Effects of landscape composition, species pool and time on grassland specialists in restored semi-natural grasslands

Biological Conservation 214 (2017) 176–183

Contents lists available at ScienceDirect

Biological Conservation journal homepage: www.elsevier.com/locate/biocon

Effects of landscape composition, species pool and time on grassland specialists in restored semi-natural grasslands

MARK

Emelie Waldéna,⁎, Erik Öckingerb, Marie Winsab, Regina Lindborga a b

Department of Physical Geography, Stockholm University, 106 91 Stockholm, Sweden Department of Ecology, Swedish University of Agriculture, Uppsala, Sweden

A R T I C L E I N F O

A B S T R A C T

Keywords: Biodiversity Landscape composition Recolonisation Restoration Semi-natural grassland Species pool

Habitat restoration is an important complement to protecting habitat for the conservation of biodiversity. Seminatural grasslands are target habitats for ecological restoration in temperate Europe. Restoration of abandoned semi-natural grasslands often relies on spontaneous colonisation of plant species from the soil seed bank or the surrounding landscape. Although many studies show that the regional species pool is important for upholding local diversity, its effect on restoration outcome in semi-natural grasslands is poorly known. In this multilandscape study, we examined grassland specialist species occurring in restored grasslands and the effect of specialist species pool, landscape composition and local temporal factors. We found that specialist richness and frequency was positively affected by specialist richness and frequency in the surrounding landscape. Specialist richness in the restored grasslands also increased with time since restoration. Moreover, specialist frequency in the restored grassland increased with the proportion of semi-natural and remnant grassland habitats in the landscape. We also found a positive relationship between the proportion of species occurring in both the restored grassland and its surrounding landscape and time since restoration, in landscapes with high proportions of seminatural grasslands. This suggests that both temporal factors, as well as the landscape composition and species pool, affect plant recolonisation in restored semi-natural grasslands.

1. Introduction Habitat restoration is included in the UN Sustainable Development Goals and the Convention for Biological Diversity (2010 Aichi Biodiversity Target D) as a key measure to counteract current losses of biodiversity and ecosystem services. One goal is to globally restore 15% of degraded natural and semi-natural habitats of high biological value by the year 2020 (CBD, 2012). Permanent semi-natural grasslands in Europe have developed through centuries by grazing and mowing to become biological hotspots of high conservation value in the agricultural landscape (Habel et al., 2013; Papanikolaou, Kühn, Frenzel, and Schweiger, 2016). However, due to agricultural intensification and abandonment, remaining semi-natural grasslands are often heavily fragmented (Cousins, Auffret, Lindgren, and Tränk, 2015) and degraded in quality (Kasari, Saar, de Bello, Takkis, and Helm, 2016). Economic compensation to restore abandoned and degraded semi-natural grasslands are therefore incorporated in agri-environment schemes (AES) in many European countries (Stoate et al., 2009). Previous studies have shown that community assembly and species persistence are influenced by processes acting at both local (Dupre and

⁎

Corresponding author. E-mail address: [email protected] (E. Waldén).

http://dx.doi.org/10.1016/j.biocon.2017.07.037 Received 3 April 2017; Received in revised form 27 July 2017; Accepted 31 July 2017 0006-3207/ © 2017 Elsevier Ltd. All rights reserved.

Ehrlen, 2002; Krauss, Klein, Steffan-Dewenter, and Tscharntke, 2004; Lindborg et al., 2012) and landscape scales (Eriksson, 1996; Kormann et al., 2015; Öckinger, Lindborg, Sjödin, and Bommarco, 2012). Although the influence of the regional species pool is widely acknowledged (Hanski, 1999; Pärtel, Bennett, and Zobel, 2016; Pärtel, SzavaKovats, and Zobel, 2011; Zobel, van der Maarel, and Dupré, 1998), studies on how it affects restoration outcome in local habitats are rare (but see Conradi and Kollmann, 2016; Prach, Fajmon, Jongepierová, and Řehounková, 2015 regarding species pool effects on recreated grasslands). Whereas creation of new grasslands on ex-arable fields often includes manually sowing target seed mixes, restoration of abandoned semi-natural grasslands in northern Europe most often relies on plants recolonising spontaneously, either from the soil seed bank or the surrounding landscape (Waldén and Lindborg, 2016). This could potentially be a cost-efficient restoration method, but requires that targeted species still occur locally within the site or regionally within the landscape so that they are able to recolonise the restored habitat (Prach and Hobbs, 2008; Török, Vida, Deák, Lengyel, and Tóthmérész, 2011). Many grassland species exhibit a time-lag before they go extinct following grassland abandonment (Bagaria, Helm, Rodà, and Pino,

Biological Conservation 214 (2017) 176–183

E. Waldén et al.

Pärtel et al., 2011). This could for example be due to their low dispersal ability and/or low competitive ability (Poschlod, Kiefer, Tränkle, Fischer, and Bonn, 1998; Riibak et al., 2015), or related to temporal and/or spatial issues, such as to not yet suitable abiotic or biotic conditions (Helsen et al., 2016; Piqueray et al., 2011) or lack of functional connectivity (Auffret et al., 2017). As a full survey of the total species pool in a landscape is extremely time consuming, using GIS-analyses of landscape features have been suggested as substitute (Perring et al., 2015), where a high proportion of semi-natural grasslands in the landscape could indicate better opportunity for community recovery within restored semi-natural grasslands. However, these methods disregard the possibilities of grassland specialists inhabiting remnant grassland habitats, such as former grasslands or midfield islets (Cousins and Aggemyr, 2008). In this study, we examined the relationship between the habitat specific plant species pool (including managed semi-natural grasslands and remnant grassland habitats) and the plant specialists species found in restored semi-natural grasslands. The questions we posed were: How is the number and frequency of grassland specialists in restored and reference grasslands affected by (1) the temporal and local factors time since restoration, abandonment time (years without management) and focal grassland area, and (2) by the landscape factors proportion of semi-natural grasslands, remnant habitats and arable fields and the number and frequency of plant specialists found in the surrounding landscape? (3) How is the proportion of shared species (i.e. grassland specialists occurring in both the focal restored/reference grassland and the surrounding landscape) affected by time since restoration, abandonment time, grassland area and semi-natural grasslands, remnant habitats and arable fields in the landscapes?

