Agriculture, Ecosystems and Environment 251 (2018) 88–98

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment

journal homepage: www.elsevier.com/locate/agee

Research Paper

Smart management is key for successful diversification of field margins in MARK

highly productive farmland
⁎
A. Kirmera, , K. Rydgrenb, S. Tischewa
a
Anhalt University of Applied Sciences, Department for Nature Conservation and Landscape Planning, Strenzfelder Allee 28, D-06406 Bernburg, Germany
b
Western Norway University of Applied Sciences, Institute of Natural Science, P.O. Box. 133, NO-6851 Sogndal, Norway

A R T I C L E I N F O A B S T R A C T

Keywords: In highly productive farmlands, field margins could offer habitats and refuges for many plant and animal species.
Field margin restoration But species-rich field margins are rapidly declining due to field enlargements and unfavourable management
Sward disturbance intensity practices. In European farmland, management is usually restricted to repeated mulching during the growing
High-diversity seed mixture season or mulching once a year between September and February. Under nutrient-rich conditions, both methods
Wild plants
benefit competitive perennial grasses at the expense of species richness. Diversification of species-poor grass
Sowing
margins on nutrient-rich sites is difficult to achieve and we lack evidence which method works best. Starting in
Management
Biodiversity late summer 2010, we implemented a large-scale field experiment on nutrient-rich Chernozem soil, where we
examined the effect of sward disturbance intensity, sowing of target species, and mowing time on the devel-
opment of field margin vegetation over seven years. After disturbance of the existing species-poor grass sward
with two intensities (tilling once or three times), a seed mixture of 49 wild plants from a regional seed propa-
gation was sown in early October 2010. After an establishment phase in 2011, the sites were mown once a year,
in either June or September, with removal of biomass. In addition, both cutting times were applied to species-
poor grass margins without disturbance and sowing treatments. We recorded the plant species composition
yearly from 2010 until 2016. Although the early establishment rate of the sown species was higher on sites
disturbed three times, the number of successfully established target species on sites with different sward dis-
turbance intensities converged during the observation period. Mowing in September resulted in higher grass
cover and considerably decreased the cover of the sown target species. On the other hand, mowing in June
resulted in significantly higher plot occupancy and cover of the sown target species. In general, the immigration
success of target forbs into adjacent undisturbed and unsown grass margins was very low even after seven years,
although mowing once a year with biomass removal increased the number of mostly ruderal species.
Diversification of grass margins was very successful with active species introduction in combination with initial
sward disturbance and management adapted to nutrient-rich site conditions. Therefore, restored field margins in
highly productive farmlands should be mown in early summer to sustain long-term biodiversity.

1. Introduction
For more than 60 years, agricultural land in Europe has experienced
a severe biodiversity decline (Baessler and Klotz, 2006; Stoate et al.,
2009; Meyer et al., 2015; Sutcliffe et al., 2015). Agricultural in-
tensification, field enlargement and mechanization (Tscharntke et al.,
2005; White and Roy, 2015), but also changes in management practices
(Alignier and Baudry, 2015), have led to a loss of semi-natural habitats,
a general decrease of plant and animal species richness, and depletion
of soil seed banks (Robinson and Sutherland, 2002; Simmering et al.,
2013; Meyer et al., 2015). Field margins can be both habitat and refuge
for many plant and animal species (Marshall and Moonen, 2002;
Tscharntke et al., 2005; Hackett and Lawrence, 2014). But, especially in
intensively used agrarian landscapes, repeated mulching or spraying
reduces plant species diversity (Marshall and Moonen, 2002; Hahn
et al., 2014), often resulting in species-poor, grass-dominated sites
(Meyer et al., 2013), which are unable to fulfill important ecosystem
services (e.g. habitat and food sources for pollinators). Contrary to
perennial wildflower strips, which are created on arable land and are
usually supported for five years by agri-environmental schemes, per-
manent field margins are situated on common land between farm tracks
and arable fields. Although they are in an unfavourable conservation
status in most European countries, only a few countries developed agri-
environmental schemes to promote field margin biodiversity (e.g. Great

⁎
Corresponding author.
E-mail address: anita.kirmer@hs-anhalt.de (A. Kirmer).

