Management mitigates the impact of urbanizationon meadow vegetation

Sirkku Manninen & Sonja Forss & Stephen Venn

# Springer Science+Business Media, LLC 2010

Abstract Urban regions often contain remnants of ecologically valuable habitats. Whilstmeadow habitats have been recognized as ecologically important and much studied, littleattention has been given to meadow assemblages of urban locations. We studied the effectsof meadow type, urbanization level, and management on vascular plant species richness,field layer diversity and soil chemistry in 18 grassland sites in the Helsinki MetropolitanArea (60°E, 25°N), on the southern coast of Finland during the summer of 2007. Werecorded a total of 252 species, though the average number of species per m2 was only 12.6.The negative effects of urbanization on forbs seemed to result in particular from increasedsoil nitrate (NO3

- -N) concentration. The highest NO3- -N and Fe concentrations and ratios

of total inorganic nitrogen (Ntot) to phosphorus (P) and potassium (K), were recorded fromthe soils of urban rocky meadows. Management by mowing decreased soil NO3

- -N and Feconcentrations, ratios of Ntot:P and Ntot:K, and increased species richness and diversity.Elevated NOx deposition is considered as a major driver of urbanization effects onvegetation, though changes in soil pH and metal concentrations, such as zinc (Zn), may alsonegatively affect the frequency of both forbs and grasses. This study shows that regularmanagement by mowing and removal of hay mitigates these effects. We also recommendincreasing the provision of dry meadows and maintaining more areas of supplementarysemi-natural grassland habitats in urban green space as concrete measures for theconservation of dry meadow assemblages and urban biodiversity.

Keywords Biodiversity . Heavymetals . Mowing . Nitrogen deposition . Semi-naturalgrassland . Urban ecology

Urban EcosystDOI 10.1007/s11252-010-0129-4

S. Manninen (*)Department of Environmental Sciences, University of Helsinki, Helsinki P.O. Box 56, 00014, Finlande-mail: sirkku.manninen@helsinki.fi

S. ForssFinnish Environment Institute, Natural Environment Centre/Biodiversity, P.O. Box 140, 00251 Helsinki,Finland

S. VennDepartment of Biological Sciences, University of Helsinki, Helsinki P.O. Box 65, 00014, Finland

Introduction

Semi-natural grasslands are among the habitats with the highest vascular plant speciesrichness in western, central and northern Europe (Willems 1982; Kull and Zobel 1991;Fischer and Stöcklin 1997; Rosén and Borgegård 1999). Many of the plant species typicalof such habitats are adapted to nutrient poor soil. Moreover, regular management keeps thevegetation low, as a result of which the habitat remains open and the high solar radiationalso warms and dries the soil quickly (Bakker 1989). A decline in the abundance of manycharacteristic plant species of such habitats has occurred generally across Europe mainlydue to the intensification of land-use practices and abandonment (Willems 1982; Erikssonet al. 1995; Fischer and Stöcklin 1997; Pykälä 2005).

In urban areas, further deterioration of such semi-natural habitats occurs due to thedeposition of air pollutants (Gilbert 1989). Whereas in the past industry and energyproduction have been the predominant sources of urban pollution, nowadays transport, inparticular road traffic, is the major source in European cities. The quality standards forurban background air will continue to be exceeded in larger conurbations in the foreseeablefuture. The most serious problems with regard to human health are peaks which exceedboth short- and long-term objectives for PM10 (particles ≤10 µm in diameter) and long-termobjective of 40 µg NO2 m−3 yr−1 for nitrogen oxides (NOx), i.e. nitric oxide (NO) andnitrogen dioxide (NO2), combined and expressed as NO2 (de Leeuw et al. 2001). However,the critical level of NOx for vegetation is even lower than these limits, at 30 µg NO2

m−3 yr−1 (UNECE 2004). In addition to direct negative effects of NOx on plants (Wellburn1990), increased nitrogen (N) deposition affects vegetation due to changes in soil pH andnutrient supply ratios (Roem and Berendse 2000; Duprè et al., 2010; Maskell et al. 2010).With regard to heavy-metal containing particle emissions from traffic, deposition of zinc(Zn) in particular seems to be increasing (Pearson et al. 2000). Changes in soil pH alsoinfluence the critical concentrations at which heavy metal toxicity occurs in both plants andsoil microorganisms (review by McGrath et al. 1995).

Remnants of semi-natural grasslands in urban areas also become excessively fragmentedand isolated (Gilbert 1989) or totally lost (Wittig et al. 2010). While biological significanceranking maybe an important explanatory variable determining patch fate, areas of higherecological value are often lost due to pressure for development (Williams et al. 2005). Inspite of or due to this, species assemblages associated with urban habitats may show bothhigh α diversity and β diversity (Niemelä 1999). Urbanization may also favour specieswith specific traits in terms of growth form, duration of life-history and nutrientrequirements. Tonteri and Haila (1990), for example, have shown that the urban vegetationin Helsinki consists predominantly of species groups that tolerate frequent disturbance, suchas annuals and grasses. Thompson and McCarthy (2008), in turn, classified 822 UKvascular plants, including both alien and native species, on the basis of their occurrencealong an urban-rural gradient and concluded that the archetype of a ‘successful urbanspecies’ that emerged from their analysis was that of robust plants of relatively fertile, dry,unshaded and alkaline habitats. The species least capable of persistance in urban areastended to be low growing species of extreme (e.g. cold, infertile) and rather open habitats,often with small seeds.

A large proportion of the threatened vascular plants found in Helsinki are associatedwith semi-natural habitats (Vähä-Piikkiö et al. 2004). In this study we investigate theecological effects of urbanization on vegetation in semi-natural grasslands in the HelsinkiMetropolitan Area, and assess whether management can ameliorate these effects.Experimental studies performed in rural meadows have shown that perennials, species

Urban Ecosyst

with N2-fixing symbionts, short-statured species, locally rare species and species thattolerate low nutrient availability, are particularly susceptible to loss via competetiveexclusion following eutrophication (Janssens et al. 1998; Suding et al. 2005; Clark andTilman 2008). Moreover, a trend towards grasses and other nitrophilic species has beenreported in the vicinity of roads (Angold 1997; Truscott et al. 2005; Bignal et al. 2007), andmetal-resistant genotypes are common in disturbed areas, especially for grass species (Ernst2004; and references therein). These observations, together with the results of Tonteri andHaila (1990), and Thompson and McCarthy (2008), led us to hypothesize that long-termincreased N deposition due to emissions of N compounds maybe one of the major driversfor a vegetational shift from forbs in favour of grasses in urban species-rich grasslandhabitats.

All the 18 urban meadows which we studied for field-layer vascular plant diversity andsoil chemistry in relation to meadow type, urbanization level and management regime in thesummer of 2007, represented Fennoscandian lowland species-rich dry to mesic grasslandsof high conservation value (Natura 2000 habitat type code 6270) (European EnvironmentAgency). Our second hypothesis was that regular management mitigates the negativeeffects of N deposition on plant diversity by removing N from the ecosystem when the hayis removed. We also hypothesized that management may mitigate the effects of particulatedeposition by reducing the levels of heavy metals reaching the soil, as elevatedconcentrations of heavy metals affect plant diversity (Bagatto and Shorthouse 1999; Vidicet al. 2006; Hernández and Pastor 2008).

Methods

Study area and sites

The study area lies in northwestern Europe, on the border of the hemiboreal and southernboreal vegetation zones (Ahti et al. 1968). The Helsinki Metropolitan Area consists of themunicipalities of Helsinki, Espoo, Vantaa and Kauniainen and covers an area of 769 km2,of which the City of Helsinki constitutes 213 km2 (Lankinen et al. 2009). The population ofthe metropolitan area was 0.96 million in 2000 and is predicted to reach 1.17 million by theend of 2025 (Helsinki Metropolitan Area Council 2006).