2015; Krauss et al., 2010; Kuussaari et al., 2009), persisting either as perennial adults or as seeds in the seed bank (Auffret and Cousins, 2011; Helm, Hanski, and Partel, 2006; Lindborg, 2007). This time delay in the extinction process, the extinction debt (Kuussaari et al., 2009), enables populations to recover rapidly when habitat conditions within restored grasslands have become suitable again (Havrdová, Douda, and Doudová, 2015; Plue and Cousins, 2013). Nevertheless, a long time between grassland abandonment and restoration has a negative effect on restoration potential in grasslands (Bossuyt, Honnay, Van Stichelen, Hermy, and Van Assche, 2001; Schrautzer, Jansen, Breuer, and Nelle, 2009; Willems, 2001). Long-term effect of abandonment often results in the dominance of few competitive plant species while grassland specialists disappear (Willems and Bik, 1998), especially if the grassland has been fertilised (Fagan, Pywell, Bullock, and Marrs, 2008; Janssens et al., 1998). Another temporal factor affecting restoration outcome is the elapsed time since restoration. Usually the overall plant species richness increase with time since restoration (e.g. Piqueray et al., 2011; Waldén and Lindborg, 2016; Winsa, Bommarco, Lindborg, Marini, and Öckinger, 2015), although this will not necessarily indicate an increase in the species that characterise the habitat (cf. Helm, Zobel, Moles, Szava-Kovats, and Pärtel, 2015). While common species often immigrate to restored grasslands (Kotiluoto, 1998), specialists and rare species often show no or only slight recovery (e.g. Helsen et al., 2013; Pykälä, 2003; Tikka et al., 2001; but see Dzwonko and Loster, 1998). Gijbels, Adriaens, and Honnay (2012) showed that target orchid species were only present in half of the estimated suitable habitats, even three decades after restoration. Due to the low colonisation potential of grassland specialists species (Pywell et al., 2003), reference sites usually have a higher fraction of grassland specialists than restored sites (Lindborg and Eriksson, 2004; Schrautzer et al., 2009). The potential species to recolonise restored habitats from the landscape species pool is determined by species dispersal ability (Riibak et al., 2015), but also on the landscape structure (Cousins, 2006). In modern European agricultural landscapes, remaining species-rich grasslands are often small and isolated, resulting in low possibilities for species to disperse between grassland fragments (Eriksson, Cousins, and Bruun, 2002). However, except for continuously managed semi-natural grasslands, populations of typical grassland species can also survive in other types of habitats, e.g. in former grasslands that now are abandoned, road verges, gardens and midfield islets (Cousins, 2006; Lindgren and Cousins, 2017; Plue and Cousins, 2013). Such habitats might facilitate species dispersal into restored grasslands, acting as stepping stone habitats (Cousins and Lindborg, 2008; Lindborg, Plue, Andersson, and Cousins, 2014). Excluding these potentially suitable habitats when surveying plant species, could therefore underestimate the available species pool. Although dispersal from surrounding habitats is a common presumption both in restoration practise and research, few studies exist that build on field surveys of the actual species pool and its potential effect on restoration outcome in restored semi-natural grasslands. One recent study of grasslands recreated on ex-arable fields showed a positive relationship between the number of target plant species in the recreated grasslands and their occurrence in the surrounding landscape (Prach et al., 2015). Remnant grassland habitats may also act as source communities when they are directly connected to recreated grasslands (Cousins and Lindborg, 2008). Even though the abiotic and biotic preconditions in recreated grasslands are fundamentally different from restored abandoned semi-natural grasslands (Fagan et al., 2008; Horrocks et al., 2016), this indicates that target species from a species pool are able to recolonise spontaneously. How large proportion of the species pool in the landscape that is represented in different types of restored semi-natural grasslands is still relatively unknown. Although the surrounding landscape might host a large habitat specific species pool, some species could still be absent in the local restored habitat (i.e. belong to the dark diversity) (Lewis, Szava-Kovats, and Pärtel, 2016;

2. Methods 2.1. Study area and geographical analyses We selected 20 circular landscapes (1 km radius, no overlap between landscapes) situated in south-central Sweden in the counties of Södermanland, Uppland and Västmanland (Geographical coordinates in Appendix A). Each landscape was centered around one focal grassland; either a restored semi-natural grassland (12 sites) or a continuously managed semi-natural grassland (8 sites). The restored grasslands were abandoned semi-natural grasslands restored 6–23 years before our study. Restoration included clearing of trees and shrubs and reintroduction of domestic grazers (cattle (10 sites), horses (1) and sheep (1)). The continuously managed grasslands here act as reference grasslands, representing intact communities and the desired state after restoration. When discussing the restored and reference grasslands together, we refer to them as ‘focal grasslands’. All focal grasslands were chosen by using information from the County Administration Boards, the Municipalities and the Uppland Foundation, combined with information from a national Swedish geographical database of semi-natural pastures (http://www.jordbruksverket.se/tuva), previously described by Winsa et al. (2017). The grasslands were selected according to standardised criteria regarding soil conditions, the state before restoration actions (degree of degradation before restoration and restoration practice and effort), and are all situated in counties with relatively similar abiotic conditions (Lindborg and Eriksson, 2004; Steiner, Öckinger, Karrer, Winsa, and Jonsell, 2016). The average grassland area was similar for restored and reference grasslands (3.23 ± 0.60 and 3.11 ± 0.62 ha, respectively). Information of time since restoration (i.e. years since restoration was initiated) and abandonment time (i.e. years without management) were obtained by asking farmers, landowners and the County Administration Boards. Time since restoration varied between 6 and 23 years and abandonment time varied between 0 and 60 years (where 0 years refers to low intensity grazing, insufficient to fully prevent succession, during the last 50 years). All focal grasslands had dry to mesic abiotic conditions and 177

Biological Conservation 214 (2017) 176–183

E. Waldén et al.

Fig. 1. Example of landscapes studied in south-central Sweden with different proportions of currently managed semi-natural grasslands (a: 15.9%, b: 0%) and remnant habitats (i.e. former semi-natural grasslands managed in the 1950ies and midfield islets) surrounding the focal restored or reference semi-natural grassland. The black dots indicate the 70 plots per landscape where grassland plant species pool was inventoried.

grasslands. To estimate the occurrence frequency of each plant species, each 1m2 plot was divided into 100 smaller squares of 10 cm × 10 cm. To examine the species pool in the landscapes, the grassland specialists were inventoried in 70 1 m × 1 m plots in each landscape. In each plot species presence/absence and frequency (scale 0–100) was recorded. The 70 plots were placed randomly on set distance intervals using ArcGIS, but according to a priority order of probable habitats (Fig. 1). Since the number of grassland specialists in the vegetation tends to decrease with time since abandonment (Plue and Cousins, 2013), the priority order was as follows; (1) semi-natural grasslands that are managed today (SNGs), (2) former semi-natural grasslands managed in the 1950ies and (3) midfield islets. Thus, if no SNGs (i.e. priority 1) occurred at the set distance interval, the plot was placed in former semi-natural grasslands managed in the 1950ies (priority 2), and so on. This was done to prevent risks of placing a plot on an obviously unsuitable habitat, such as on a road or arable field. All plots were placed at a distance of minimum 50 m from each other. In total, 1400 plots were inventoried in the surrounding landscape, whereof 986 plots were placed in SNGs, 362 plots in former semi-natural grasslands and 52 plots in midfield islets.

were continuously distributed along a gradient of managed semi-natural grassland (SNG) area in the surrounding landscape (1 km radius), ranging from 0% to 15.8% (Fig. 1). Some of the landscapes did not contain any SNGs besides the focal grassland according to the national register (TUVA). However, when controlling in field a few additional SNG patches were discovered and thus included in the following data analyses. We also extracted geographical information about remnant habitats and arable fields in each landscape (using aerial photographs and the Economic map provided by the Swedish National Land Survey, Lantmäteriet). The remnant habitats consisted of former semi-natural grasslands managed in the 1950ies and current midfield islets (small bedrock outcrops surrounded by crop-fields, historically grazed after harvest). For each landscape, the proportions of SNGs, remnant habitats and arable fields were calculated based on coverage area in ArcGIS.

