be far from obvious from a purely floristic analysis.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Data and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Urbanization is a major cause of plant species extinctions, and one that is likely to accelerate in coming decades (Duncan & Young 2000; Preston 2000; Williams et al. 2005). Much of the predicted global growth of the human population is likely to be accommodated in cities, leading to an increase in the total area covered by urban and other built-up areas (Varis 2006) and the intensification of land use within existing urban boundaries. Moreover, land modification by urbanization is much more permanent than that by agriculture, with little or no opportunity for recovery following abandonment. Despite their intrinsic interest, urban floras are often regarded as having little or no conservation value. But as cities grow to occupy larger and larger areas, and more and more of the world's population lives in them, interest is growing in the kinds of plants and floras likely to survive in these areas, from the perspective of conservation, evolution and the ecosystem services provided by urban vegetation (Pickett & Cadenasso 2008). Urban floras also have social implications since for an increasing proportion of humans, the biota that survive in cities will be their only contact with nature.

During the process of urbanization, existing habitats are destroyed or profoundly modified, and many species are lost. In addition, new anthropogenic habitats are created, providing habitats for new species and for novel plant communities (Godefroid & Koedam 2007). Some have suggested that these novel communities are similar in cities throughout the world (McKinney 2006; Schwartz et al. 2006), with increasing interest in discovering the functional traits that are both favoured and disadvantaged by urbanization, particularly in European cities (e.g. Landolt 2000; Chocholouskova & Pysek 2003; Kuhn et al. 2004; Kuhn & Klotz 2006; Godefroid & Koedam 2007).

Here we ask whether urbanization selects for particular plant traits in the British flora, by examining the composition of samples of vegetation along an urban–rural gradient (Hill et al. 2002; McDonnell et al. 1993). We compare the traits of species defined by their locations on two different versions of this gradient: mean proportion of urban land cover in the 1-km grid squares in which the species occurs (urbanity), and absolute frequency in highly urban grid squares (urban frequency). Note that urbanity essentially measures the degree to which species are confined to urban habitats, since species with high values cannot be abundant outside urban areas, while urban frequency is a measure of how common species are in urban habitats; species with high values of frequency may be equally common outside towns.

Data and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Data and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

quadrat and trait data

The quadrat data set is described fully in Hill et al. (2002). Briefly, we combined five existing data sets, three of them almost wholly rural, one largely urban and one with elements of both (see Table 2 in Hill et al. 2002). The urban data were from two of England's largest cities, Sheffield in Yorkshire and Birmingham in the West Midlands. The combined data set included 26 710 samples of vegetation from 2508 UK 1-km grid squares, of which 2595 samples (10%) were from ‘highly urban’ grid squares, defined as those with > 40% urban land cover (Fuller et al. 1994; Bunce et al. 1996). Different surveys used quadrat sizes varying from 1 × 1 m to 14 × 14 m, but since the data were used only to locate species on an urban–rural gradient, these differences are immaterial. The overwhelming majority of the floristic records are for spontaneous vegetation, but in a fundamentally anthropogenic landscape such as the UK, crop plants, plantation trees and ornamentals (e.g. in urban derelict land) were inevitably also recorded. All quadrat data were originally quantitative, but were transformed to presence and absence before analysis.

Trait data were available for Ellenberg L (light), F (moisture) and R (soil pH), life-form, plant height, clonality, seed mass, specific leaf area (SLA), soil seed bank, seed dispersal (by wind, water, on fur or feathers, inside grazing animals or by birds (fleshy fruits)). Note that strictly speaking, Ellenberg numbers are not actual plant traits, but habitat associations (Ellenberg et al. 1992). Data for Ellenberg numbers, life-form, plant height and clonality were obtained from plantatt (Hill et al. 2004), although clonality categories were simplified and amalgamated for this analysis. Soil seed bank data were obtained from Thompson et al. (1997), but for this analysis were simplified to a binary classification (seeds persistent for more than 1 year or not). Seed mass and SLA were obtained from the LEDA data base (Knevel et al. 2003). Seed dispersal data were obtained from W. Ozinga (unpubl. data) and subsequently modified by reference to seed morphology and plant habitat, height and growth form. Dispersal was a binary classification, that is, species were classed as dispersed by a particular vector or not, and classes were not mutually exclusive.