http://dx.doi.org/10.1016/j.agee.2017.09.028
Received 11 May 2017; Received in revised form 20 September 2017; Accepted 25 September 2017
0167-8809/ © 2017 Elsevier B.V. All rights reserved.
A. Kirmer et al.
Britain, Switzerland). In Germany, as well as in 15 other EU-28 coun-
tries, field margins are eligible as Ecological Focus Areas (EFA; Hart,
2015). Even though the Common Agricultural Policy (CAP, 2014–2020)
requires that 5% of arable land be designated as EFA, most farmers
favor fallows or cultivation of nitrogen-fixing crops instead of main-
taining landscape elements like field margins (Pe’er et al., 2016).
In intensively managed agricultural landscapes, relying on sponta-
neous colonization of target plants is usually not successful (Kleijn
et al., 1998; Török et al., 2011; Prach et al., 2015) due to depleted seed
banks and the limited long-distance dispersal ability of many plant
species (Bakker and Berendse, 1999; Nathan et al., 2008). Hence, active
introduction of target species in restoration sites is recommended (Kiehl
et al., 2010; Rydgren et al., 2010). Although grassland diversification
experiments showed that the establishment success of target species is
dependent on initial sward disturbance (Hofmann and Isselstein, 2004;
Edwards et al., 2007; Pywell et al., 2007; Schmiede et al., 2012), there
are still uncertainties about sward disturbance intensity.
After successful species re-introduction, implementing a suitable
management regime is crucial for the maintenance of species diversity
(Pywell et al., 2007; Öster et al., 2009; Auestad et al., 2016; John et al.,
2016). In many European countries, field margins are usually mown
once a year between September and February (Kleijn et al., 1998:
France, Netherlands, UK; Bokenstrand et al., 2004: Sweden) or not at all
(Hovd and Skogen, 2005: Norway). This might work well on marginal
sites, but under nutrient-rich conditions it benefits competitive per-
ennial grasses (Hansson and Fogelfors, 1998; Amiaud et al., 2008). In
addition, late mowing destroys hibernating structures for invertebrates
(Blake et al., 2013) and removes winter feeding sources for farmland
birds (Vickery et al., 2009). By contrast, many forbs are able to re-
generate quickly after early mowing, thus prolonging the flowering
period until early autumn.
While many studies document the early establishment phase after
species introduction into field margins (de Cauwer et al., 2005; Carvell
et al., 2007; Blake et al., 2013), long-term studies are scarce
(Bokenstrand et al., 2004; Smith et al., 2010). Successful examples of
sustainable conversions of species-poor grass strips into flower-rich,
highly-diverse field margins are completely missing. Even though grass
strips are beneficial for some arthropod groups (Meek et al., 2002;
Badenhausser and Cordeau, 2012), they meet neither the demands of
nectar and pollen-feeding invertebrates nor those of farmland birds
(Vickery et al., 2009). Converting species-poor grass margins into
structurally and floristically diverse vegetation during summer and
seed-rich habitats in winter are optimal prerequisites to provide
feeding, nesting, and hibernating habitats for many animal species.
In general, problems with sowing high-diversity seed mixtures of
regional wild plants include their rather high cost and difficulty in
obtaining regionally produced seeds (Kiehl et al., 2010; Tischew et al.,
2011). A cost efficient option is to introduce target species only in parts
of restoration sites (Rayburn and Laca, 2013; Valkó et al., 2016).
Combined with appropriate management, this might promote coloni-
zation of adjacent areas, but little is known about the immigration
success of target species into adjacent species-poor grass margins (but
see Smith et al., 2010).
In a large-scale field experiment in Germany, we examined the ef-
fect of sward disturbance intensity, mowing time and sowing on the
development of field margin vegetation over seven years. We focused
on the following questions: (Q1) Is sowing of a species-rich seed mix-
ture a successful method to diversify species-poor grass margins? (Q2)
Can adjacent target forbs establish in regularly mown, but unsown grass
margins, and does mowing time influence their establishment success?