2.2. Field inventories In the field study, we inventoried 30 vascular plant species typical for semi-natural grasslands in this region (hereafter called ‘grassland specialists’) (Appendix B). The grassland specialists were selected according to the following criteria (Ekstam and Forshed, 1997; Krauss et al., 2010); (1) they typically decrease in abundance in an early to intermediate successional phase (3–15 years after grassland abandonment), (2) their optimal occurrence are in traditionally managed seminatural grasslands (3) with dry to mesic abiotic conditions, and (4) they are easy to identify in the field even after flowering. We did not select the rarest grassland specialists since their occurrence in semi-natural grassland nature reserves is too low to be analysed (Lindborg and Eriksson, 2004). We also avoided non-typical species not aimed for in grassland restoration, such as generalist and ruderal species. Here, we chose to inventory typical perennial grassland specialists that have the potential to occur also in remnant habitats, such as road verges and midfield islets. By selecting according to these criteria, the 30 grassland specialists consequently have more similar species trait composition, compared to ruderal and generalist species (cf. Conradi and Kollmann, 2016). We surveyed the presence of the selected grassland specialists in the landscape species pool and in the focal grasslands separately. Data of grassland specialist presence and frequency in the focal grasslands were collected from June to September in 2011 (Winsa et al., 2017). In each grassland, grassland specialists were recorded within 10 plots of 1 m × 1 m (i.e. 10m2/grassland). The locations of the plots were randomly plotted on a map to obtain an even distribution over the entire

2.3. Data analyses All statistical analyses were performed in R 3.3.0 (R Core Team, 2016). Prior to analyses, we balanced the landscape diversity data sampled in the 70 plots to be able to compare it with the 10 plots sampled in the focal grasslands. This was done for all 20 landscapes by using rarefaction with 999 permutations (i.e. extracting 10 plots 999 times and calculating an average value for each landscape). For the species occurrence frequency data (scale 0–100/plot), we calculated an average for all 30 species in the 10 focal grassland plots and 10 rarefied landscape plots. Although some species were not found in the focal grasslands, they do occur in some of the surrounding landscapes (Waldén, unpublished data) and in the region, according to regional floras. Since presenting absence of species is interesting in a restoration context (cf. Pärtel et al., 2011), all species are included in the analyses. By calculating an average for all 30 species combined, we were able to analyse average species frequency per grassland, despite the non-occurrence of some species. The calculated frequency was log-transformed prior to analysis. Before answering our posed questions, we used Welch t-tests to analyse whether there was a difference between restored and reference grasslands regarding number, frequency and proportion of shared grassland specialists. We also analysed whether 178

Biological Conservation 214 (2017) 176–183

E. Waldén et al.

the specialists occurred more often in the landscape plots placed in SNGs, compared to plots placed in remnant habitats (i.e. former seminatural grasslands and midfield islets) (calculated average proportion of occurrence for the 30 specialist species in each habitat type for all landscapes, analysed in a Generalized Linear Mixed Model with ‘Landscape’ set as random factor, followed by Anova type III-test for significance, using the R package ‘lme4’ (Bates, Maechler, Bolker, and Walker, 2015)). Our two first questions regarding the potential relationship between the diversity metrics and the temporal and local factors (time since restoration, abandonment time and focal grassland area, question 1) and landscape factors (proportion of SNGs, remnant habitats and arable fields in the surrounding landscape and the number and frequency, respectively, of the species in the surrounding landscape, question 2) were tested in multiple analyses. We addressed these two questions separately for the number versus frequency of grassland specialists, using multiple Spearman Monte Carlo tests (R package ‘coin’ (Hothorn, Hornik, van de Wiel, and Zeileis, 2008)). This was analysed both for restored and reference grasslands combined and separately for the restored grassland sites. For our final question (3) regarding grassland specialists occurring in both the focal grassland and the surrounding landscape, we calculated an index of the proportion of shared species between the focal grassland and the landscape for each site. To test if the proportion of shared species differed between the restored and reference sites and if it correlated with any of the predictor variables, we used Spearman Monte Carlo tests. Further, we also tested eventual effects of the interaction between time since restoration and proportion of SNGs and remnant habitats in the surrounding landscape, on the proportion of shared species. This was done using a GLM (Gaussian distribution) where the marginal effect of each variable was analysed using Anova type II-test (Chi2-test) on the full model (R package ‘car’ (Fox and Weisberg, 2011)).

Fig. 2. Venn-diagrams showing the average proportion of grassland specialist plant species occurring only in the focal restored/reference (a/b) semi-natural grassland (i.e. Unique focal), only in the surrounding landscape (i.e. Unique landscape) and the average proportion occurring in both the focal grassland and the landscape (i.e. Shared).

occurred proportionally more frequently in the SNG plots (3.33 ± 0.04% CI) than in the remnant habitat plots (i.e. former seminatural grassland and the midfield islet plots (1.70 ± 0.03% CI)) (GLMM, p-value < 0.05). Most (8 out of 11) of the specialists that did not occur in any of the remnant habitat plots, were also rare in the SNG plots (occurred in < 1% of all inventoried plots). Similarly, the four most abundant specialists in the remnant habitat plots (Anthoxanthum odoratum, Campanula rotundifolia, Festuca ovina and Pimpinella saxifraga) were also more abundant in the SNGs (occurred in > 6% of all plots). Detailed information of grassland specialist occurrence in the focal grasslands and landscapes is provided in Appendix B. Out of the total specialist species pool, 27.4 ± 7.7% was shared between the restored grassland and surrounding landscape (i.e. occurred in both the restored grassland and its corresponding surrounding landscape, Fig. 2a). The proportion of shared species was similar for sites with a focal reference grassland (26.8 ± 7.8%, Fig. 2b, Table 1) (data for each site in Appendix A). The three most commonly shared species were also the most common grassland specialists overall (Anthoxanthum odoratum, Campanula rotundifolia and Festuca ovina, Appendix B and C). Although the restored grasslands had a slightly lower proportion of grassland specialists occurring uniquely in the focal grassland than the reference grasslands, the proportions were not significantly different (Fig. 3, Welch t-test, t = 1.13, p-value = 0.29). No specific specialist were overall uniquely found in the focal grassland, whereas eight specialists often (in ≥ 50% of all sites) were found only in the landscape and not in the corresponding focal restored grassland (Appendix C). Out of these eight specialists, five were never, or very seldom, shared between focal grassland and corresponding landscape (categorised as Unique, Appendix D). However, there were no significant differences regarding the dispersal or persistence traits between these five specialists more often found uniquely in the landscape, and four specialists more often shared (Appendix D).