statistical model

Analyses were conducted only on species found in 10 or more quadrats. Linear models were developed for the measures of urbanity and urban frequency, with plant traits as explanatory variables. Plant height (cm), SLA (mmmg−1), Ellenberg numbers and seed mass (mg) were continuous variables. Quadratic and linear terms were included for these to accommodate the possible homogenizing influence of urbanization, with variables centred to improve the efficiency of parameter estimation. Life-form, specialist dispersal mechanisms, clonality and the presence of a seed bank were categorical variables. Based on assessments of model fit, plant height and seed mass were log-transformed, and SLA was square-root-transformed. Similarly, the response variables were log-transformed to improve model fit (log(x + 0.1) for urbanity and log(x + 1) for urban frequency).

Data for some plant traits were not available for some species (seed mass, SLA and seed bank). These cases were accommodated by estimating the missing values using a hierarchical model based on information from other species within families. Submerged and floating water plants, for which plant height is not a meaningful concept, were assigned an arbitrary height of 1 cm. The probability of a plant species having a persistent seed bank was assumed to vary among families. The transformed continuous variables (seed mass, SLA) were assumed to have normal distributions for each family, the mean and variance of which were drawn from common distributions for all families. This hierarchical structure meant that missing values for species in families with few data were drawn from the range of values for other species irrespective of family. In contrast, missing values for species in data-rich families were drawn from values that were representative of that family. These models can be written as:

  • Yi~ norm(mi, s)
  • image
  • image
  • image
  • image
  • υ j ~ normM, σM)
  • ψ i ~ normSLA, σSLA)
  • ϕi ~ invgamma(αM, βM)
  • ω i ~ invgammaSLA, βSLA)

where Yi is the measured response variable for species i for the particular model, s is the residual variance, mi is the predicted response, a is the intercept term, b1 to b12 are the regression coefficients for the continuous variables, lj is the effect of having life-form j, LFi is the life-form of species i, dk is the effect of having dispersal mode k, Di is the dispersal mode of species i, p is the effect of having a persistent seed bank, and Si = 1 when species i has a seed bank and zero otherwise, cf is the effect of having clonality type f, and Ci is the clonality type of species i. The hierarchical component of the model has parameters (µSLA, σSLA, α SLA, βSLA) that describe how the mean (ψj) and variance (ωj) of species’ SLA (within families) varies among families, with equivalent parameters for seed mass (µM, σM, αM, βM). The parameter Fami identifies which family species i belongs to, and γj is the probability of a species in family j having a persistent seed bank.

We fitted the full model to the data and did not attempt to select simpler models (using information criteria, for example) because we were interested in measuring the effect size for each parameter, and wanted to compare aliens with natives. Therefore, a common statistical model was desirable. Effect sizes were calculated as the largest predicted change in the response variable across the range of the explanatory variable.

Because of the ease of analysing complicated hierarchical models in a Bayesian framework (Clark 2005), the parameter estimates were obtained using WinBUGS (Spiegelhalter et al. 2007). Flat priors were assigned to the model parameters to ensure that the data had by far the greatest influence on the posterior distribution. These priors were normally-distributed with mean zero and SD of 1000 for the fixed effects parameters and the hyper-parameter means (a and bi, and µM and µSLA), uniform between 0 and 10 for the SD hyper-parameters (σM, σSLA), gamma-distributed with mean 1 and variance 1000 for the parameters of the inverse gamma-distributions (αM, βM, αSLA, βSLA), and uniform between 0 and 1 for γi, the probability of family i having a seed bank. Inspection of the samples and insensitivity to initial values indicated that 90 000 Monte Carlo samples adequately described the posterior distribution of the parameters after an initial burn-in of 10 000 samples.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Data and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We had data on 151 alien species from 42 families, and 671 native species from 107 families. Data were available for 100% of species for seed dispersal mechanism (except for two sterile hybrids), Ellenberg numbers, height, clonality and life-form, 86% for seed bank, 96% for seed mass and 81% for SLA.

The regression coefficients of the statistical models (Table 1) describe the relative influence of the different traits. The R2 values of the models were 0.58 and 0.31 for urbanity and frequency of natives, and 0.66 and 0.38 for aliens. The effect size for each trait was summarized by the largest predicted change in the response variable across the range of the data, expressed on a log-scale (Fig. 1). Therefore, a value of one for the effect size means that the response variable changes by a factor of 10 across the range of the data. Height, Ellenberg R (soil pH), life-form, seed mass, SLA, dispersal mode, Ellenberg L (light), and Ellenberg F (moisture) were relatively important predictors in at least one of the statistical models. Clonality and the presence of a persistent seed bank were relatively unimportant predictors in all the statistical models.