(Q3) Does sward disturbance intensity influence the establishment of
sown target species? (Q4) Which mowing time (June or September) is
most successful in maintaining species-richness of restored field mar-
gins?
89
Agriculture, Ecosystems and Environment 251 (2018) 88–98
2. Material and methods
2.1. Study area
Our study area is situated on the border of the Strenzfeld campus of
the Anhalt University of Applied Sciences in Saxony-Anhalt (N
51°49′01.00″, E 11°42′14.09″, 90–93 m above sea level). The climate is
dry with 511 mm annual precipitation and a mean annual temperature
of 9.7 °C (long-term mean 1981–2010, Deutscher Wetterdienst).
Adjacent arable fields are used for conventional wheat, rape or maize
production. The soil is a very productive Chernozem developed on loess
substrate. Before the experiment started, the field margins were char-
acterized by thick litter layers and a few competitive grasses, such as
Dactylis glomerata, Elymus repens, and Poa angustifolia, as a result of
repeated mulching.
2.2. Experimental design and sampling
In August 2010, we established 30 blocks (3 m × 17 m) 1 m apart
along a 540 m long field margin. A 1 m × 8 m permanent plot was
positioned in the centre of each block. Species abundance, measured as
percentage cover, was recorded in all permanent plots to document the
initial conditions before site preparation started (Tables 1 and 2), then
the whole site was mown with removal of biomass and litter. Soil
samples were taken at the plot level, to a depth of 0–15 cm (mixing of
12 soil core samples per plot), in August 2010. Overall, the soil had a
pH of 7.4 ( ± 0.05) (after German DIN 19684-1) and contained 0.25%
( ± 0.03) total nitrogen (after German DIN ISO 13878), 106 mg kg−1
soil ( ± 16.8) CAL exchangeable phosphorous and 318 mg kg−1 soil
( ± 75.2) exchangeable potassium (both after VDLUFA A 6 2.1.1.), for
further details see Kiehl et al. (2014).
The total 540 m field margin was divided into five equal sections of
108 m, each of which contained six blocks, to which the following six
treatments were randomly assigned: (T1) sward disturbance once be-
fore sowing, with mowing in mid-June, (T2) sward disturbance once
before sowing, with mowing in mid-September, (T3) sward disturbance
three times before sowing, with mowing in mid-June, (T4) sward dis-
turbance three times before sowing, with mowing in mid-September;
(T5) without sward disturbance or sowing, and with mowing in mid-
June, and (T6) without sward disturbance or sowing, and with mowing
in mid-September.
We used a rototiller to loosen the compacted soil surface (only the
first time), immediately followed by a grubber to pull out grass rhi-
zomes. In 10 blocks, the grass sward was disturbed three times at c. 14-
day intervals starting on September 9th, 2010. The last intervention
took place on October 6th, 2010, at which time sward disturbance
occurred in the 10 blocks that were disturbed only once. The remaining
10 blocks were left undisturbed and unsown. During the observation
time, the 1 m space between blocks was regularly disturbed by roto-
tiller. In the year after sowing, all blocks were mown at the beginning of
June as well as at the end of August to reduce arable weeds emerging
from the soil seed bank. Regular management (mowing either mid-June
or mid-September with removal of biomass) started in 2012.
For the seed mixture, we selected 49 species from 15 plant families,
belonging to different plant communities: dry grasslands (16 species),
mesic grasslands (12 species), fringes (8 species), and dry ruderal
communities (13 species) (Table 1). These species have a long blooming
period and are moderately competitive under nutrient-rich conditions
(no highly competitive ruderals, tall grasses or weeds). On October 7th,
2010, a seed mixture of 5 grasses, 5 legumes, and 39 nonlegume forbs
from a regional agricultural propagation of wild plants was sown by
hand with a density of 20 kg seeds ha−2, corresponding to 1917 seed
per m−1, calculated due to the thousand grain weight of the selected
species. Seed number for each species (see Table 1, column SN) was
based on ecological (seed size, seed longevity, germination rate) and
economical (costs) requirements. After sowing, the soil surface was
A. Kirmer et al. Agriculture, Ecosystems and Environment 251 (2018) 88–98