3. Results 3.1. Grassland specialist diversity in the restored and reference grasslands, and surrounding landscapes Out of the 30 grassland specialists studied, the restored grasslands had an average of 4.50 ( ± 1.45 CI) specialists, whereas the reference grasslands had an average of 5.38 ( ± 1.85 CI) specialists (no significant difference between restored and reference grasslands, Table 1). There was no significant difference in average occurrence frequency of the grassland specialists between the restored (0.031 ± 0.012 CI) and reference (0.032 ± 0.019 CI) grasslands (Table 1). On average, the grassland specialists found in the landscape Table 1 Differences between restored (Rest., n = 12) and reference (Ref., i.e. Continuously managed, n = 8) focal semi-natural grasslands in south-eastern Sweden, regarding number of grassland specialist plant species (max. 30), average occurrence frequency of grassland specialists (calculated average for all 30 species per grassland) and proportion of shared grassland specialists between focal grassland and surrounding landscape (i.e. species occurring in both the focal grasslands and its corresponding landscape, circular radius 1 km). Analysed in Welch t-tests. Biodiversity measure

Mean ± 95% CI

Diff. in mean (rest. vs. ref.)

Welch t-test

No. grassland specialists

Rest.: 4.50 ± 1.45 Ref.: 5.38 ± 1.85

− 0.88

Frequency grassland specialists (average) Prop. shared grassland specialists

Rest.: 0.03 ± 0.01 Ref.: 0.03 ± 0.02

0

Rest.: 0.27 ± 0.08 Ref.: 0.27 ± 0.08

0

t = 0.80, df = 14.7, p = 0.44 t = −0.30, df = 11.2, p = 0.77 t = −0.11, df = 17.0, p = 0.91

179

Biological Conservation 214 (2017) 176–183

E. Waldén et al.

grasslands (Fig. 5). The proportion of arable fields in the surrounding landscape did not have a significant effect on any of the tested biodiversity metrics (Table 2). 4. Discussion Restoration of semi-natural grasslands has high priority in Europe. However, prioritising among different restoration actions is necessary due to constraints in budget, time and degree of threat. Although a landscape perspective is important in restoration (Perring et al., 2015), few studies on how restored semi-natural grasslands is affected by landscape composition and species pool exists. Here we show that specialist plant diversity in restored semi-natural grasslands was affected by both temporal factors (time since restoration and abandonment time), as well as spatial and landscape factors (grassland area, specialist species pool and proportion of semi-natural grasslands and remnant grassland habitats in the landscape). This highlights the importance of considering both spatial and temporal factors when planning for, and prioritising between, future restoration prospects. Moreover, we found no difference between restored and reference grasslands concerning grassland specialist richness, frequency or proportion of specialists occurring in both grassland and surrounding landscape. These findings suggest that grassland specialist species with time could become equally common as in reference grasslands, with preconditions such as short abandonment time and a specialist-rich species pool in the surrounding landscape. The grassland specialist richness increased with time since restoration and decreased with abandonment time. Although not directly tested in this study, these two temporal variables could reflect different colonisation patterns; dispersal in space through recolonisation from surrounding landscape, and in time from the seed bank. Dispersal from surrounding species pool may take time, resulting in a higher specialist species richness in older restored grasslands (Waldén and Lindborg, 2016). Long-lived grassland plants may also be positively affected by local habitat area (Lindborg et al., 2012). Congruently, we found a positive trend between semi-natural grassland area and specialist species richness and frequency, which might be due to the increased probability of suitable microhabitats in larger grasslands (Deák et al., 2015; Pykälä, 2003; Öckinger and Smith, 2006).In addition, a short abandonment time may preserve specialist species in the seed bank (Kalamees, Püssa, Zobel, and Zobel, 2012) and facilitate local recruitment after restoration. However, since these two temporal variables were slightly correlated in this study, it may be difficult to disentangle the effect of one from the other. Nevertheless, our results indicates that temporal factors needs to be considered when targeting possibilities for

Fig. 3. Relationship between the number of grassland specialist plant species found in focal restored semi-natural grasslands in south-central Sweden, and time since restoration (in years) to the left (Spearman Monte Carlo, z = 2.62, p = 0.003).

3.2. Temporal and spatial factors affecting restored and reference grassland diversity The number of grassland specialists found in the restored grasslands was positively affected by time since restoration (Fig. 3, Table 2) and negatively affected by abandonment time (Table 2, although it should be noted that time since restoration and abandonment time correlated (Pearson correlation, r = − 0.64, p = 0.03). The specialist richness in focal grasslands was also positively related to the focal grassland area (Table 2), the specialist richness in the landscape (Fig. 4, Table 2) and the proportion of SNGs and remnant habitats in the surrounding landscape (Table 2). Similarly, the average grassland specialist frequency was positively affected by specialist frequency in the surrounding landscape, as well as by the grassland area and proportion of SNGs and remnant habitats in the surrounding landscape (Table 2). The proportion of shared specialist species between focal grassland and surrounding landscape, was significantly affected by grassland area (Table 2). Moreover, the proportion of shared specialists was also positively related to the interaction between time since restoration and proportion of SNGs in the surrounding landscape (GLM, marginal effect of each variable analysed in Anova type II-test on the full model, Chi2 = 3.7, p = value 0.05), where time since restoration had a positive effect in landscapes with high proportion of semi-natural

Table 2 Relationship between different biodiversity metrics measured in restored (Rest., n = 12) and reference (Ref., i.e. Continuously managed, n = 8) focal semi-natural grasslands in southeastern Sweden and different temporal and landscape variables. Diversity metrics: Number (max. 30) and Frequency (average for 30 species per grassland) of grassland specialist species found in the focal (restored and reference) grasslands, and Proportion of shared specialist species between focal grassland and its surrounding landscape (i.e. grassland specialists occurring in both the focal grasslands and its corresponding landscape, circular radius 1 km). Temporal and landscape variables: Time since restoration (in years), Abandonment time (years without management), Focal grassland area, Proportion of managed semi-natural grasslands (SNGs), Proportion of SNGs and remnant habitats (i.e. former SNGs managed in 1950ies and midfield islets) and Proportion of arable fields in the surrounding landscape, and Number of grassland specialist species (max. 30) in surrounding landscape (balanced sample by using rarefaction and calculating average of 999 permutations), Frequency of grassland specialists in surrounding landscape (balanced sample by rarefaction). Analysed in Spearman Monte Carlo correlation tests. Z-value and p-value (inside parenthesis). Time since restoration and Abandonment time correlated (Pearson correlation, r = −0.64, p = 0.03). Significant results (p < 0.05) indicated by bold letters. Predictor variable

No. spec (rest.)

No. spec (Rest. + Ref.)

Freq. spec (rest.)

Freq. spec (rest. + ref.)

Prop. Shared spec (rest.)

Prop. shared spec (rest. + ref.)

Time since restoration Abandonment time Area Prop. SNGs landscape (1 km) Prop. SNGs + remn hab landscape (1 km) Prop. arable fields landscape (1 km) No. specialists in landscape Freq. specialists in landscape

2.62 (0.003) − 2.32 (0.02) 1.78 (0.08) 0.70 (0.51) 1.53 (0.13)

– – 2.40 (0.01) 1.25 (0.22) 2.05 (0.04)

1.28 (0.21) −1.19 (0.24) 1.69 (0.09) 1.07 (0.30) 2.22 (0.02)

– – 2.17 (0.03) 1.41 (0.16) 2.77 (0.003)

0.97 (0.34) − 1.73 (0.08) 2.11 (0.03) 0.06 (0.97) 0.30 (0.78)

– – 3.11 (< 0.001) 1.17 (0.26) 1.52 (0.13)

− 0.46 (0.67) 2.23 (0.03) –

−0.46 (0.65) 2.33 (0.02) –

0.15 (0.89) – 1.00 (0.32)

−0.73 (0.47) – 1.86 (0.05)

− 0.51 (0.64) – –

− 0.85 (0.41) – –

180

Biological Conservation 214 (2017) 176–183

E. Waldén et al.

Fig. 4. Relationship between the number (to the left) and average frequency (to the right) of grassland specialist plant species found in focal semi-natural grasslands (restored and reference) and the number (left) and average frequency (right) of specialists found in the surrounding landscape (1 km radius, balanced sample by rarefaction) in south-central Sweden (Spearman Monte Carlo tests; z = 2.33 and p = 0.02 (No. species) and z = 1.86, p = 0.05 (Specialist frequency, log-transformed prior analysis)).