Table 1.  Mean (± SD) of the posterior distributions of the regression coefficients for the four different models of urbanity and urban frequency, with plant traits as explanatory variables
ParameterCoefficientFrequency nativeFrequency alienUrbanity nativeUrbanity alien
Intercepta0.46 (0.13)0.44 (0.34)0.52 (0.078)1.0 (0.16)
Height (linear)b10.19 (0.070)0.41 (0.15)0.19 (0.042)0.19 (0.073)
Height (quadratic)b20.16 (0.056)0.035 (0.13)0.054 (0.034)0.050 (0.063)
Seed mass (linear)b30.12 (0.041)–0.27 (0.094)0.074 (0.023)–0.155 (0.045)
Seed mass (quadratic)b4–0.011 (0.019)0.011 (0.037)–0.0025 (0.011)0.0074 (0.019)
SLA (square root)b50.062 (0.031)0.024 (0.086)0.10 (0.019)–0.013 (0.045)
SLA (linear)b6–0.049 (0.017)–0.081 (0.053)–0.070 (0.0089)–0.048 (0.020)
L (linear)b70.20 (0.041)0.0024 (0.12)0.070 (0.024)0.0036 (0.060)
L (quadratic)b8–0.067 (0.013)0.012 (0.031)–0.022 (0.0077)0.020 (0.015)
F (linear)b9–0.023 (0.026)–0.039 (0.064)–0.072 (0.015)–0.048 (0.030)
F (quadratic)b10–0.013 (0.0079)0.0061 (0.024)0.014 (0.0046)–0.0043 (0.012)
R (linear)b110.041 (0.019)0.15 (0.094)0.11 (0.012)0.18 (0.045)
R (quadratic)b12–0.038 (0.0089)–0.018 (0.032)–0.022 (0.0054)–0.043 (0.016)
Clonality (short)c20.047 (0.074)–0.086 (0.28)–0.041 (0.044)–0.045 (0.13)
Clonality (long)c30.24 (0.066)–0.042 (0.18)–0.014 (0.040)–0.023 (0.085)
Dispersal (fur/feather)d10.10 (0.074)0.16 (0.14)–0.0025 (0.044)0.0060 (0.064)
Dispersal (fruit)d2–0.17 (0.13)0.0896 (0.31)–0.018 (0.0798)0.043 (0.14)
Dispersal (grazing)d30.29 (0.056)0.11 (0.12)0.083 (0.033)–0.11 (0.056)
Dispersal (water)d4–0.14 (0.098)–0.53 (0.44)–0.034 (0.058)–0.25 (0.22)
Dispersal (wind)d50.28 (0.086)0.19 (0.18)0.17 (0.051)0.011 (0.086)
Life-form (geophyte)l20.29 (0.15)0.11 (0.37)0.24 (0.089)0.15 (0.17)
Life-form (hemicryptophyte)l30.29 (0.11)0.19 (0.27)0.19 (0.07)0.028 (0.12)
Life-form (hydrophyte)l40.61 (0.19)1.4 (0.82)0.55 (0.11)0.99 (0.38)
Life-form (liana)l50.41 (0.23)0.11 (0.14)
Life-form (phanerophyte)l6–0.038 (0.19)–0.19 (0.41)0.13 (0.11)–0.090 (0.19)
Life-form (therophyte)l70.34 (0.13)0.35 (0.28)0.31 (0.079)–0.0081 (0.13)
Life-form (parasitic)l80.31 (0.26)0.22 (0.15)
Seed bankp0.16 (0.059)0.043 (0.16)0.12 (0.036)0.044 (0.081)
image

Figure 1. Effect sizes of the independent variables on alien and native urbanity and frequency, expressed as the largest predicted change in the response variable across the range of the data, on a log10-scale, so a value of 1 represents a predicted 10-fold change across the range of the data. All effect sizes are shown as positive for ease of comparison. Error bars represent one SE each side of the mean (circles).