Table 1
Mean percentage cover of layers and sown species in 2010 and 2016 on permanent plots (1 m × 8 m, n = 5). Mean cover of species ≥ 4% is marked in bold. SoG (phyto-sociological
groups): DG = dry grassland, MG = mesic grassland, FG = fringe communities, RD = dry ruderal communities. SpG (species groups): OSS = only sown species, SRS = sown resident
species, FG (functional groups): G = grasses, L = legumes, F = nonlegume forbs. SN (seed number in mixture): seed per m-1. Mix: x = recommended for seed mixture, (x) = only for
drier sites, x! = highly competitive, x!! = small distribution areal. Sown species not present on permanent plots in 2016: Campanula patula L., MG, SN = 100, (x); Campanula glomerata L.,
DG, SN = 100; Eryngium campestre L., DG, SN = 10; Cynoglossum officinale L., RD, SN = 10; Melica transsilvanica Schur, SN = 30; Reseda luteola L., RD, SN = 10.

Treatment T5 T1 T3 T6 T2 T4 T5 T1 T3 T6 T2 T4

Observation year 2010 2010 2010 2010 2010 2010 2016 2016 2016 2016 2016 2016
Management starting in 2012: Jun = June-mown, Sep = September-mown Jun Jun Jun Sep Sep Sep Jun Jun Jun Sep Sep Sep
Sward disturbance intensity before sowing: 0 = none,1 = once,3 = three times 0 1 3 0 1 3 0 1 3 0 1 3
Sowing: 0 = no, 1 = yes (begin of October 2010) 0 0 0 0 0 0 0 1 1 0 1 1
cover grasses (%) 45.2 52.6 47.0 46.3 48.3 43.1 52.6 21.0 18.0 54.4 41.0 51.0
cover legumes (%) 0.0 0.0 0.0 0.0 0.0 0.0 0.2 3.0 2.6 0.2 2.1 2.2
cover nonlegume forbs (%) 0.8 0.4 1.0 1.7 0.7 0.9 12.2 56.0 51.4 9.4 26.9 23.8
cover total herb layer (%) 46.0 53.0 48.0 48.0 49.0 44.0 65.0 80.0 74.0 64.0 69.0 76.0
cover mosses/liches (%) 0.1 0.0 0.1 0.2 0.1 0.0 4.8 2.7 2.6 4.6 3.2 1.8
cover litter layer (%) 88.0 86.0 87.0 88.0 85.0 84.0 38.0 12.6 11.4 25.0 27.0 21.0
cover bare soil (%) 1.0 1.5 1.1 1.0 1.2 1.6 9.0 6.0 6.8 11.2 8.0 5.8

SoG SpG FG SN Mix Sown species

MG SRS F 30 x Achillea millefolium L. 0.2 0.1 0.6 1.2 0.3 0.6 1.8 1.2 1.2 0.1 0.2 0.2
MG SRS F 50 x Daucus carota L. 0.02 0.02 0.02 0.02 0.1 0.4 0.3 0.1 0.3 0.6
MG SRS F 50 x Galium album Mill. 0.2 0.1 0.1 0.1 4.5 4.1 4.1 4.6 8.6 8.5
MG SRS F 20 x Pastinaca sativa L. 0.02 0.02 0.3 0.04 1.5 0.9 0.04 0.6 1.0
MG SRS L 10 x! Trifolium pratense L. 0.02 1.5 0.7 0.2 0.02
MG OSS F 50 x Leucanthemum ircutianum DC. 0.5 23.4 17.0 0.02 1.0 1.4
MG OSS F 20 x Knautia arvensis (L.) J.M. Coult. s. str. 0.1 7.0 13.6 0.8 1.6
MG OSS F 30 x Prunella vulgaris L. 0.04 2.4 1.8 0.3 0.1
MG OSS F 20 x Centaurea jacea L. s. str. 1.8 2.1 1.8 0.8
MG OSS F 10 x Crepis biennis L. 0.1 0.1 0.02 0.1
MG OSS L 30 x Lotus corniculatus L. 0.1 0.2 0.4 0.3