(Auffret et al., 2017; Hooftman, Edwards, and Bullock, 2015; Kormann et al., 2015). Here, we found that the proportion of shared species was positively related to time since restoration in landscapes with high proportions of SNGs, suggesting a combined effect of temporal and spatial factors. Grassland specialists occurring in landscapes with high cover of SNGs consequently have a higher probability to repopulate restored habitats when given time. A similar pattern has been observed for recreated grasslands on ex-arable fields in the Czech Republic, where the establishment of target species was significantly related to their occurrence in the surrounding landscape and time since restoration (Prach et al., 2015). Thus, recently restored grasslands in landscapes with a high proportion of SNGs may exhibit a “colonisation credit” (Cristofoli, Piqueray, Dufrêne, Bizoux, and Mahy, 2010; Gijbels et al., 2012), in comparison to older restored grasslands, due to a time lag for specialist species dispersal. In line with other studies (e.g. Cousins, 2006; Cousins and Eriksson, 2001), we found that remnant grassland habitats, such as former grasslands and midfield islets, also harbour grassland specialist species. In the landscapes containing no other semi-natural grasslands than the focal grassland, remnant grassland habitats might be important sources for specialist species recolonising restored grasslands (Cousins and Aggemyr, 2008). This could explain why the proportion of both managed and remnant habitats in the landscapes affected grassland specialist richness and frequency positively, whereas the proportion of solely SNGs in the landscape did not. Although the SNGs overall had a higher number and frequency of specialists, it should be noted that the plot distribution in different habitat types naturally differed between the landscapes, whereas any conclusions drawn from this must consider the possible effect of differences in overall landscape composition. Nevertheless, some of the grassland specialists were also found in the remnant habitats. Many of these species are known to persist in unfavourable conditions, e.g. Primula veris, even for decades (Lehtilä et al., 2016). Two species (Carex pallescens and Lotus corniculatus) occurring in almost all landscapes were rarely found in their associated restored sites. Other studies have also detected this dark diversity pattern (Piqueray et al., 2011), where some specialist species have difficulties recolonising restored grasslands, despite close source populations. Species belonging to the dark diversity often display low dispersal ability and/or low stress-tolerance (Riibak et al., 2015). Here, we found no difference regarding dispersal or persistence traits between specialist species more often found uniquely in surrounding landscape, and species often occurring in both the restored grassland and its corresponding landscape. The selected species in this study are however selected as grassland specialists and are thus more likely have similar sets of traits, than if compared to ruderal and/or generalist species (Conradi and Kollmann, 2016), which is often the case in other studies (e.g. Marteinsdottir and Eriksson, 2014; Riibak et al., 2015). An alternative explanation to that some specialists were absent in the restored

Fig. 5. Relationship between the proportion of grassland specialist plant species shared between restored semi-natural grasslands and surrounding landscapes in south-central Sweden (i.e. occurred in both the restored grassland and its corresponding landscape) and the time since restoration (in years). Proportion of semi-natural grasslands (SNG) in surrounding landscapes (circular radius 1 km) is indicated by black (high prop., > 0.08) and white (low, < 0.05) circles. The proportion of shared specialists was positively related to the interaction between restoration age and proportion of SNGs in the surrounding landscape (GLM, marginal effect of each variable analysed in Anova type II-test on the full model, Chi2 = 3.7, p = value 0.05).

grassland specialists to recolonise restored grasslands from the local seed bank and/or landscape species pool (c.f. Winsa et al., 2015). For restoration purposes, it is also important to know if the desired species are able to disperse from the surrounding habitats into the restored site. We found a positive relationship between the number of grassland specialists occurring in restored grasslands and in the surrounding landscape, suggesting that surrounding grassland specialist species pool could be important for the restoration outcome. The frequency of specialist plant species in the landscape also had a positive effect on frequency in the focal grasslands. However, the relation between grassland specialists in the surrounding landscape and the number of specialists in restored grasslands, does not reveal if it is the same species or not. Nevertheless, we did find that on average 27% of the specialists occurring in the species pool were shared between the restored grassland and its corresponding landscape. Similar proportion of shared species was also found for the reference grasslands, implying dispersal limitations for some of the studied species, rather than unsuitable local conditions. These limitations might be caused by the species' dispersal traits (Marteinsdottir and Eriksson, 2014), but could also depend on landscape characteristics, such as grassland connectivity 181