The 18 meadows we studied during May-August 2007 were located within 15 km of theHelsinki city centre (60°10′N, 24°57′E). These sites were selected from a larger number ofmeadows that were surveyed during 2005–2006 and found to contain forbs typical of thespecified semi-natural grassland type (H. Tuovila, personal communication), and for whichinformation on management regime was available. Six of the meadows had been mownregularly (i.e. once or twice each summer and the hay removed, except in Bemböle), nineirregularly (once or twice during the last 5 years, hay not removed always) and three wereunmanaged (Table 1). The meadows were located along an urban-rural gradient anddistributed as evenly as possible across the study area. This approach was adopted tominimize any bias that might arise due to geographical factors, such as a coastal gradient.Other criteria for inclusion were tree canopy cover of <30% and ownership by themunicipality or a public organization.

The surface areas of the studied meadows ranged from 0.06 ha to 1.12 ha (Table 1). Data fordetermining the surface areas and borders of each meadow and the urbanization level of theirsurroundings were obtained from SeutuCD’03 (Helsinki Metropolitan Area Council 2003)and Uusimaa CD (Uusimaa Regional Council and Helsinki Metropolitan Area Council

Urban Ecosyst

2002), and mapped using MapInfo Professional 8.0. In addition to population density(number of inhabitants per km2), we recorded total length of roads within a buffer zone of 1-km around each site and the distance of each meadow from the nearest trunk road, asindicators of urbanization level. Google Earth (using satellite images from the period 2002–2006) was used to estimate the total area of open habitats (tree canopy cover <30%) withinthe buffer zone around each site, as a number of the sites used were not covered by theSeutuCD’03 and Uusimaa CD resources.

The main gaseous air pollutants in the Helsinki Metropolitan Area are NO, NO2 and thesecondary pollutant ozone (O3). The annual mean NO concentrations ranged between 6 and43 µg m−3 and those of NO2 between 0.3 and 46 µg m−3 during 2000–2006. The majorsources of NOx emissions in 2006 (16,483 tons) were power plants (52%) and road traffic(34%), followed by ships (7%) and aeroplanes (4%) (Helsinki Metropolitan Area Council2007). Nitrogen deposition is not measured in the area, but the modelled NOx-N depositionin Helsinki is 0.2–0.3 g m−2 yr−1 (Syri et al. 2004), while the total N deposition is ca.0.5 g m−2 yr−1 in southern Finnish background areas (Kangas and Syri 2002). The annualmean O3 concentrations were 18–29 ppb, the highest 1-h means reaching 53–94 ppb andthe 24-h highest means 40–63 ppb. There is no data available for heavy metals (other thanlead), but the highest concentrations of particles (both PM10 and PM2.5) were measured atthe same locations as those for NOx, i.e. in areas with the heaviest traffic, while the highestO3 concentrations were recorded at the most rural location. Particle emissions from the useof fireplaces in domestic houses accounted for c. 29%, while the contributions of powerplants and road traffic were 33% and 28% respectively of the total particle emissions of 100tons in 2006 (Helsinki Metropolitan Area Council 2007).

Table 1 Type, management and area of the studied meadows, together with distance from nearest trunkroad, number of inhabitants, length of roads and total area of open habitats within 1-km buffer zone of eachmeadow. The meadows are listed from most urban to most rural in terms of population density

Site Type Management Area(ha)

Inhabitants(km−2)

Roads inbuffer (km)

Distance toroad (km)

Open habitat inbuffer (ha)

Reijolankallio rocky irregular 0.57 5,104 36.8 0.06 22.7

Ormusmäki rocky irregular 0.20 3,366 38.7 1.28 51.9

Tuorinniemi rocky irregular 0.23 3,329 31.2 0.97 48.4

Herttoniemi mesic regular 0.32 3,193 40.6 0.63 57.8

Pajamäki rocky unmanaged 1.12 2,786 32.1 0.34 51.6

Håkansberg rocky unmanaged 0.71 2,783 30.9 0.28 37.5

Vaskipelto mesic regular 0.42 2,692 30.0 0.99 41.9

Linnavuori dry irregular 0.12 2,167 32.9 0.22 125.4

Kurkimoisio dry irregular 0.13 2,138 27.8 0.39 88.8

Hilapelto mesic unmanaged 0.14 1,999 30.7 0.51 30.0

Uussilta dry regular 0.35 1,961 2.3 0.05 88.6

Jollas rocky regular 0.25 678 14.4 4.40 73.2

Bemböle dry irregular 0.18 211 18.5 0.70 136.9

Talosaari mesic regular 0.39 57 7.6 1.52 61.6

Tuupakka dry regular 0.08 52 21.4 0.78 146.5

Kunnarla dry irregular 0.06 36 15.4 3.98 82.3

Westerkulla mesic irregular 0.20 8 8.2 0.19 172.0

Porvarinlahti dry irregular 0.62 4 8.2 1.35 84.9

Urban Ecosyst

The average temperature in March–August 2007 was 11.7°C and the total rainfall333 mm, compared to 9.9°C and 295 mm respectively in the years 1971–2000. The 1.8°Chigher temperature was evident in the early onset of thermal spring (1st cf. 23rd March) andthe 13% higher rainfall as a doubled precipitation in May compared to the 30-year averages(Drebs et al. 2002; Finnish Meteorological Institute 2007).

Vegetation survey

Cover of vascular plant species in the field layer was estimated visually with a 1% precisionbetween 4th June and 3rd July 2007, using a 1 m×1 m quadrat that was divided into 100sub-squares. The number of quadrats varied from 5 to 15, depending linearly on the size ofthe meadow, and they were distributed so as to best represent microsite variation withineach meadow. These data were furthermore used to calculate the average cover of eachspecies per m2, as well as the Shannon-Wiener diversity index for each meadow.

The cover of the perennial forbs, Dianthus deltoides, Campanula rotundifolia, Galiumverum, Hypericum perforatum and Knautia arvensis, which have Ellenberg values of 2–4for N, were summed to indicate the relative abundance of nitrophobic species, and those ofAegopodium podagraria, Anthriscus sylvestris, Cirsium arvense, Lamium album and Urticadioica, with Ellenberg N values of 7–9 (Ellenberg et al. 2001), were summed to indicate therelative abundance of nitrophilic species. To estimate the total number of species for eachmeadow, species that were not found in the quadrats were also listed. Therefore we scoredthe study sites for the first time on 11th–15th May 2007, when we recorded all of the springflowering forbs in each meadow, and then we re-checked the sites on 5th–11th July 2007.The height of the vegetation was measured for each quadrat on 11th–13th July. Speciesnomenclature and ranking of species according to their functional type, typical biotope/habitat, origin and life history, followed that of Hämet-Ahti et al. (1998, 2005).

Soil chemistry

Soil samples were collected between 5th–11th July by taking seven subsamples from the soilsurface (≤10 cm thick/deep) of each meadow using a soil borer. We removed any litter, pooledthe subsamples and sent them to the commercial laboratory Viljavuuspalvelu Oy for analysis ofpHH2O, as well as nutrients and heavy metals. Concentrations of easily soluble (i.e. plantavailable) calcium (Ca), potassium (K), magnesium (Mg), phosphorus (P), sulphur (S),copper (Cu), zinc (Zn), manganese (Mn) and iron (Fe), were measured from acidicammonium acetate extracts (with and without EDTA), using an ICP emission spectrometer.Ammonium (NH4

+ -N) was extracted with 0.1 M K2SO4 and analysed using the Kjeldahlmethod. Nitrate (NO3

- -N) was analysed from a water-soil suspension with an ion-selectiveelectrode. The sum of NH4

+ -N and NO3- -N was used as an indicator of inorganic total

nitrogen (Ntot) concentration. The samples were dried and sieved prior to analysis.

Statistical analyses

Soil chemistry results were log transformed to achieve approximate normality prior to thestatistical analyses. The classification of meadows into rocky and non-rocky types was based onthe relative proportion of bare bedrock (≥30% and <30%, respectively) and the non-rocky classwas further divided into dry and mesic types, according to general features of vegetation andtopography (e.g. exposure) in each meadow and its surroundings. Accordingly, seven of themeadows were classified as dry, six as rocky, and five as mesic meadows (Table 1).