DG OSS G 200 x Festuca rupicola Heuff. 5.6 8.0 0.2 12.6 18.6
DG OSS G 200 x Briza media L. 0.5 1.9 0.04 0.2
DG OSS F 10 x Centaurea scabiosa L. 0.7 0.7 1.0 0.5
DG OSS G 200 Festuca ovina agg. p.p. 0.4 0.5 0.3 0.2
DG OSS F 50 x Galium verum L. s. str. 0.3 0.7 0.1 1.1
DG OSS L 20 x Medicago falcata L. 1.3 0.9 0.1 0.7
DG OSS F 30 (x) Pimpinella saxifraga L. 0.1 0.04 0.02 0.02
DG OSS F 20 x Sanguisorba minor Scop. 0.7 0.6 1.1 0.7
DG OSS F 30 (x) Salvia pratensis L. 0.2 0.2 0.02
DG OSS F 15 x Scabiosa ochroleuca L. 0.4 0.2 0.1
DG OSS F 30 (x) Plantago media L. 0.3 0.2 0.1
DG OSS F 30 (x) Filipendula vulgaris Moench 0.2 0.04
DG OSS F 20 (x) Leontodon hispidus L. 0.1 0.02
DG OSS F 30 (x) Dianthus carthusianorum L. 0.1

FC OSS F 10 x Agrimonia eupatoria L. 0.2 0.7 0.4 0.8 1.0

FC OSS F 30 x Hypericum perforatum L. 0.2 0.7 0.2 0.04 0.2
FC OSS F 20 x Betonica officinalis L. 0.8 0.9 0.4 0.2
FC OSS F 50 x Clinopodium vulgare L. 0.7 1.0 0.7 0.5
FC OSS F 50 x Origanum vulgare L. 2.5 1.8 0.6 0.5
FC OSS F 20 x Silene vulgaris (Moench) Garcke 0.1 0.5 0.3 0.3
FC OSS F 20 x Stachys recta L. 0.9 1.2 0.7 0.3
FC OSS L 1 x! Securigera varia (L.) Lassen 0.2 0.1 0.2

RD SRS G 120 x Poa angustifolia L. 20.4 25.0 28.2 29.0 14.4 18.7 11.6 9.0 4.4 20.0 10.6 11.0
RD OSS F 10 x Silene latifolia ssp. alba (Mill.) Greuter et Burdet 0.9 0.5 0.3 0.3 0.6 1.9
RD OSS F 10 x Falcaria vulgaris Bernh. 0.8 0.1 0.1 0.1 0.6 0.1
RD OSS F 20 x Anthemis tinctoria L. 0.1 0.1 0.1 0.5 0.2
RD OSS F 10 x Malva moschata L. 0.2 2.6 1.1 1.7 1.9
RD OSS F 20 x Cichorium intybus L. 0.8 0.1 0.2 0.4
RD OSS F 20 x!! Lavatera thuringiaca L. 0.6 1.0 0.3 0.2
RD OSS F 30 x Linaria vulgaris Mill. 0.3 0.6 0.2 0.2
RD OSS L 1 Lathyrus tuberosus L. 0.02 0.02
RD OSS F 10 x Malva sylvestris L. 0.1 0.1

rolled to ensure soil contact of the seeds.
We recorded functional group cover and species abundance as
percentage cover (0–100 with 0.01% as the lowest value for occur-
rence) in permanent plots from 2011 to 2016. Cover measurements
began in August 2011, and then between mid-May and the beginning of
June starting in 2012 and for the subsequent years.
2.3. Statistical analyses
For the statistical analyses, we distinguished sown species that were
not present in the grass margins before sowing (OSS = only sown
species). Six species from the seed mixture (Achillea millefolium, Daucus
carota, Galium album, Pastinaca sativa, Poa angustifolia, Trifolium pra-
tense) were analysed separately, because they were already present in
the grass margins before the experiment started (SRS = sown resident
species). Sown resident and only sown species were pooled for the
analysis of plot occupancy of sown species. Cover of grasses (CoGr) was
used to show the development of functional groups for different
mowing times.
Statistical analyses were conducted using R version 3.2.5 (R

90
A. Kirmer et al. Agriculture, Ecosystems and Environment 251 (2018) 88–98

Table 2
Mean percentage cover of resident/spontaneously immigrating species in 2010 and 2016 on permanent plots (1 m × 8 m, n = 5). Mean cover ≥ 4% is marked in bold. SoG (phyto-
sociological groups): MG = mesic grassland, RD = dry ruderal communities; R = nitrophilic ruderal communities, A = arable weed communities. SpG (species groups): RTS = resident/
spontaneous target species, RNT = resident/spontaneous non-target species; FG (functional groups): G = grasses, L = legumes, F = nonlegume forbs. LS (life span): a = annual,
b = biennial, p = perennial. Additional species with very low cover: Glechoma hederacea L. (p, T5: 2010, 2016), Alopecurus pratensis L. (p, T4, 2016), Artemisia absinthium L. (p, T1, 2016).