Download figure to PowerPoint

The propensity to be confined to urban areas (urbanity) and frequency in urban areas were positively related to height for both alien and native species (Fig. 2). Urban frequency changed by a factor of 15 for native plants across the range of heights, and 30 times for alien plants. This ratio was approximately six for urbanity for both native and alien plants.

image

Figure 2. Examples of response of alien and native urbanity and frequency to height, SLA and seed mass. Natives on left, aliens on the right. Solid lines are mean of fitted response, and dashed lines are 95% credible intervals.

Download figure to PowerPoint

Generally, urban areas appeared to favour plants of base-rich soils, although the model for urban frequency of natives indicated highest frequency at intermediate pH values. Urbanity varied approximately ninefold across the range of the data for native species and 40-fold for alien species. Frequency varied approximately sixfold across the range of the data for native species and 10-fold for alien species.

Life-form had a relatively large effect for the alien species (effect sizes of approximately 14 and 60 for urbanity and urban frequency), with smaller effects for native species (effect sizes of approximately four for urbanity and six for frequency). In all four cases, hydrophytes appeared to be more favoured by urbanization, with relatively small differences among the other life-forms. The uncertainty for the hydrophyte effect was large for alien species because there are very few alien hydrophytes in the UK flora.

There was a negative relationship with seed mass for urban frequency and urbanity of aliens (Fig. 2; effect sizes of 28 and 7). The effect sizes for urbanity and frequency of natives were relatively small and positive (three and five). Intermediate SLA was consistently favoured by urbanization (Fig. 2). The effect sizes were nine and four for urbanity, and 4 and 10 for frequency of natives and aliens, respectively.

Species with high moisture requirements tended to be adversely affected by urbanization in all four models, with effect sizes for Ellenberg F being two and three for urbanity for natives and aliens, and six and three for urban frequency. There was also a tendency for a positive relationship between light requirements (Ellenberg L) and response to urbanization for all four models, but the effect sizes were relatively small (two for urbanity and six for frequency of natives, and three for frequency and urbanity of aliens).

The presence of specialized dispersal mechanisms had a relatively small effect on the response variables (effect sizes ≤ 4, except for 18 for alien urban frequency). In all cases, species dispersed by water were most adversely affected by urbanization, with little difference between the other dispersal mechanisms. There was little effect of clonality (effect sizes ≤ 2) in all four models, and no consistent direction of the relationship. There was a positive effect of seed bank on the response to urbanization in all four models but the effect sizes were small (≤ 1.5).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Data and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

These results provide striking evidence that some traits are consistently favoured by urban environments in Britain. Urbanization strongly favours robust plants of base-rich habitats, and to a lesser extent those of dry, unshaded and at least moderately fertile habitats. To a large extent, these results were similar for aliens and natives and for different definitions of ‘urban’; only seed mass behaved very differently for natives and aliens.

At first sight, the strong tendency for success in urban habitats to be linked positively to plant height is surprising, given that cities are generally regarded as rather disturbed places, which might be expected to favour ruderality and/or the annual life-history (i.e. relatively small plants). However, to some extent this view may derive from the origins of urban ecology in the bombed cities of post-WWII Europe, and modern cities may be less disturbed than is sometimes thought; in the Czech Republic, for example, weedy vegetation of towns and cities was highly enriched in perennials and biennials relative to more-disturbed arable fields (Lososova et al. 2006). Trends in abundance of different growth forms may be far from simple; in Plzeň in the Czech Republic, the city has more annuals than its immediate surroundings, but at the same time the number of woody species has increased by 280% since the 19th century, largely due to escape of woody neophytes from cultivation (Chocholouskova & Pysek 2003). Similar increases in woody species have been observed in Brussels (Godefroid & Koedam 2007) and Italian cities (Celesti-Grapow & Blasi 1998). At the same time, species with the lowest levels of both urbanity and urban frequency tend to be short species of extreme (e.g. cold, infertile) and rather open habitats (e.g. Drosera intermedia, Saxifraga stellaris, Thalictrum alpinum, Huperzia selago, Carex dioica, Alchemilla alpina, Carex ornithopoda, Saxifraga aizoides). In Middlesex in southern England, Preston (2000) found that locally extinct species were predominantly short plants of open, infertile habitats. Not surprisingly, such species are almost exclusively natives. Short – and particularly very short (< 30 cm) – species were more likely to have suffered local extinction during European settlement of Auckland, New Zealand (Duncan & Young 2000).