Urban Ecosyst

Relationships between vegetation, soil and other variables (meadow type, urbanizationindicators, management regime) were analysed using Spearman’s rank correlation test andfactor analysis based on principal component axes and varimax rotation, with the exceptionthat regression analysis was also used to assess the relationship between the number ofspecies and the surface areas of the meadows. The Kruskal-Wallis H-test was used toanalyse the main effect of management regime on the studied vegetation and soil variables.Pairwise comparisons between the meadow types and management classes in terms of themeasured variables were performed using the Mann-Whitney U-test.

Non-parametric tests were used because of the low number of study sites (n=3–9 permeadow type and per management regime), statistically significant differences in equalityof error variances (Levene’s test) in many cases, and the fact that the soil NO3

- -Nconcentrations at seven meadows were below the detection limit of 10 mg l−1 and thereforeno soil NO3

- -N results were available for those meadows. The soil S and P concentrationswere similarly below the detection limits (9 mg l−1 and 3.5 mg l−1) in four and twomeadows respectively. For these reasons, the number of samples varied from two to six forthe different meadow types and management regimes. The results were consideredsignificant at p<0.05 and as trends at p<0.1. The analyses were performed using SPSSStatistics 16.0 and 17.0 for Mac.

Results

Characteristics of vegetation and soil

We scored the total number of field-layer vascular plant species (excluding trees andshrubs) in order to get an overview of species richness in the urban meadows. Of the 252species found, 209 were forbs, 30 grasses, 10 pteridophytes, and 3 dwarf shrubs. Three forbspecies, Achillea millefolium, Anthriscus sylvestris and Stellaria graminea, appeared in allof the meadows and 70 were only found in one meadow. The grasses Alopecurus pratensis,Elymus repens, Festuca rubra and Poa pratensis, as well as the forbs Rumex acetosa,Taraxacum ssp., Veronica chamaedrys and Vicia cracca, were also among the mostfrequent species when assessed on the basis of the number of meadows in which theyoccurred (i.e. in ≥12 out of 18 meadows). Perennial species (70.9%) clearly dominated overannuals and biennials, and native species (46.5%) and archeophytes (36.7%) over speciesthat have become established in the 1700s or later, such as Impatiens glandulifera (inHåkansberg) and Lupinus polyphyllus (in Reijolankallio and Pajamäki).

The total number of species increased with increasing area of meadow, while the averagenumber of species per m2 and other vegetation variables did not depend on the meadow size(Fig. 1). There were 12.6±3.0 species per m2, the range being 7.6–18.0 (Table 2). In spiteof the general dominance of forbs (i.e. forb:grass ratio >1.0 in 12 out of 18 meadows), thegrasses Agrostis capillaris, Alopecurus pratensis, Elymus repens and Festuca rubra had thehighest average covers per m2 (9.44%, 7.42%, 7.34% and 6.38%, respectively), whenaveraged over all 18 meadows. The total cover of grasses was highest in the dry, regularlymanaged Uussilta meadow (83.6%) and lowest in the rocky, irregularly managedOrmusmäki meadow (5.0%). In addition to Achillea millefolium (4.44%), the most frequentforbs in terms of their average 1-m2 covers over all meadows were Viola tricolor (3.08%),Galium album (2.59%) and Galium verum (2.33%), though that of Galium verum resultedfrom its high cover of 22.13% in the Uussilta meadow. The highest total cover of forbs wasfound in Porvarinlahti (86.2%) and the lowest in Kunnarla (27.4%). The former also had

Urban Ecosyst

the highest Shannon-Wiener value (4.14), and the latter had the lowest one (1.37). Both ofthese meadows are dry, irregularly managed and belong to the most rural class, though thearea of the Porvarinlahti meadow is 10-fold that of the Kunnarla meadow. Agrostisstolonifera was one of the species that was only found in Porvarinlahti.

The above-mentioned Galium verum was the most frequent and Knautia arvensis(0.03%) the least frequent of the nitrophobic species in the 1 m2 quadrats. In comparison,the threatened Dianthus deltoides had an average cover of 0.14% and Campanularotundifolia 0.43%. Aegopodium podagraria was, in turn, the most frequent nitrophilicspecies, with an average cover of 1.17% and Lamium album the least frequent, with a coverof 0.12%, when averaged over all the meadows.

The surface soil was fine sandy moraine with an organic matter content of 3–20% and aweight:volume ratio of 0.8–1.25 kg l−1 in all except two meadows, in which the surface soilwas sandy moraine (Kunnarla) or mull (Reijolankallio). All of the meadows had acidic soil,with pH varying between 5.0 and 6.0. The chemical analyses showed wide variation in thenutrient and heavy metal concentrations among the meadows (Table 2). While the highestconcentrations of NO3

- -N, NH4+ -N and Ntot were 2 to 4-fold compared to the lowest ones,

the highest P concentration was 13-fold compared to the lowest. The differences betweenthe highest and lowest heavy metal concentrations ranged from 2.5-fold for Fe to 16-foldfor Cu. Soil from the rocky, unmanaged Håkansberg had the lowest concentrations of P, K,Mg and Cu, and the highest concentrations of NO3

- -N and Ntot, and consequently thehighest Ntot:P and Ntot:K ratios. In turn, soil from the rocky, irregularly managedOrmusmäki was characterized by the highest Ca, S and Cu concentrations. Soil pH was

y = 13.971ln(x)+97.64 R2 = 0.371p = 0.007

y = -0.516ln(x)+11.88 R2 = 0.017p = 0.585

Fig. 1 The total number of spe-cies and number of species perm2 in relation to surface area ofstudied meadows in the HelsinkiMetropolitan Area in the summerof 2007

Urban Ecosyst

positively correlated with Ca, Mg, NH4+ -N and Cu concentrations (p=0.003, p=0.032, p<

0.001 and p=0.068, respectively) and negatively with Mn concentration (p=0.001).

Differences between meadow types

The classification of meadows into rocky, dry and mesic types, is mainly considered todescribe soil moisture and productivity of the meadows. Vegetation height was measured tosubstantiate this and the Kruskal-Wallis test showed a significant main effect of meadowtype (χ2=8.864, p=0.012) on vegetation height, the means ± SDs being 39±8 cm in mesic,27±9 cm in dry, and 22±4 cm in rocky meadows. Similarly, the total cover of field layervascular plants was highest in the mesic meadows (118±12%), followed by dry (110±38%)and rocky meadows (66±14%), the latter type differing at p<0.05 from both the formergroups. The Shannon-Wiener index and the cover of nitrophilic species were also greater inmesic than rocky meadows, and cover of grasses was greater in both mesic and drymeadows than in rocky meadows (all p<0.05). In contrast, the forb:grass ratio and the soil

Table 2 Mean ± SD and number of samples for the studied vegetation and soil chemistry variables, andSpearman rank correlation coefficients significant at <0.05 (bold) or <0.1 (regular), for relationships withmeadow type, urbanization level indicators and management regime

Variable Mean ±SD

n Meadowtype

Inhabitants(km−2)

Roads(km)

Distance toroad (km)

Managementregime

Species m−2 12.6±3.0 18 ns ns ns ns 0.474

Shannon-Wiener 2.70±0.75 18 0.515 ns ns ns 0.410

Nitrophobic cover (%) 3.22±5.25 18 ns −0.647 −0.570 ns ns

Nitrophilic cover (%) 3.82±5.54 18 0.610 −0.590 ns ns ns

Legume cover (%) 2.31±2.37 18 ns ns ns ns 0.437

Forb cover (%) 52.6±17.4 18 ns ns ns ns ns

Grass cover (%) 43.2±24.0 18 0.696 ns ns ns ns

Forb:grass 1.98±2.50 18 −0.526 ns ns ns ns

pH 5.58±0.32 18 ns ns ns 0.526 ns

NO3- -N (mg l−1) 19±7 11 −0.604 ns 0.615 ns −0.803

NH4+ -N (mg l−1) 7±3 18 ns ns ns 0.488 ns

Ntot 26.0±6.3 11 −0.630 ns 0.561 ns −0.829P (mg l−1) 13.0±12.3 16 ns −0.607 −0.541 ns ns