Treatment T5 T1 T3 T6 T2 T4 T5 T1 T3 T6 T2 T4

Observation year 2010 2010 2010 2010 2010 2010 2016 2016 2016 2016 2016 2016
Management starting in 2012: Jun = June-mown, Sep = September-mown Jun Jun Jun Sep Sep Sep Jun Jun Jun Sep Sep Sep
Sward disturbance intensity before sowing: 0 = none,1 = once,3 = three times 0 1 3 0 1 3 0 1 3 0 1 3
Sowing: 0 = no, 1 = yes (begin of October 2010) 0 0 0 0 0 0 0 1 1 0 1 1

SoG SpGr FG LS Resident/spontaneously immigrating species

MG RTS G p Arrhenatherum elatius (L.) J. Presl et C. Presl 0.6 0.2 0.2 0.2 24.0 2.6 2.5 18.1 10.2 1.2
MG RTS G p Dactylis glomerata L. 1.7 1.4 1.9 0.3 0.7 4.0 9.1 4.4 4.0 9.6 5.0 4.0
MG RTS G p Lolium perenne L. 0.4 0.1 0.4 0.2 0.2
MG RTS G p Festuca rubra L. 0.1 0.7 0.3 0.2 0.04 0.5
MG RTS G p Poa trivialis L. 2.1 1.0 1.6 2.0 0.7 1.0
MG RTS G p Poa pratensis L. 0.02 0.02 2.0 3.0
MG RTS L p Medicago lupulina L. 0.02 0.1 0.3 0.1
MG RTS F p Plantago lanceolata L. 0.1 0.1 0.1
RD RTS F a Papaver rhoeas L. 0.1 4.2 0.3 1.0
RD RTS F b Verbascum lychnitis L. 0.2 0.5
R RNTS G p Elymus repens (L.) Gould 24.2 28.0 18.0 19.0 34.6 21.2 6.6 1.0 1.2 5.6 4.4 11.4
R/MG RNTS F p Taraxacum sect. ruderalia Kirschner et al. 0.2 0.2 0.2 0.1 0.2 0.1 0.3 0.2 0.1 0.2 0.1 0.1
R RNTS F p Rumex crispus L. 0.02 0.02 0.02 0.02 0.02 0.1 0.02 0.1 0.1 0.1
R RNTS F b Cirsium vulgare (Savi) Ten. 0.02 0.02 0.3 0.1 0.02 0.2 0.1
R RNTS F p Cirsium arvense (L.) Scop. 0.02 0.1 0.1 0.1 0.02 0.2 0.04
R RNTS F a Geranium pusillum Burm. f. 0.02 0.02 0.02 1.6 0.3 0.7 0.6 0.1 0.4
R RNTS F b Carduus acanthoides L. 0.1 0.1 0.1 0.5 0.5
A RNTS F a Veronica arvensis L. 0.1 0.1 0.1 0.3 0.1 0.1
A RNTS L a Vicia tetrasperma (L.) Schreb. 0.1 0.1 0.1 0.04 0.1
A RNTS G a Bromus sterilis L. 0.04 0.1 0.8 0.2 2.8
A RNTS L a Vicia sativa L. 0.04 1.0
RD RNTS G a Bromus hordeaceus L. 0.4 0.02 0.1
R RNTS F a Galium aparine L. 0.04 0.2