A positive relationship with Ellenberg R (high soil pH) has an obvious source, in that the most highly modified parts of cities are dominated by man-made building materials, which tend to be strongly alkaline (Sukopp & Starfinger 1999). Not surprisingly therefore, previous studies have often found that urban areas favour calcicolous species (e.g. Chocholouskova & Pysek 2003). However, cities may be far from uniform in this respect; studies that have looked in detail at different types of urban land use (e.g. Godefroid & Koedam 2007) have found that the relationship between urban plants and high pH is confined to the most built-up areas. Preston (2000) found that urban expansion in Middlesex had led to the local extinction of a disproportionate number of species at both extremes of the pH range, a pattern that has also been reported from Zurich (Landolt 2000). Therefore it is perhaps not surprising that for natives at least, frequency (rather than urbanity) peaks at intermediate values of Ellenberg R in this study.

Seed mass was negatively related, with relatively large effect sizes, to both frequency and urbanity of aliens. It is tempting to speculate that in the disturbed urban environment, this relationship is driven more by seed production than by seed size per se. In other words, small-seeded species are more successful than large-seeded ones because they are more fecund. In one of the few examples in this study of inconsistent behaviour of aliens and natives, the effect size for both frequency and urbanity of natives was positive, although only small. To some extent this result seems to be related to that for plant height; the least urban species are small plants of short, open vegetation, often with correspondingly small seeds. Thus we surmise that the negative effect of seed size in natives is largely an artefact of the very poor performance of small (and small-seeded) plants. This effect is not apparent in aliens because this latter type of species is essentially absent from the alien flora; of the one hundred smallest-seeded species in the data base analysed here, only eleven are alien, a far lower proportion than in the whole data set.

Many previous studies have reported that urban species are generally plants of fertile habitats (Landolt 2000; Preston 2000; Chocholouskova & Pysek 2003; Van der Veken et al. 2004; Lososova et al. 2006), but in this respect, as in many others, cities are highly heterogeneous (Godefroid & Koedam 2007). As Sukopp & Starfinger (1999) point out, anthropogenic activity can profoundly modify soils; at one extreme, deeply cultivated soils of gardens or former gardens may be highly fertile, while other urban soils may be highly compacted and/or mixed with rubble or subsoil. These latter soils may be high in phosphorus and bases, but are usually low in nitrogen and may also be very dry. SLA is positively correlated with growth rate, nutrient content and other indicators of fertility (Grime et al. 1997), and in this study the most urban species were characterised by intermediate levels of SLA, suggesting that urban habitats are fertile, but not extremely so. It is possible that some other studies have failed to note this unimodal relationship by not including quadratic terms in their analyses, essentially assuming that all relationships were linear.

Our data are consistent with most other studies that have concluded that cities are generally dry and unshaded (Chocholouskova & Pysek 2003), but in our study both effects were relatively weak. One reason for these weak effects may be that only the most built-up areas consistently lack trees and consist largely of coarse substrates and well-drained, sealed surfaces. In other, less-developed urban areas, there is likely to be both more shade and more water, with correspondingly different effects on plant traits (Godefroid & Koedam 2007). Given the generally negative impact of urbanization on plants of moist habitats, it is at first sight surprising that the only large effect of urbanization on life-form is a positive one for hydrophytes, particularly aliens. The reason for this apparently anomalous result is quickly revealed by inspection of the data. There are only three alien hydrophytes in the data set, two of which (Elodea canadensis and E. nuttallii) either are or have been highly invasive, particularly in disturbed eutrophic water bodies. Both frequently escape from aquaria or garden ponds, are strongly favoured by eutrophication, and are moderately to highly urban in their distribution. In contrast, urbanization is disastrous for many native water plants, especially those preferring oligotrophic conditions (Preston et al. 2003).

With the single exception of a strong negative effect of dispersal by water on alien frequency, neither clonality, seed dispersal nor seed persistence in the soil were strongly linked to success in urban habitats, either positively or negatively. Given the importance of reproduction by seed in disturbed urban habitats, these latter results are at first sight surprising. The conclusion seems to be that as long as species are tolerant of the urban environment in other respects, success or failure of dispersal – at least in the traditional sense – is not an important determinant of success. One reason for this may be that non-standard dispersal pathways dependent on human activity are increasingly important in the modern world (Hodkinson & Thompson 1997; Von der Lippe & Kowarik 2007), and this is likely to be particularly true in densely-populated urban environments. For example Fallopia japonica and Petasites fragrans survive, and have even become invasive, in the urban environment despite not reproducing sexually in the UK.