Biological Conservation 214 (2017) 176–183

E. Waldén et al.

Bossuyt, B., Honnay, O., Van Stichelen, K., Hermy, M., Van Assche, J., 2001. The effect of a complex land use history on the restoration possibilities of heathland in central Belgium. Belgian J. Bot. 134, 29–40. Brückmann, S.V., Krauss, J., Steffan-Dewenter, I., 2010. Butterfly and plant specialists suffer from reduced connectivity in fragmented landscapes. J. Appl. Ecol. 47, 799–809. http://dx.doi.org/10.1111/j.1365-2664.2010.01828.x. CBD, 2012. Strategic Plan for Biodiversity 2011–2020 and the Aichi Targets. Conradi, T., Kollmann, J., 2016. Species pools and environmental sorting control different aspects of plant diversity and functional trait composition in recovering grasslands. J. Ecol. 104, 1314–1325. http://dx.doi.org/10.1111/1365-2745.12617. Cousins, S.A.O., 2006. Plant species richness in midfield islets and road verges - the effect of landscape fragmentation. Biol. Conserv. 127, 500–509. http://dx.doi.org/10. 1016/j.biocon.2005.09.009. Cousins, S.A.O., Aggemyr, E., 2008. The influence of field shape, area and surrounding landscape and plant species richness in grazed ex-fields. Biol. Conserv. 141, 126–135. http://dx.doi.org/10.1016/j.biocon.2007.09.004. Cousins, S.A.O., Eriksson, O., 2001. Plant species occurrences in a rural hemiboreal landscape: effects of remnant habitats, site history, topography and soil. Ecography (Cop.). 24, 461–469. http://dx.doi.org/10.1034/j.1600-0587.2001.d01-202.x. Cousins, S.A.O., Lindborg, R., 2008. Remnant grassland habitats as source communities for plant diversification in agricultural landscapes. Biol. Conserv. 141, 233–240. http://dx.doi.org/10.1016/j.biocon.2007.09.016. Cousins, S.A.O., Auffret, A.G., Lindgren, J., Tränk, L., 2015. Regional-scale land-cover change during the 20th century and its consequences for biodiversity. Ambio 44 (Suppl. 1), S17–27. http://dx.doi.org/10.1007/s13280-014-0585-9. Cristofoli, S., Piqueray, J., Dufrêne, M., Bizoux, J.-P., Mahy, G., 2010. Colonization credit in restored wet heathlands. Restor. Ecol. 18, 645–655. http://dx.doi.org/10.1111/j. 1526-100X.2008.00495.x. Deák, B., Valkó, O., Török, P., Kelemen, A., Miglécz, T., Szabó, S., Szabó, G., Tóthmérész, B., 2015. Micro-topographic heterogeneity increases plant diversity in old stages of restored grasslands. Basic Appl. Ecol. 16, 291–299. http://dx.doi.org/10.1016/j. baae.2015.02.008. Dupre, C., Ehrlen, J., 2002. Habitat configuration, species traits and plant distributions. J. Ecol. 90, 796–805. http://dx.doi.org/10.1046/j.1365-2745.2002.00717.x. Dzwonko, Z., Loster, S., 1998. Dynamics of species richness and composition in a limestone grassland restored after tree cutting. J. Veg. Sci. 9, 387–394. http://dx.doi.org/ 10.2307/3237103. Ekstam, U., Forshed, N., 1997. Om hävden upphör - Kärlväxter som indikatorarter i ängsoch hagmarker. Sweden. (in Swedish). Eriksson, O., 1996. Regional dynamics of plants: a review of evidence for remnant, source-sink and metapopulations. Oikos 77, 248–258. http://dx.doi.org/10.2307/ 3546063. Eriksson, O., Cousins, S.A.O., Bruun, H.H., 2002. Land-use history and fragmentation of traditionally managed grasslands in Scandinavia. J. Veg. Sci. 13, 743–748. http://dx. doi.org/10.1111/j.1654-1103.2002.tb02102.x. Fagan, K.C., Pywell, R.F., Bullock, J.M., Marrs, R.H., 2008. Do restored calcareous grasslands on former arable fields resemble ancient targets? The effect of time, methods and environment on outcomes. J. Appl. Ecol. 45, 1293–1303. http://dx.doi. org/10.1111/j.1365-2664.2008.01492.x. Fox, J., Weisberg, S., 2011. An {R} Companion to Applied Regression, Second. Ed. Sage, Thousand Oaks {CA}. Gijbels, P., Adriaens, D., Honnay, O., 2012. An orchid colonization credit in restored calcareous grasslands. Ecoscience 19, 21–28. http://dx.doi.org/10.2980/19-1-3460. Habel, J.C., Dengler, J., Janišová, M., Török, P., Wellstein, C., Wiezik, M., 2013. European grassland ecosystems: threatened hotspots of biodiversity. Biodivers. Conserv. 22, 2131–2138. http://dx.doi.org/10.1007/s10531-013-0537-x. Hanski, I., 1999. Habitat connectivity, habitat continuity, and metapopulations in dynamic landscapes. Oikos 87, 209–219. http://dx.doi.org/10.2307/3546736. Havrdová, A., Douda, J., Doudová, J., 2015. Local topography affects seed bank successional patterns in alluvial meadows. Flora - Morphol. Distrib. Funct. Ecol. Plants 217, 155–163. http://dx.doi.org/10.1016/j.flora.2015.10.007. Helm, A., Hanski, I., Partel, M., 2006. Slow response of plant species richness to habitat loss and fragmentation. Ecol. Lett. 9, 72–77. http://dx.doi.org/10.1111/j.1461-0248. 2005.00841.x. Helm, A., Zobel, M., Moles, A.T., Szava-Kovats, R., Pärtel, M., 2015. Characteristic and derived diversity: implementing the species pool concept to quantify conservation condition of habitats. Divers. Distrib. 21, 711–721. http://dx.doi.org/10.1111/ddi. 12285. Helsen, K., Hermy, M., Honnay, O., 2013. Spatial isolation slows down directional plant functional group assembly in restored semi-natural grasslands. J. Appl. Ecol. 50, 404–413. http://dx.doi.org/10.1111/1365-2664.12037. Helsen, K., Hermy, M., Honnay, O., 2016. A test of priority effect persistence in seminatural grasslands through the removal of plant functional groups during community assembly. BMC Ecol. 16, 22. http://dx.doi.org/10.1186/s12898-016-0077-9. Hooftman, D.A.P., Edwards, B., Bullock, J.M., 2015. Reductions in connectivity and habitat quality drive local extinctions in a plant diversity hotspot. Ecography (Cop.). http://dx.doi.org/10.1111/ecog.01503. Horrocks, C.A., Heal, K.V., Harvie, B., Tallowin, J.B., Cardenas, L.M., Dungait, J.A.J., 2016. Can species-rich grasslands be established on former intensively managed arable soils? Agric. Ecosyst. Environ. 217, 59–67. http://dx.doi.org/10.1016/j.agee. 2015.10.015. Hothorn, T., Hornik, K., van de Wiel, M.A., Zeileis, A., 2008. Implementing a class of permutation tests: the {coin} package. J. Stat. Softw. 28, 1–23. Janssens, F., Peeters, A., Tallowin, J.R.B., Bakker, J.P., Bekker, R.M., Fillat, F., Oomes, M.J.M., 1998. Relationship between soil chemical factors and grassland diversity. Plant Soil 202, 69–78. http://dx.doi.org/10.1023/a:1004389614865.

grassland despite presence in surrounding landscape, could be the low population densities of the specific species in the species pool. Although they did occur in the surrounding landscapes, their relatively low abundances could negatively affect their recruitment and dispersal performance (Knight et al., 2005; Soons and Heil, 2002). For rare species colonising from adjacent habitat, very strict habitat requirements not met in restored grasslands could also hinder establishment (Waldén and Lindborg, 2016). 5. Conclusions In planning of grassland restorations, it is important to consider both temporal and spatial processes for improved success. We propose that restoration should prioritise large, newly abandoned grasslands situated in landscapes containing high amounts of semi-natural grasslands. However, in landscapes with few semi-natural grasslands remaining, high amounts of remnant grassland habitats, such as formerly managed grasslands and midfield islets could act as potential source habitats for grassland specialist species. By forming direct connections between existing target habitats, other potential source habitats and restored grasslands (Brückmann, Krauss, and Steffan-Dewenter, 2010; Cousins and Lindborg, 2008; Sojneková and Chytrý, 2015), or increasing functional connectivity between grasslands (Auffret et al., 2017; Auffret and Cousins, 2013), propagule dispersal could be further facilitated. However, restoring a degraded grassland in a homogeneous landscape would probably increase the overall landscape biodiversity more than a restored grassland in a heterogeneous landscape would (cf. Rundlöf and Smith, 2006). Our results suggest that restoring seminatural grasslands in a more intact heterogeneous landscape holding a larger species pool could speed up the restoration process, and as an effect of more available habitats in the landscape ensure long-term population survival. This could have implications for how to prioritise future restorations of natural and semi-natural habitats. Acknowledgements We would like to thank landowners and farmers for letting us work on their land and E. Lindgren and A. Knöppel for assistance with fieldwork. We are also grateful to E. Thorsén, J. Plue and S. Jakobsson for advice on the statistical analyses and three anonymous reviewers who gave valuable comments on a previous version of the manuscript. This study was funded by the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (215-20091105) (FORMAS). Appendix A–D. Supplementary data Supplementary data associated with this article can be found in the online version, at https://doi.org/10.1016/j.biocon.2017.07.037. These data include site descriptions, specialist species occurrences and species traits. References Auffret, A.G., Cousins, S.A.O., 2011. Past and present management influences the seed bank and seed rain in a rural landscape mosaic. J. Appl. Ecol. 48, 1278–1285. http:// dx.doi.org/10.1111/j.1365-2664.2011.02019.x. Auffret, A.G., Cousins, S.A.O., 2013. Grassland connectivity by motor vehicles and grazing livestock. Ecography (Cop.). 36, 1150–1157. http://dx.doi.org/10.1111/j. 1600-0587.2013.00185.x. Auffret, A.G., Rico, Y., Bullock, J.M., Hooftman, D.A.P., Pakeman, R.J., Soons, M.B., Suárez-Esteban, A., Traveset, A., Wagner, H.H., Cousins, S.A.O., 2017. Plant functional connectivity - integrating landscape structure and effective dispersal. J. Ecol. doi. http://dx.doi.org/10.1111/1365-2745.12742. Bagaria, G., Helm, A., Rodà, F., Pino, J., 2015. Assessing coexisting plant extinction debt and colonization credit in a grassland-forest change gradient. Oecologia 179, 823–834. http://dx.doi.org/10.1007/s00442-015-3377-4. Bates, D., Maechler, M., Bolker, B., Walker, S., 2015. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67. http://dx.doi.org/10.18637/jss.v067.i01.