K (mg l−1) 133±58 18 0.403 ns ns ns ns

Ntot:P 3.54±1.93 9 −0.699 0.708 ns ns −0.543Ntot:K 0.22±0.11 11 −0.645 ns 0.845 ns −0.526Ca (mg l−1) 1,540±980 18 ns ns ns ns ns

Mg (mg l−1) 115±48 18 ns ns ns ns ns

S (mg l−1) 16±7 14 −0.604 ns ns ns ns

Cu (mg l−1) 9.4±8.7 18 ns ns ns ns ns

Mn (mg l−1) 64±42 18 ns −0.638 −0.412 ns ns

Zn (mg l−1) 23.5±14.1 18 −0.503 ns ns ns ns

Fe (mg l−1) 836±203 18 −0.621 0.553 0.480 −0.571 −0.484

Meadow type: 1 = rocky, 2 = dry, 3=mesic

Management: 0 = unmanaged, 1 = irregularly managed, 2 = regularly managed

Ntot = NO3- -N + NH4

+ -N

Urban Ecosyst

NO3- -N, Ntot, S, Zn, and Fe concentrations, as well as Ntot:P and Ntot:K ratios, were lowest

in the mesic meadows (Table 2, Fig. 2). For example, the soil Zn concentration increasedfrom 13.3 mg l−1 in mesic meadows to 25.1 mg l−1 in dry meadows and still further to30.1 mg l−1 in rocky meadows.

0

5

10

15

20

Nu

mb

er o

f sp

ecie

s m

-2

0

2

4

6

8

10

Fo

rb:g

rass

0

10

20

30

40

50

So

il N

O3- -

N

(mg

l-1)

0

5

10

15

So

il N

tot:P

0

0.2

0.4

0.6

0.8

So

il N

tot:K

Rocky Dry Mesic

Meadow type

a a

a

a

ab

b

aa

a

a

b ab

a

a a

Fig. 2 Mean ± SD for totalnumber of species, number ofspecies per m2, forb:grass ratio,and soil NO3

- -N concentrationand the soil Ntot:P and Ntot:Kratios of the studied meadowtypes. Letters indicate differencesat p<0.05 between the meadowtypes (Mann-Whitney U)

Urban Ecosyst

Agrostis capillaris and Galium verum were characteristic species of dry meadows, whileViola tricolor thrived in rocky meadows. Alopecurus pratensis and Achillea millefoliumwere frequent in both dry and mesic meadows, while Elymus repens was most frequent inmesic meadows. The highest site-specific covers of the nitrophilic species Lamium album(11.17%), Aegopodium podagraria (10.00%) and Anthriscus sylvestris (9.17%) were foundfrom dry meadows.

Urbanization indicators and effects

Of the urbanization indicators tested, population density and length of roads within a 1-kmbuffer zone were positively correlated with each other (rS=0.851, p<0.001) and negativelycorrelated with the total area of open habitat in the buffer zone (rS= −0.680, p=0.002 andrS= −0.505, p=0.033, respectively). The indicators of urbanization level were correlated tovarious plant and soil variables. The frequency of nitrophobic species and concentrations ofsoil P and Mn decreased, and NO3

- -N and Fe concentrations, together with Ntot:P and Ntot:K ratios, increased with increasing population density and length of roads in the buffer zone(Table 2, Fig. 3). The significant positive correlation indicated by the Spearman rankcorrelation analysis between soil NO3

- -N concentration and road length clearly suggeststhat one of the major sources of soil N is traffic. Secondly, soil Fe concentration alsoincreased with decreasing distance to trunk roads, while soil pH and NH4

+ -N concentrationwere highest, the farther away from roads the meadows were located. The cover of bothnitrophobic and nitrophilic species increased with increasing area of open habitat in thebuffer zone (rS=0.804, p<0.001 and rS=0.501, p=0.034, respectively). The latter probablyarose from the high frequencies of Anthriscus sylvestris (9.17%) and Lamium album(11.17%) in the (rural) meadow at Bemböle, where the hay is not removed after mowing.

Differences between management regimes

Management regime did not correlate with meadow type or urbanization indicators (datanot shown) thus allowing us to focus on main effects of management on vegetation andsoil. Regular management increased the number of species per m2 (Table 2, Fig. 4).Moreover, the Shannon-Wiener diversity index was highest in the managed meadows(3.04±0.53), followed by irregularly managed (2.61±0.90) and thirdly unmanagedmeadows (2.29±0.42), the difference between the regularly managed and unmanagedmeadows being significant at p=0.071. The pairwise comparisons suggested that the coverof nitrophobic species increased with management, the unmanaged meadows (0.2%)differing from the irregularly managed meadows (3.3%) at p=0.077 and from the regularlymanaged ones (5.0%) at p=0.092. Management also tended to increase the cover oflegumes (unmanaged 0.9%, irregularly managed 2.0% and regularly managed 3.5%). Itshould be noted that the mesic meadow at Herttoniemi, which was the only one of theregularly managed meadows that had been mown twice in each summer, did not differ fromall the other meadows in terms of species richness and diversity or soil characteristics.

The soil NO3- -N, Ntot and Fe concentrations in particular decreased and the Ntot:P and

Ntot:K ratios tended to decrease with increasing management frequency (Table 2, Fig. 4).The Kruskal-Wallis analysis also showed significant main effects of management on soilNO3

- -N (χ2=6.572, p=0.037), NH4+ -N (χ2=5.429, p=0.066) and Ntot (χ

2=6.868, p=0.032) concentrations. The average NO3

- -N concentration was double and the Ntot:P ratioabout 4-fold in the unmanaged meadows compared to the regularly managed ones(27.5 mg l−1 cf. 12.5 mg l−1, and 8.33 cf. 2.20, respectively).

Urban Ecosyst

Combined effects of meadow type, urbanization, management and soil chemistry

Factor analysis (Table 3, Fig. 5) produced six principal components (PCs) that supportedthe results of the Spearman rank correlation analyses (Table 2) on the effects ofurbanization level indicators, meadow type and management regime on the studied

0

5

10

15

20

Nu

mb

er o

f sp

ecie

s m

-2

0

2

4

6

8

10

Fo

rb:g

rass

0

10

20

30

40

50

So

il N

O3- -

N

(mg

l-1)

0

5

10

15

So

il N

tot:P

0

0.2

0.4

0.6

0.8

So

il N

tot:K

<1000 1000-2500 >2500

Inhabitants km-2

a a a

a

b

a

a

a a

a ab

b

a a

a

Fig. 3 Mean ± SD for totalnumber of species, number ofspecies per m2, forb:grass ratio,soil NO3

- -N concentration andthe soil Ntot:P and Ntot:K ratios ofmeadows classified according tothe population density within a1-km buffer zone around theirborders. Letters indicate differ-ences at p<0.05 between theurbanization levels in terms ofpopulation density(Mann-Whitney U)

Urban Ecosyst

vegetation and soil variables. Thus PC1 can be considered to represent the “urban-rural”gradient that relates especially to urbanization effects on soil pH. It explained 28.6%, whilstPC2 explained 21.4% of the total variation. PC2 corresponds to plant biodiversity inrelation to soil Zn in particular, since the number of species per m2, Shannon-Wiener index,cover of legumes and total cover of forbs, decreased with increasing soil Zn concentration(Table 4).

0

5

10

15

20

Nu

mb

er o

f sp

ecie

s m

-2

0

2

4

6

8

10

Fo

rb:g

rass

0

5

10

15

So

il N

tot:P

0

0.2

0.4

0.6

0.8

So

il N

tot:K

Unmanaged Irregular Regular

0

10

20

30

40

50

So

il N

O3- -

N

(mg

l-1)

Management

a

ab b

a

a

a

ab

b

a

a

b ab

a

a a

Fig. 4 Mean ± SD for totalnumber of species, number ofspecies per m2, forb:grass ratio,and soil NO3

- -N concentrationand the soil Ntot:P and Ntot:Kratios of meadows belonging todifferent management classes inthe Helsinki Metropolitan Areaduring the summer of 2007. Let-ters indicate differences at p<0.05 between the managementregimes (Mann-Whitney U)

Urban Ecosyst

Table 4 also shows that nitrophobic species were more frequent the lower the soil pHwas and that the cover of forbs and legumes, i.e. N2-fixing forbs (Lathyrus pratensis,Trifolium pratense, Trifolium repens, Trifolium spadiceum, Vicia cracca, Vicia hirsuta,Vicia sepium, Vicia tetrasperma), were negatively correlated with soil NO3

- -N and Ntot

concentrations. The effect of soil K concentration on vegetation was evident in thesignificant positive correlation between the frequency of nitrophilic species and soilK concentration, and as negative trends between the frequency of nitrophilic speciesand soil Ntot:K ratio, and between the forb:grass ratio and soil K concentration (Tables 3and 4).