Urban areas are far from homogeneous, with densely built-up areas favouring very different species and traits compared to urban areas with more natural vegetation, parks, gardens and plantations (Godefroid & Koedam 2007). Our approach implicitly includes this variability, with the relatively few species with very high values of urbanity (e.g. Buddleja davidii, Conyza canadensis, Lactuca serriola, Aster novi-belgii) known to be largely confined to the most highly-modified urban areas (Preston et al. 2002), while the much more numerous species with moderately high values of urbanity are more widely distributed. It is also clear that the species with the highest values of urbanity are quite different from those that are actually most frequent in urban areas. The former (e.g. the four species mentioned above) are exclusively alien, while the latter are uniformly native and abundant everywhere (e.g. Agrostis stolonifera, Holcus lanatus, Cirsium arvense, Plantago lanceolata, Lolium perenne, Taraxacum officinale, Rubus fruticosus, Trifolium repens). Indeed, if we rank the entire data set by urbanity and urban frequency, irrespective of native or alien status, the 100 highest values in the two lists have only 19 species in common. Given this variability, it is remarkable that a single, coherent picture of the ‘successful urban species’ emerges quite clearly from our analysis of plant functional traits: robust plants of relatively fertile, dry, unshaded, base-rich habitats. The similarity between trends for aliens and natives, and for urbanity and urban frequency, despite the very different species concerned, is a powerful argument for the utility of an approach to plant ecology based on traits rather than on floristics (Thompson et al. 1995; Fukami et al. 2005; Smart et al. 2006).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Data and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This paper is a product of an Urbanization and Plant Functional traits workshop hosted by the Australian Research Centre for Urban Ecology at the University of Melbourne and funded by the ARC-NZ Research Network for Vegetation Function. Wim Ozinga kindly provided access to an unpublished dispersal data base. We are grateful to Mark Hill and David Roy for assistance in assembling the original floristic data set.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Data and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Bunce, R.G.H., Barr, C.J., Gillespie, M.K. & Howard, D.C. (1996) The ITE Land Classification: Providing an environmental stratification of Great Britain. Environmental Monitoring and Assessment, 39, 3946.
  • Celesti-Grapow, L. & Blasi, C. (1998) A comparison of the urban flora of different phytoclimatic regions in Italy. Global Ecology and Biogeography, 7, 367378.
  • Chocholouskova, Z. & Pysek, P. (2003) Changes in composition and structure of urban flora over 120 years: a case study of the city of Plzen. Flora, 198, 366376.
  • Clark, J.S. (2005) Why environmental scientists are becoming Bayesians. Ecology Letters, 8, 214.
  • Duncan, R.P. & Young, J.R. (2000) Determinants of plant extinction and rarity 145 years after European settlement of Auckland, New Zealand. Ecology, 81, 30483061.
  • Ellenberg, H., Weber, H.E., Düll, R., Wirth, V., Werner, W. & Paulißen, D. (1992) Zeigwerte von Pflanzen in Mitteleuropa Erich Goltze, Göttingen.
  • Fukami, T., Bezemer, T.M., Mortimer, S.R. & Van Der Putten, W.H. (2005) Species divergence and trait convergence in experimental plant community assembly. Ecology Letters, 8, 12831290.
  • Fuller, R.M., Groom, G.B. & Jones, A.R. (1994) The land cover map of Great-Britain – an automated classification of Landsat Thematic Mapper data. Photogrammetric Engineering and Remote Sensing, 60, 553562.
  • Godefroid, S. & Koedam, N. (2007) Urban plant species patterns are highly driven by density and function of built-up areas. Landscape Ecology, 22, 12271239.
  • Grime, J.P., Thompson, K., Hunt, R., Hodgson, J.G., Cornelissen, J.H.C., Rorison, I.H. et al . (1997) Integrated screening validates primary axes of specialisation in plants. Oikos, 79, 259281.