182

Biological Conservation 214 (2017) 176–183

E. Waldén et al.

rising to the challenges of the coming decades. Ecosphere 6, art131. doi:http://dx. doi.org/10.1890/ES15-00121.1. Piqueray, J., Bottin, G., Delescaille, L.M., Bisteau, E., Colinet, G., Mahy, G., 2011. Rapid restoration of a species-rich ecosystem assessed from soil and vegetation indicators: The case of calcareous grasslands restored from forest stands. Ecol. Indic. 11, 724–733. http://dx.doi.org/10.1016/j.ecolind.2010.06.007. Plue, J., Cousins, S.A.O., 2013. Temporal dispersal in fragmented landscapes. Biol. Conserv. 160, 250–262. http://dx.doi.org/10.1016/j.biocon.2013.02.010. Poschlod, P., Kiefer, S., Tränkle, U., Fischer, S., Bonn, S., 1998. Plant species richness in calcareous grasslands as affected by dispersability in space and time. Appl. Veg. Sci. 1, 75–90. http://dx.doi.org/10.2307/1479087. Prach, K., Hobbs, R.J., 2008. Spontaneous succession versus technical reclamation in the restoration of disturbed sites. Restor. Ecol. 16, 363–366. http://dx.doi.org/10.1111/ j.1526-100X.2008.00412.x. Prach, K., Fajmon, K., Jongepierová, I., Řehounková, K., 2015. Landscape context in colonization of restored dry grasslands by target species. Appl. Veg. Sci. 18, 181–189. http://dx.doi.org/10.1111/avsc.12140. Pykälä, J., 2003. Effects of restoration with cattle grazing on plant species composition and richness of semi-natural grasslands. Biodivers. Conserv. 12, 2211–2226. http:// dx.doi.org/10.1023/a:1024558617080. Pywell, R.F., Bullock, J.M., Roy, D.B., Warman, L.I.Z., Walker, K.J., Rothery, P., 2003. Plant traits as predictors of performance in ecological restoration. J. Appl. Ecol. 40, 65–77. http://dx.doi.org/10.1046/j.1365-2664.2003.00762.x. R Core Team, 2016. R: A Language and Environment for Statistical Computing. Riibak, K., Reitalu, T., Tamme, R., Helm, A., Gerhold, P., Znamenskiy, S., Bengtsson, K., Rosén, E., Prentice, H.C., Pärtel, M., 2015. Dark diversity in dry calcareous grasslands is determined by dispersal ability and stress-tolerance. Ecography (Cop.). 38, 713–721. http://dx.doi.org/10.1111/ecog.01312. Rundlöf, M., Smith, H.G., 2006. The effect of organic farming on butterfly diversity depends on landscape context. J. Appl. Ecol. 43, 1121–1127. http://dx.doi.org/10. 1111/j.1365-2664.2006.01233.x. Schrautzer, J., Jansen, D., Breuer, M., Nelle, O., 2009. Succession and Management of Calcareous Dry Grasslands in the Northern Franconian Jura, Germany, in: Papers From the 5th Dry Grassland Meeting 2008. Kiel. Floristisch-Soziologische Arbeitsgemeinschaft e.V.pp. 339–351. Sojneková, M., Chytrý, M., 2015. From arable land to species-rich semi-natural grasslands: succession in abandoned fields in a dry region of central Europe. Ecol. Eng. 77, 373–381. http://dx.doi.org/10.1016/j.ecoleng.2015.01.042. Soons, M.B., Heil, G.W., 2002. Reduced colonization capacity in fragmented populations of wind-dispersed grassland forbs. J. Ecol. 90, 1033–1043. http://dx.doi.org/10. 1046/j.1365-2745.2002.00729.x. Steiner, M., Öckinger, E., Karrer, G., Winsa, M., Jonsell, M., 2016. Restoration of seminatural grasslands, a success for phytophagous beetles (Curculionidae). Biodivers. Conserv. 25, 3005–3022. http://dx.doi.org/10.1007/s10531-016-1217-4. Stoate, C., Baldi, A., Beja, P., Boatman, N.D., Herzon, I., van Doorn, A., de Snoo, G.R., Rakosy, L., Ramwell, C., 2009. Ecological impacts of early 21st century agricultural change in Europe - a review. J. Environ. Manag. 91, 22–46. http://dx.doi.org/10. 1016/j.jenvman.2009.07.005. Tikka, P.M., Heikkila, T., Heiskanen, M., Kuitunen, M., 2001. The role of competition and rarity in the restoration of a dry grassland in Finland. Appl. Veg. Sci. 4, 139–146. Török, P., Vida, E., Deák, B., Lengyel, S., Tóthmérész, B., 2011. Grassland restoration on former croplands in Europe: an assessment of applicability of techniques and costs. Biodivers. Conserv. 20, 2311–2332. http://dx.doi.org/10.1007/s10531-011-9992-4. Waldén, E., Lindborg, R., 2016. Long term positive effect of grassland restoration on plant diversity - success or not? PLoS One 11, e0155836. http://dx.doi.org/10.1371/ journal.pone.0155836. Willems, J.H., 2001. Problems, approaches, and, results in restoration of Dutch calcareous grassland during the last 30 years. Restor. Ecol. 9, 147–154. http://dx.doi.org/10. 1046/j.1526-100x.2001.009002147.x. Willems, J.H., Bik, L.P.M., 1998. Restoration of high species density in calcareous grassland: the role of seed rain and soil seed bank. Appl. Veg. Sci. 1, 91–100. http:// dx.doi.org/10.2307/1479088. Winsa, M., Bommarco, R., Lindborg, R., Marini, L., Öckinger, E., 2015. Recovery of plant diversity in restored semi-natural pastures depends on adjacent land use. Appl. Veg. Sci. 18, 413–422. http://dx.doi.org/10.1111/avsc.12157. Winsa, M., Öckinger, E., Bommarco, R., Lindborg, R., Roberts, S.P.M., Wärnsberg, J., Bartomeus, I., 2017. Sustained functional composition of pollinators in restored pastures despite slow functional restoration of plants. Ecol. Evol. 7, 3836–3846. http://dx.doi.org/10.1002/ece3.2924. Zobel, M., van der Maarel, E., Dupré, C., 1998. Species pool: the concept, its determination and significance for community restoration. Appl. Veg. Sci. 1, 55–66. http:// dx.doi.org/10.2307/1479085.