Table 3 Rotated component matrix of principal component analysis (PCA; varimax rotation with Kaisernormalization) showing variable loading values for the six PC axes. Loading values higher than 0.5 andlower than−0.5 are given in bold

Variable PC1 PC2 PC3 PC4 PC5 PC6

nitrophilic −0.902 0.280

P −0.832 −0.277 −0.242inhabitantskm2 0.734 0.348 0.356 0.443

openhabitat −0.714 0.417 −0.427 0.287

roadsbuffer 0.701 0.259 0.421 0.347

NtoP 0.665 0.420 −0.291 0.433

ShannonWiener 0.894 0.272

S 0.879 −0.221speciesm2 0.459 0.831

forbs −0.424 0.825 0.238 −0.225Ca 0.991

Mg 0.374 0.762 0.439

pH 0.322 −0.335 0.736 −0.431Cu 0.695 0.670

NH4 −0.390 0.490 −0.397 −0.477 −0.358K 0.247 0.908

grasses 0.341 0.855

forbstograsses −0.236 0.235 0.229 −0.817 −0.245NtoK 0.307 −0.807 0.409

legumes 0.556 −0.785 0.209

nitrophobic −0.360 0.383 −0.346 0.756

NO3 0.390 0.754 0.478

Ntot 0.525 0.746 0.327

Mn −0.473 −0.579 0.592 −0.246Zn −0.426 −0.467 0.234 −0.434 0.553

Fe 0.339 0.877

distancetoroad 0.251 −0.572 −0.345 −0.640managementregime 0.457 −0.224 0.310 −0.440 −0.636meadowtype −0.444 0.571 −0.617Cumulative variance (%) 28.6 50.0 68.0 80.3 88.4 94.1

Urban Ecosyst

Discussion

The main focus of our field study was to yield essential ecological knowledge that can beintegrated into ecosystem and site-specific management procedures and into urban planningfor the conservation of threatened species and their habitats. Our sites included rocky, dryand mesic meadows, of which the rocky meadows represent the driest and the leastproductive type, require the least management and are probably least threatened bydevelopment. With this approach, we try to provide an overall view on the long-term

Fig. 5 Plot of the first two principal components (PC) from the principal component analysis of vegetationand soil variables, meadow type and management regime, as well as population density (inhabitantskm2),road length and area of open habitat within a 1-km radius of the sites (roadsbuffer and openhabitat,respectively), and distance of meadows to the nearest trunk road (distancetoroad)

Table 4 Spearman rank correlation coefficients significant at <0.05 (bold) or <0.1 (regular), for relationshipsbetween vegetation and soil variables

pH NO3- -N

(mg l−1)Ntot

(mg l−1)K(mg l−1)

Ntot:K Ca(mg l−1)

Mn(mg l−1)

Zn(mg l−1)

Species m−2 −0.505Shannon-Wiener −0.527Nitrophobic cover (%) −0.527 −0.437 0.747

Nitrophilic cover (%) 0.475 −0.587 0.432

Legume cover (%) −0.674 −0.694 −0.516Forb cover (%) −0.683 −0.694Grass cover (%) −0.495Forb:grass −0.436 0.431

Ntot = NO3- -N + NH4

+ -N

Urban Ecosyst

interacting effects of various abiotic factors on vascular plant diversity in acidic northernEuropean semi-natural grasslands of high conservation value.

Although the number of sites that could be included in the survey was low and we werelacking some data for individual sites, we found significant negative effects of urbanizationon forbs and positive effects of management on species number per m2, Shannon-Wienerindex and the cover of legumes. As hypothesized, these effects seemed to result fromeutrophication due to elevated NOx deposition and to be modified by meadow type and theeffect of management on soil NO3

- -N concentration. Soil pH and elevated concentrationsof metals such as Zn, may also play an important role as factors determining plant diversityin urban meadows.

Species diversity and nitrogen

Pykälä (2003) reported altogether 252 species in mesic semi-natural meadows in southernFinland, with 44 species per m2 in old pastures followed by 31 and 26 species per m2 innew and abandoned pastures, respectively. Compared to these figures, the urban meadowsin the Helsinki Metropolitan Area show high β diversity, whereas the average number ofonly 13 species per m2 indicates low α diversity. The former can be mainly attributed to thefact that our sites represent a more diverse range than those of Pykälä (2003), while thelatter indicates a general degradation of semi-natural grasslands in the area. The averageShannon-Wiener indices for dry and mesic meadows (both 2.98) were also slightly lowerthan those reported for rural meadows in southern Sweden (3.33 and 3.45) by Linusson etal. (1998).

Although the majority of N which is added to acid grasslands maybe immobilized(Phoenix et al. 2003), our results suggest that increased soil NO3

- -N concentration inparticular negatively affected the frequencies of forbs, including N2 fixing legumes. Stevenset al. (2004) reported species richness in 68 acid grasslands across Great Britain to havedeclined as a linear function of the rate of inorganic N deposition. The number of smallforbs in particular has declined in British habitats (Maskell et al. 2010), whilst Agrostiscapillaris, Festuca rubra, Achillea millefolium and Rumex acetosa, which were alsocommon in our urban meadows, have been found to clearly increase in managed butunfertilized semi-natural grasslands on nutrient-poor, acidic soils in Europe (Stevens et al.2004; Duprè et al. 2010). According to Clark and Tilman (2008), the decline in plantspecies number under low chronic rates of N addition result mainly from the loss of rarespecies. However, common species may also be sensitive to N fertilization (Suding et al.2005) and abandonment (Pykälä et al. 2005), and the loss of species may occur randomlywith respect to traits (Suding et al. 2005). The low average number of species per m2 andthe high β diversity in our meadows support these findings. Thus management for theenhancement of locally susceptible functional groups and generally vulnerable rare species,is essential to maintain biodiversity in these meadows (Suding et al. 2005).

Two out of the three highest soil NO3- -N concentrations (c. 26 and 43 mg kg−1 dry soil)

were measured from rocky, unmanaged meadows from the most urban area (>2,500inhabitants per km2). Higher soil total N concentrations were also found from urban sites ina study of forest sites in the Helsinki Metropolitan Area by Nikula et al. (2010). Moreover,Pouyat et al. (1997) and Pouyat and Turechek (2001) reported higher N mineralization andnitrification (oxidation of NH4

+ to NO3−) rates in acidic urban soils compared to rural forest

soils. Such results suggest increased N concentration in leaf litter (Morecroft et al. 1994;Clark et al. 2009; Nikula et al. 2010), especially in the vicinity of roads (Laffray et al.2010), as well as changes in other factors that affect microbial processes related to N

Urban Ecosyst

cycling, such as soil temperature, moisture, microbial C availability and availabilites ofother nutrients (Hart et al. 1994; West et al. 2006).

Dianthus deltoides was the only one of the selected nitrophobic species that is classifiedas an acidity indicator (i.e. having an R value of 3), the others being indicative of weaklyacid to weakly basic conditions or indifferent (Ellenberg et al. 2001). Thus the negativecorrelation between the frequency of nitrophobic species with soil acidity, which rangedbetween pH5.0–6.0, probably does not alone explain the strong negative loading ofnitrophobic species on PC1. We used Ellenberg N value (Ellenberg et al. 2001) as the basisfor selecting nitrophobic and nitrophilic indicator species. However, species reactions to N,as well as to soil acidity and trace metals, depend on all the environmental factors at the site(Ernst 2004; Clark et al. 2007), as is also shown by the present study. With regard to salttolerance of plants, we found Agrostis stolonifera, which typically occurs in verges of roadssubjected to winter salting (Truscott et al. 2005), only at the Porvarinlahti meadow, which islocated just a few hundred metres away from the sea.