  • Hill, M.O., Preston, C.D. & Roy, D.B. (2004) PLANTATT. Attributes of British and Irish Plants: Status, Size, Life History, Geography and Habitats CEH, Huntingdon.
  • Hill, M.O., Roy, D.B. & Thompson, K. (2002) Hemeroby, urbanity and ruderality: bioindicators of disturbance and human impact. Journal of Applied Ecology, 39, 708720.
  • Hodkinson, D.J. & Thompson, K. (1997) Plant dispersal: the role of man. Journal of Applied Ecology, 34, 14841496.
  • Knevel, I.C., Bekker, R.M., Bakker, J.P. & Kleyer, M. (2003) Life-history traits of the Northwest European flora: The LEDA database. Journal of Vegetation Science, 14, 611614.
  • Kuhn, I. & Klotz, S. (2006) Urbanization and homogenization – Comparing the floras of urban and rural areas in Germany. Biological Conservation, 127, 292300.
  • Kuhn, I., Brandl, R. & Klotz, S. (2004) The flora of German cities is naturally species rich. Evolutionary Ecology Research, 6, 749764.
  • Landolt, E. (2000) Some results of a floristic inventory within the city of Zurich (1984–1998). Preslia, 72, 441455.
  • Lososova, Z., Chytry, M., Kuhn, I., Hajek, O., Horakova, V., Pysek, P. & Tichy, L. (2006) Patterns of plant traits in annual vegetation of man-made habitats in central Europe. Perspectives in Plant Ecology Evolution and Systematics, 8, 6981.
  • McDonnell, M.J., Pickett, S.T.A. & Pouyat, R.V. (1993). The application of the ecological gradient paradigm to the study of urban effects. Humans as Components of Ecosystems: Subtle Human Effects and the Ecology of Populated Areas (edS M.J. McDonnell & S.T.A. Pickett), pp. 175189. Springer-Verlag, New York.
  • McKinney, M.L. (2006) Urbanization as a major cause of biotic homogenization. Biological Conservation, 127, 247260.
  • Pickett, S.T.A. & Cadenasso, M.L. (2008) Linking ecological and built components of urban mosaics: an open cycle of ecological design. Journal of Ecology, 96, 812.
  • Preston, C.D. (2000) Engulfed by suburbia or destroyed by the plough: the ecology of extinction in Middlesex and Cambridgeshire. Watsonia, 23, 5981.
  • Preston, C.D., Pearman, D.A. & Dines, T.D. (2002) New Atlas of the British and Irish Flora Oxford University Press, Oxford.
  • Preston, C.D., Sheail, J., Armitage, P. & Davy-Bowker, J. (2003) The long-term impact of urbanisation on aquatic plants: Cambridge and the River Cam. Science of the Total Environment, 314, 6787.
  • Schwartz, M.W., Thorne, J.H. & Viers, J.H. (2006) Biotic homogenization of the California flora in urban and urbanizing regions. Biological Conservation, 127, 282291.
  • Smart, S.M., Thompson, K., Marrs, R.H., Le Duc, M.G., Maskell, L.C. & Firbank, L.G. (2006) Biotic homogenization and changes in species diversity across human-modified ecosystems. Proceedings of the Royal Society B – Biological Sciences, 273, 26592665.
  • Spiegelhalter, D., Thomas, A., Best, N. & Lunn, D. (2007) Open BUGS User Manual, Version 3.0.2. MRC Biostatistics Unit, Cambridge, UK.
  • Sukopp, H. & Starfinger, U. (1999). Disturbance in urban ecosystems. Ecosystems of Disturbed Ground (ed. L.R. Walker), Vol. 16, pp. 397412. Elsevier, Amsterdam.
  • Thompson, K., Bakker, J.P. & Bekker, R.M. (1997) The Soil Seed Banks of North West Europe: Methodology, Density and Longevity. Cambridge University Press, Cambridge.
  • Thompson, K., Hodgson, J.G. & Rich, T.C.G. (1995) Native and alien invasive plants: more of the same? Ecography, 18, 390402.
  • Van der Veken, S., Verheyen, K. & Hermy, M. (2004) Plant species loss in an urban area (Turnhout, Belgium) from 1880 to 1999 and its environmental determinants. Flora, 199, 516523.
  • Varis, O. (2006) Megacities, development and water. International Journal of Water Resources Development, 22, 199225.
  • Von der Lippe, M. & Kowarik, I. (2007) Long-distance dispersal of plants by vehicles as a driver of plant invasions. Conservation Biology, 21, 986996.
  • Williams, N.S.G., Morgan, J.W., McDonnell, M.J. & McCarthy, M.A. (2005) Plant traits and local extinctions in natural grasslands along an urban–rural gradient. Journal of Ecology, 93, 12031213.