Kalamees, R., Püssa, K., Zobel, K., Zobel, M., 2012. Restoration potential of the persistent soil seed bank in successional calcareous (alvar) grasslands in Estonia. Appl. Veg. Sci. 15, 208–218. http://dx.doi.org/10.1111/j.1654-109X.2011.01169.x. Kasari, L., Saar, L., de Bello, F., Takkis, K., Helm, A., 2016. Hybrid ecosystems can contribute to local biodiversity conservation. Biodivers. Conserv. 1–19. http://dx.doi. org/10.1007/s10531-016-1218-3. Knight, T.M., Steets, J.A., Vamosi, J.C., Mazer, S.J., Burd, M., Campbell, D.R., Dudash, M.R., Johnston, M.O., Mitchell, R.J., Ashman, T.-L., 2005. Pollen limitation of plant reproduction: pattern and process. Annu. Rev. Ecol. Evol. Syst. 36, 467–497. http:// dx.doi.org/10.1146/annurev.ecolsys.36.102403.115320. Kotiluoto, R., 1998. Vegetation changes in restored semi-natural meadows in the turku Archipelago of SW Finland. Plant Ecol. 136, 53–67. http://dx.doi.org/10.1023/ a:1009781217847. Kormann, U., Rösch, V., Batáry, P., Tscharntke, T., Orci, K.M., Samu, F., Scherber, C., 2015. Local and landscape management drive trait-mediated biodiversity of nine taxa on small grassland fragments. Divers. Distrib. 21, 1204–1217. http://dx.doi.org/10. 1111/ddi.12324. Krauss, J., Klein, A.-M., Steffan-Dewenter, I., Tscharntke, T., 2004. Effects of habitat area, isolation, and landscape diversity on plant species richness of calcareous grasslands. Biodivers. Conserv. 13, 1427–1439. http://dx.doi.org/10.1023/B:BIOC.0000021323. 18165.58. Krauss, J., Bommarco, R., Guardiola, M., Heikkinen, R.K., Helm, A., Kuussaari, M., Lindborg, R., Öckinger, E., Partel, M., Pino, J., Poyry, J., Raatikainen, K.M., Sang, A., Stefanescu, C., Teder, T., Zobel, M., Steffan-Dewenter, I., 2010. Habitat fragmentation causes immediate and time-delayed biodiversity loss at different trophic levels. Ecol. Lett. 13, 597–605. http://dx.doi.org/10.1111/j.1461-0248.2010.01457.x. Kuussaari, M., Bommarco, R., Heikkinen, R.K., Helm, A., Krauss, J., Lindborg, R., Öckinger, E., Partel, M., Pino, J., Roda, F., Stefanescu, C., Teder, T., Zobel, M., Steffan-Dewenter, I., 2009. Extinction debt: a challenge for biodiversity conservation. Trends Ecol. Evol. 24, 564–571. Lehtilä, K., Dahlgren, J.P., Garcia, M.B., Leimu, R., Syrjänen, K., Ehrlén, J., 2016. Forest succession and population viability of grassland plants: long repayment of extinction debt in Primula veris. Oecologia 181, 125–135. http://dx.doi.org/10.1007/s00442016-3569-6. Lewis, R.J., Szava-Kovats, R., Pärtel, M., 2016. Estimating dark diversity and species pools: an empirical assessment of two methods. Methods Ecol. Evol. 7, 104–113. http://dx.doi.org/10.1111/2041-210X.12443. Lindborg, R., 2007. Evaluating the distribution of plant life-history traits in relation to current and historical landscape configurations. J. Ecol. 95, 555–564. http://dx.doi. org/10.1111/j.1365-2745.2007.01232.x. Lindborg, R., Eriksson, O., 2004. Effects of restoration on plant species richness and composition in Scandinavian semi-natural grasslands. Restor. Ecol. 12, 318–326. http://dx.doi.org/10.1111/j.1061-2971.2004.00334.x. Lindborg, R., Helm, A., Bommarco, R., Heikkinen, R.K., Kühn, I., Pykälä, J., Pärtel, M., 2012. Effect of habitat area and isolation on plant trait distribution in European forests and grasslands. Ecography (Cop.). 35, 356–363. http://dx.doi.org/10.1111/j. 1600-0587.2011.07286.x. Lindborg, R., Plue, J., Andersson, K., Cousins, S.A.O., 2014. Function of small habitat elements for enhancing plant diversity in different agricultural landscapes. Biol. Conserv. 169, 206–213. http://dx.doi.org/10.1016/j.biocon.2013.11.015. Lindgren, J.P., Cousins, S.A.O., 2017. Island biogeography theory outweighs habitat amount hypothesis in predicting plant species richness in small grassland remnants. Landsc. Ecol. 1–12. http://dx.doi.org/10.1007/s10980-017-0544-5. Marteinsdottir, B., Eriksson, O., 2014. Trait-based filtering from the regional species pool into local grassland communities. J. Plant Ecol. 7, 347–355. http://dx.doi.org/10. 1093/jpe/rtt032. Öckinger, E., Smith, H.G., 2006. Landscape composition and habitat area affects butterfly species richness in semi-natural grasslands. Oecologia 149, 526–534. http://dx.doi. org/10.1007/s00442-006-0464-6. Öckinger, E., Lindborg, R., Sjödin, N.E., Bommarco, R., 2012. Landscape matrix modifies richness of plants and insects in grassland fragments. Ecography (Cop.). 35, 259–267. http://dx.doi.org/10.1111/j.1600-0587.2011.06870.x. Papanikolaou, A.D., Kühn, I., Frenzel, M., Schweiger, O., 2016. Semi-natural habitats mitigate the effects of temperature rise on wild bees. J. Appl. Ecol. http://dx.doi.org/ 10.1111/1365-2664.12763. Pärtel, M., Szava-Kovats, R., Zobel, M., 2011. Dark diversity: shedding light on absent species. Trends Ecol. Evol. 26, 124–128. http://dx.doi.org/10.1016/j.tree.2010.12. 004. Pärtel, M., Bennett, J.A., Zobel, M., 2016. Macroecology of biodiversity: disentangling local and regional effects. New Phytol. 211, 404–410. http://dx.doi.org/10.1111/ NPH.13943. Perring, M.P., Standish, R.J., Price, J.N., Craig, M.D., Erickson, T.E., Ruthrof, K.X., Whiteley, A.S., Valentine, L.E., Hobbs, R.J., 2015. Advances in restoration ecology:

183