The P and K concentrations in the surface soils of the studied meadows were also highcompared to those that have been measured from sandy soils in rural meadows (Kanerva etal. 2005). In spite of this, our results suggest that K restricts the growth of nitrophilicspecies in particular. Phosphorus is known to be both a limiting factor (e.g. Raatikainen etal. 2007) and to decrease species richness in grasslands (Tracy and Sanderson 2000). In thecase of the studied Helsinki sites, it is highly likely that P will become a limiting factor inthe most urbanized meadows unless they are managed.

Management and vegetation of urban meadows

The cover of nitrophobic species increased with decreasing population density and lengthof roads in the buffer zone, and was highest in the regularly managed meadows, thus alsosupporting our hypothesis that mowing and removal of hay mitigate the effects ofurbanization on grassland vegetation by reducing the soil NO3

- -N concentration inparticular. It is likely that the effect of mowing on soil NO3

- -N concentration will be highlyyear and site specific (Ilmarinen et al. 2009). The present study was conducted over a singlegrowing season and does not, therefore, enable us to assess the effects of weather andclimatic factors on soil NO3

- -N concentration, which vary both temporally and spatiallyalong the urban-rural gradient. Although the thin surface soil layer in rocky meadows,where the vegetation was also sparsest, may accumulate the most deposited N, the risk oflow forbs being overgrown by taller forbs and grasses, shrubs and eventually trees, isapparently highest in dry meadows. This deduction is based on the higher frequencies of thenitrophilic species, Lamium album, Anthriscus sylvestris and Aegopodium podagraria indry meadows, where K availability may limit their growth less than in more productivemesic meadows.

We found the three highest total numbers of species (≥100) in the most urban meadows.Two of these, Pajamäki and Tuorinniemi, are rocky meadows which are exposed to frequentrecreational disturbance. Such disturbance, as long as it does not become too intensive, mayhave a positive effect on forb richness, because it re-establishes earlier successional stages(Schadek et al. 2009) and enhances habitat heterogeneity. In the more moist meadows,where graminoids become dominant in the longer term, especially if affected by increasedN deposition, frequent management is essential (Berlin et al. 2000). However, lateflowering perennials that have transient seed banks and are reliant upon seed production forregeneration, might be lost if repetitively subjected to early cutting, especially if there is norecruitment from adjacent uncut habitats (Smith and Jones 1991). Additionally, two cuts

Urban Ecosyst

maybe required to gain control over coarse grasses and herbs such as Anthriscus sylvestris(Parr and Way 1988). The low incidence of the alien invasive species Impatiensglandulifera and Lupinus polyphyllus (Beerling 1993; Valtonen et al. 2006) suggests thatit is still possible to control them in the Helsinki meadow network if appropriatemanagement is initiated early enough.

The role of metals

Our results also show that management decreases soil Fe concentration and may thusmitigate the negative effects of metal deposition from traffic emissions, for instance, onvegetation. There is no correlation of Zn with road variables in this study. This leads us toassume that former industrial activity and/or energy production are more likely to be themain sources of Zn in this case (Harrison et al. 2003).

Our suggestions that relatively low Zn concentrations (about 5–57 mg kg−1 dry soil) areharmful to grassland vegetation is also supported by Hernández and Pastor (2008), whoreport negative correlations at p<0.05 between plant available Zn (0.2–198 ppm) andShannon-Wiener index, total number of species and number of Compositae species, in soilswith average pH5.9–6.7. Hernández and Pastor (2008) also demonstrated correlationsbetween plant available Zn and number of Graminae species and plant cover significant atp<0.1. Zn-sensitivity of both grasses (e.g. Agrostis capillaris, Festuca ovina) and forbs(e.g. Campanula rotundifolia, Viola tricolor) relates to their degree of mycorrhization (seereviews by Bååth 1989; Ernst 2004). Relatively low levels of Zn and Cu are also known tobe harmful to symbiotic organisms, such as Rhizobium leguminosarum bv. trifolii, and thisis supported by the negative correlation between the cover of legumes and soil Znconcentration.

Our results did not suggest any negative effects of soil Cu, though an increase in floraldensity with decreasing soil water-soluble concentration of Cu (1–5 µg g−1 dry mass) hasbeen found in mine tailings which contained elevated levels of Cu, Ni and sulphides andhad an average pH5–6 in the upper horizons (Bagatto and Shorthouse 1999). This may beexplained by the positive correlations between soil Cu and pH, Ca, Mg and Fe (Table 3,Fig. 5). Furthermore, Rooney et al. (2006) have shown that exchangeable Ca and soil cationexchange capacity are the best single predictors of Cu toxicity to plants. Inclusion of othersoil properties, such as iron oxide concentration, pH, clay, or organic carbon content,further improved their predictions.

Future prospects and research needs

Ecological management of existing vulnerable grassland habitats is of major importance.We conclude that the existing diversity of (perennial) vascular plants, especially short-statured forbs, can only be maintained in these urban remnant meadows by theimplementation of regular mowing in terms of the timing and frequency of cuts, and theremoval of hay which otherwise decomposes releasing nutrients and leading to overgrowthand reduction in plant species richness (Parr and Way 1988; Morgan and Lunt 1999).

Management removes excess NO3- -N and heavy metals and ameliorates competition for

light, which maybe one of the major mechanisms of plant diversity loss after eutrophication(Hautier et al. 2009), especially in mesic grasslands with their dense and high vegetation.Direct negative effects of NOx emissions on plants can also not be exluded, especially inthe vicinity of roads (Wellburn 1990; Honour et al. 2009), and the significance of elevatedtropospheric O3 levels as a stress factor may also be increasing. Annual mean O3

Urban Ecosyst

concentrations have increased in the Helsinki Metropolitan Area during the period 1994–2007 by 2.5% and 0.2% at urban and rural locations, respectively (Anttila and Tuovinen2010) and Rämö et al. (2006, 2007), for instance, have demostrated experimentally that O3

has negative effects on the growth and reproductive development of forbs such asCampanula rotundifolia and Vicia cracca in particular.

It should also be noted that the average size of the studied meadows was only 0.34 haand only one of the sites exceeded 1 ha. As the number of species increased with increasingarea of meadows and the cover of nitrophobic species increased with increasing area ofopen habitat in the vicinity of meadows, isolation and scarcity of suitable habitat clearlyalso constitute a problem for the species assemblages of these urban meadows. A study ofDianthus deltoides populations from the same meadows in the summer of 2008 showed thatits seed production is positively correlated with the area of adjacent meadow habitat and theabundance of certain nectar plant species (Berghem 2009). We consider that this is due tothe trophic interactions with assemblages of pollinating insects supported by thesesupplementary meadow habitats. This suggests that further fragmentation and loss ofsemi-natural grassland habitats may also threaten the persistence of Dianthus deltoides dueto decreasing abundance of pollinators (Jennersten 1988; Ekroos 2010). Additional habitatprovision is essential for the conservation of declining species in particular, because spatialfactors such as site location within a landscape and proximity to species-rich grasslands canoverride the effects of local soil properties (Cousins et al. 2009).

In addition to vegetation, this project includes studies on pollinating and regulatoryinsect taxa (e.g. aculeate Hymenoptera, Carabidae, Lepidoptera and Formicidae) from thesemeadows. Combining air quality and deposition data, patch level metrics (area, perimeter,perimeter:area ratio and nearest neighbour distances) (Williams et al. 2005), together withecological information on vegetation, insects, and soil biota and processes, would help todesign management and conservation startegies to protect ecosystem structure and function,and thereby enhance the provision of ecosystem services provided by these urban, semi-natural grasslands, such as pollination (see Dearborn and Kark 2010).

Acknowledgements We thank MSc Paula Vehko, MSc Päivi Ketola, Ms Laura Sandholm and Ms EveliinaTuurnas for helping with data collection. The project was funded by the City of Helsinki, Nylands nation vidHelsingfors universitet, Finnish Konkordia Fund, Societas Biologica Fennica Vanamo Societas pro Fauna etFlora Fennica and Academy of Finland (project 126915). We also acknowledge Mrs Inkeri Vähä-Piikkiöfrom Helsinki Urban Facts, and Mrs Tuuli Ylikotila and Mr Markku Heinonen from the Street and Park Unitof the Public Works Department for all their help.

References

Ahti T, Hämet-Ahti L, Jalas J (1968) Vegetation zones and their sections in northwestern Europe. Ann BotFenn 5:169–211

Angold PG (1997) The impact of a road upon adjacent heathland vegetation: effects of plant speciescomposition. J Appl Ecol 34:409–417

Anttila P, Tuovinen J-P (2010) Trends in primary and secondary pollutant concentrations in Finland in 1994–2007. Atmos Environ 44:30–41

Bååth E (1989) Effects of heavy metals in soil on microbial processes and populations (a review). Water AirSoil Pollut 47:335–379

Bagatto G, Shorthouse JD (1999) Biotic and abiotic characteristics of ecosystems on acidic metalliferousmine tailings near Sudbury, Ontario. Can J Bot 77:410–425

Bakker JP (1989) Nature management by grazing and cutting. Kluwer Academic Publishers, DordrechtBeerling DJ (1993) The impact of temperature on the northern distribution limits of the introduced species

Fallopia japonica and Impatiens glandulifera in north-west Europe. J Biogeogr 20:45–53

Urban Ecosyst

Berghem F (2009) Ketoneilikan (Dianthus deltoides L.) populaatiorakenne, fenologia ja siementuottopääkaupunkiseudun niityillä [Population structure, phenology and seed production of maiden pink(Dianthus deltoides L.) in the meadows of Helsinki Metropolitan Area]. MSc thesis, University ofHelsinki (In Finnish)

Berlin GAI, Linusson A-C, Olsson EGA (2000) Vegetation changes in semi-natural meadows withunchanged management in southern Sweden, 1965–1990. Acta Oecologia 21:125–138

Bignal KL, Ashmore MR, Headley AD, Stewart K, Weigert K (2007) Ecological impacts of air pollutionfrom road transport on local vegetation. Appl Geochem 22:1265–1271

Clark CM, Tilman D (2008) Loss of plant species after chronic low-level nitrogen deposition to prairiegrasslands. Nature 451:712–715

Clark CM, Cleland EE, Collins SL, Fargione JE, Gough L, Gross KL, Pennings SC, Suding KN, Grace JB(2007) Environmental and plant community determinants of species loss following nitrogen enrichment.Ecol Lett 10:596–607

Clark CM, Hobbie SE, Venterea R, Tilman D (2009) Long-lasting effects on nitrogen cycling 12 years aftertreatments cease despite minimal long-term nitrogen retention. Glob Chang Biol 15:1755–1766

Cousins SA, Lindborg R, Mattson S (2009) Land use history and site location are more important forgrassland species richness than local soil properties. Nord J Bot 27:483–489

Dearborn DC, Kark S (2010) Motivations for conserving urban biodiversity. Conserv Biol 24:432–440De Leeuw FAAM, Moussiopoulos N, Sahm P, Bartonova A (2001) Urban air quality in larger conurbations

in the European Union. Environ Model Softw 16:399–414Drebs A, Nordlund A, Karlsson P, Helminen J, Rissanen P (2002) Climatological statistics of Finland 1971–

2000. Finnish Meteorological Institute, HelsinkiDuprè C, Stevens CJ, Ranke T, Bleekers A, Peppler-Lisbach C, Gowing DJG, Dise NB, Dorland E, Bobbink R,

Diekmann M (2010) Changes in species richness and composition in European acidic grasslands over past70 years: the contribution of cumulative atmospheric nitrogen deposition. Glob Chang Biol 16:344–357

Ekroos J (2010) Effects of management and landscape structure on biodiversity in boreal agriculturalfarmland. Ph.D. thesis. University of Helsinki, Faculty of Biological and Environmental Sciences, http://ethesis.helsinki.fi

Ellenberg H, Weber HE, Düll R, Wirth V, Werner W (2001) Zeigerwerte von Pflanzen in Mitteleuropa, 3rdedn. Scr Geobot 18:1–251

Eriksson Å, Eriksson O, Berglund H (1995) Species abundance patterns of plants in Swedish semi-naturalpastures. Ecography 18:310–317

Ernst WHO (2004) The use of higher plants as bioindicators. In: Markert BA, Breure AM, Zechmeister HG(eds) Bioindicators & biomonitors. Principles, concepts and applications. Trace metals and othercontaminants in the environment, vol 6. Elsevier, the Netherlands, pp 423–463

European Environmental Agency, http://eunis.eea.europa.eu/habitats-factsheet.jsp?idHabitat=10126.Accessed 12 Mar 2010

Finnish Meteorological Institute (2007) Säätilastoja maalis-elokuu 2007 (Weather statistics March–August2007). http://www.fmi.fi/weather/. Accessed 12 Sep 2007, 8 Nov 2007, 3 Jan 2008

Fischer M, Stöcklin J (1997) Local extinctions of plants in remnants of extensively used calcareousgrasslands 1950–1985. Conserv Biol 87:319–325

Gilbert OL (1989) The ecology of urban habitats. Chapman & Hall, LondonHämet-Ahti L, Suominen J, Ulvinen T, Uotila P (eds) (1998) Retkeilykasvio (Field Flora of Finland), 4th

edn. Finnish Museum of Natural History, Botanical Museum, HelsinkiHämet-Ahti L, Kurtto A, Lampinen R, Piirainen M, Suominen J, Ulvinen T, Uotila P, Väre H (2005)

Lisäyksiä ja korjauksia Retkeilykasvion neljänteen painokseen (Additions and errata on the 4th editionof Field Flora of Finland). Lutukka 21:41–85 (In Finnish)

Harrison RM, Tilling R, Callén Romero MS, Harrad S, Jarvis K (2003) A study of trace metals andpolycyclic aromatic hydrocarbons in the roadside environment. Atmos Environ 37:2391–2402

Hart SC, Nason GE, Myrold DD, Perry DA (1994) Dynamics of gross nitrogen transformations in an old-growth forest: the carbon connection. Ecology 75:880–891

Hautier Y, Niklaus PA, Hector A (2009) Competition for light causes plant biodiversity loss aftereutrophication. Science 324:636–638

Helsinki Metropolitan Area Council (2003) SeutuCD’03 (RegionCD’03) (In Finnish)Helsinki Metropolitan Area Council (2006) Helsinki Metropolitan Area Vision 2025 — summary. YTV,

HelsinkiHelsinki Metropolitan Area Council (2007) Ilman laatu pääkaupunkiseudulla vuonna 2006 (Air quality in the

Helsinki Metropolitan Area in 2006). YTV, Helsinki (In Finnish)Hernández AJ, Pastor J (2008) Relationship between plant biodiversity and heavy metal bioavailability in

grasslands overlying an abandoned mine. Environ Geochem Health 30:127–133

Urban Ecosyst

Honour SL, Bell JNB, Ashenden TW, Cape JN, Power SA (2009) Responses of herbaceous plants to urbanair pollution: effects on growth, phenology and leaf surface characteristics. Environ Pollut 157:1279–1286

Ilmarinen K, Mikola J, Nissinen K, Vestberg M (2009) Role of soil organisms in the maintenance of species-rich seminatural grassland through mowing. Restor Ecol 17:78–88

Janssens F, Peeters A, Tallowin JRB, Bakker JP, Bekker RM, Fillat F, Oomes MJM (1998) Relationshipbetween soil chemical factors and grassland diversity. Plant Soil 202:69–78

Jennersten O (1988) Pollination in Dianthus deltoides (Caryophyllaceae): effects of habitat fragmentation onvisitation and seed set. Conserv Biol 2:359–366

Kanerva T, Regina K, Rämö K, Karhu K, Ojanperä K, Manninen S (2005) Mesocosms mimic naturalmeadows as regards greenhouse gas fluxes and potential activities of nitrifying and denitrifying bacteria.Plant and Soil 276:287–299

Kangas L, Syri S (2002) Regional nitrogen deposition model for integrated assessment of acidification andeutrophication. Atmos Environ 36:1111–1122

Kull K, Zobel M (1991) High species richness in an Estonian wooded meadow. J Veg Sci 2:711–714Laffray X, Rose C, Garrec J-P (2010) Biomonitoring of traffic-related nitrogen oxides in the Maurienne

valley (Savoie, France), using purple moor grass growth parameters and leaf 15N/14N ratio. EnvironPollut 158:1652–1660

Lankinen L, Selander P, Tikkanen T (eds) (2009) The State of Helsinki Region 2009. EuropeanComparisons, City of Helsinki Urban Facts

Linusson A-C, Berlin GAI, Olsson EGA (1998) Reduced community diversity in semi-natural meadows insouthern Sweden, 1965–1990. Plant Ecol 136:77–94

Maskell LC, Smart SM, Bullock JM, Thompson K, Stevens CJ (2010) Nitrogen deposition causeswidespread loss of species richness in British habitats. Glob Chang Biol 16:671–679

McGrath SP, Chaudri AM, Giller KE (1995) Long-term effects of metals in sewage sludge on soils,microorganisms and plants. J Ind Microbiol 14:94–104

Morecroft MD, Sellers EK, Lee JA (1994) An experimental investigation into effects of atmospheric nitrogendeposition on two semi-natural grasslands. J Ecol 82:475–483

Morgan JW, Lunt ID (1999) Effects of time-since-fire on the tussock dynamics of a dominant grass(Themeda triandra) in a temperate Australian grassland. Biol Conserv 88:379–386

Niemelä J (1999) Ecology and urban planning. Biodivers Conserv 8:119–131Nikula S, Vapaavuori E, Manninen S (2010) Urbanization-related changes in European aspen (Populus

tremula L.): leaf traits and litter decomposition. Environ Pollut 158:2132–2142Parr TW, Way JM (1988) Management of roadside vegetation: the long-term effects of cutting. J Appl Ecol

25:1073–1087Pearson J, Wells DM, Seller KJ, Bennett A, Soares A, Woodall J, Ingrouille MJ (2000) Traffic exposure

increases natural 15N and heavy metal concnetrations in mosses. New Phytol 147:317–326Phoenix GK, Booth RE, Leake JR, ReadDJ, Grime JP, Lee JA (2003) Effects of enhanced nitrogen deposition and

phosphorus limitation on nitrogen budgets of semi-natural grasslands. Glob Chang Biol 9:1309–1321Pouyat RV, Turechek WW (2001) Short- and long-term effects of site factors on net N-mineralization and

nitrification rates along an urban-rural gradient. Urban Ecosyst 5:159–178Pouyat RV, McDonnell MJ, Pickett STA (1997) Litter decomposition and nitrogen mineralization in oak

stands along an urban-rural land use gradient. Urban Ecosyst 1:117–131Pykälä J (2003) Effects of restoration with cattle grazing on plant species composition and richness of semi-

natural grasslands. Biodivers Conserv 12:2211–2226Pykälä J (2005) Plant species responses to cattle grazing in mesic semi-natural grassland. Agric Ecosyst

Environ 108:109–117Pykälä J, Luoto M, Heikkinen RK, Kontula T (2005) Plant species richness and persistence of rare plants in

abandoned semi-natural grasslands in northen Europe. Basic Appl Ecol 6:25–33Raatikainen KM, Heikkinen RK, Pykälä J (2007) Impacts of local and regional factors on vegetation of

boreal semi-natural grasslands. Plant Ecol 189:155–173Rämö K, Kanerva T, Nikula S, Ojanperä K, Manninen S (2006) Influences of elevated ozone and carbon

dioxide in growth responses of lowland hay meadow mesocosms. Environ Pollut 144:101–111Rämö K, Kanerva T, Ojanperä K, Manninen S (2007) Growth onset, senescence, and reproductive

development of meadow species in a competitive environment exposed to elevated O3 and CO2. EnvironPollut 145:850–860

Roem WJ, Berendse F (2000) Soil acidity and nutrient supply ratio as possible factors determining changesin plant species diversity in grassland and heathland communities. Biol Conserv 92:151–161

Rooney CP, Zhao F-J, McGrath SP (2006) Soil factors controlling the expression of copper toxicity to plantsin wide range of European soils. Environ Toxicol Chem 25:726–732

Urban Ecosyst

Rosén E, Borgegård S-O (1999) The open cultural landscape. Acta Phytogeogr Suec 84:113–134Schadek U, Strauss B, Biedermann R, Kleyer M (2009) Plant species richness, vegetation structure and soil

resources of urban brownfield sites linked to successional age. Urban Ecosyst 12:115–126Smith RS, Jones L (1991) The phenology of mesotrophic grassland in the Pennine Dales, Northern England:

Historic hay cutting dates, vegetation variation and plant species phenologies. J Appl Ecol 28:42–59Stevens CJ, Dise NB, Mountford JO, Goving DJ (2004) Impact of nitrogen deposition on the species

richness of grasslands. Science 303:1876–1879Suding KN, Collins SL, Gough L, Clark C, Cleland EE, Gross KL, Milchunas DG, Pennings S (2005)

Functional- and abundance-based mechanisms explain diversity loss due to N fertilization. PNAS102:4387–4392

Syri S, Fronzek S, Karvosenoja N, Forsius M (2004) Sulphur and nitrogen oxides emissions in Europe anddeposition in Finland during the 21st century. Boreal Environ Res 9:185–198

Thompson K, McCarthy MA (2008) Traits of British alien and native urban plants. J Ecol 96:853–859Tonteri T, Haila Y (1990) Plants in a boreal city: ecological characteristics of vegetation in Helsinki and its

surroundings, southern Finland. Ann Bot Fenn 27:337–352Tracy BF, Sanderson MA (2000) Patterns of plant species richness in pasture lands of the northeast United

States. Plant Ecol 149:169–180Truscott AM, Palmer SCF, McGowan GM, Cape JN, Smart S (2005) Vegetation composition of roadside

verges in Scotland: the effects of nitrogen deposition, disturbance and management. Environ Pollut136:109–118

UNECE (2004) Mapping Manual 2004. Manual on methodologies and criteria for modelling and mappingcritical loads and levels. Chapter III mapping critical levels for vegetation. http://icpmapping.org/cms/zeigeBereich/11/manual-english.html. Accessed 4 Apr 2010

Uusimaa Regional Council and Helsinki Metropolitan Area Council (2002) Uusimaa CD (In Finnish)Vähä-Piikkiö I, Kurtto A, Hahkala V (2004) Species number, historical elements and protection of threatened

species in the flora of Helsinki, Finland. Landsc Urban Plan 68:357–370Valtonen A, Jantunen J, Saarinen K (2006) Flora and lepidoptera fauna adversily affected by invasive

Lupinus polyphyllus along road verges. Biol Conserv 133:389–396Vidic T, Jogan N, Drobne D, Vilhar B (2006) Natural revegetation in the vicinity of the former lead smelter

in Zerjav, Slovenia. Environ Sci Technol 40:4119–4125Wellburn AR (1990) Why are atmospheric oxides of nitrogen usually phytotoxic and not alternative

fertilizers. New Phytol 115:395–429West JB, Hobbie SE, Reich PB (2006) Effects of plant species diversity, atmospheric [CO2), and N addition

on gross rates of inorganic N release from soil organic matter. Glob Chang Biol 12:1400–1408Willems JH (1982) Phytosociological and geographical survey of Mesobromion communities in Western

Europe. Vegetatio 48:227–240Williams NSG, McDonnell MJ, Seager EJ (2005) Factors influencing the loss of an endangered ecosystem in

an urbanising landscape: a case study of native grasslands from Melbourne, Austrialia. Landsc UrbanPlan 71:35–49

Wittig R, Becker U, Nawrath S (2010) Grassland loss in the vicinity of a highly prospering metropolitan areafrom 1867/68 to 2000 — The example of the Taunus (Hesse, Germany) and its Vorland. Landsc UrbanPlan 95:175–180

Urban Ecosyst