Development Core Team, 2015).
For answering Q1, we extracted the gradient structure of the species
composition of treatments T1–T6 for all observation years (2010–2016)
by use of detrended correspondence analysis (DCA – Hill and Gauch,
1980) and global non-metric multidimensional scaling (GNMDS –
Minchin, 1987) as implemented in the vegan package version 2.4.0
(Oksanen et al., 2016). Prior to the ordination analyses, we weighted
each matrix element with a power function (Økland, 1986; van Son and
Halvorsen, 2014) obtaining a scalar range up to 100 by using a
weighting parameter w = 0.500. GNMDS was run using function
monoMDS in the vegan package, with the Bray-Curtis dissimilarity
index, maximum number of iterations = 500 and convergence toler-
ance criterion = 10–5. GNMDS solutions were obtained from 2000
different random starting configurations, of which more than two had
to give identical minimum stress solutions. This best solution was
subjected to varimax rotation using Principal Component Analysis
(PCA) and the resulting axes were re-scaled to half-change (H.C.) units.
We solved for two-, three- and four-dimensional GNMDS solutions. We
compared pairs of axes using the Kendall’s rank correlation coefficient
τ. The two first axes of the DCA and the two first axes of the three
GNMDS solutions were closely similar (0.81 ≤ |τ| ≤ 0.86), whereas |τ|
between the DCA axis 3 and GNMDS-axes 3 was 0.39–0.43. The cor-
responding axes of the three GNMDS solutions were closely similar (axis
1: 0.94 ≤ |τ| ≤ 0.97; axis 2: 0.88 ≤ |τ| ≤ 0.95; axis 3: τ = 0.87), and
the axes of the third dimensional GNMDS solution was always most
identical to the axes of the other GNMDS solutions. We therefore used
the 3-dimensional GNMDS solution to present the gradient structure.
In order to answer Q2, we used the number of only-sown forbs
(NoOSF) in the last observation year (2016) to analyse immigration
success of this group into the unsown grass margins (T5–T6), and
whether mowing time (the main factor with two levels) influenced their
establishment success. We analysed NoOSF with Poisson errors using
the log link function.
We used generalized linear mixed-effect models, package lme4
(Bates et al., 2015), to answer questions Q3 and Q4, whether the per-
centage cover of only sown species (CoOSS), percentage cover of
grasses (CoGr), and number of only sown species (NoOSS) changed with
different sward disturbance intensities and different mowing times
between 2011 and 2016 (T1–T4). Since percentage cover data are
strictly bounded, we fitted linearized models with binomial errors for
CoOSS and CoGr, applying a logit link function (Warton and Hui,
2011). NoOSS was analysed with Poisson errors using the log link
function. For all analyses, we used time (six levels), mowing time (two
levels) and sward disturbance intensity (two levels) as the main factors.
The possibility of temporal pseudoreplication arising from repeated
measurements in the same plots was accounted for by adding plot as a
random variable in the models. We simplified the models as much as
possible to find the minimal adequate model by using a backward se-
lection procedure with likelihood ratio tests (Crawley, 2013).
Related to Q3 and Q4, we tested whether plot occupancy of sown
species belonging to fringe communities, dry or mesic grassland com-
munities, or dry ruderal communities differed between mowing time
and sward disturbance intensities (two main factors) in the last ob-
servation year (2016). Plot occupancy was measured by the percentage
of sown plots (n = 10) with successful establishment of sown species in
the last observation year. Plots with different sward disturbance in-
tensities were pooled because the effect of sward disturbance intensity
was not significant in 2016. Since these data are proportions, we used
GLM with binomial error, except for the dry grassland data where we
used quasi-binomial error to account for over-dispersion, and the logit-
link function (Crawley, 2013).

91
A. Kirmer et al. Agriculture, Ecosystems and Environment 251 (2018) 88–98

Fig. 1. GNMDS ordination, axes 1 and 2, scaled in half-change units, of the 30 sample plots from six treatments for all observation years (2010–2016). The mean resultant displacement of
each combination of treatment by years along the two first GNMDS axes is visualized with different symbols on connecting lines.
T1: sward disturbance once before sowing, T3: sward disturbance three times before sowing, both mown in June;
T2: sward disturbance once before sowing, T4: sward disturbance three times before sowing, both mown in September;
T5: mown in June, T6: mown in September, both without sward disturbance and without sowing.

3. Results
3.1. General vegetation development
The original species composition of the plots was rather similar in
2010 before application of the treatments (Tables 1 and 2). However,
already in 2011, the species composition in all sown plots (T1–T4)
shifted considerably away from their initial positions along the first
GNMDS-axis, whereas the plot trajectory of the T5 and T6 treatments
(only mown, no sward disturbance, and no seeding) was much lower
(Fig. 1). From 2011 until 2014, plots with a similar sward disturbance
intensity remained similar. However, starting in 2015, September-
mown plots, regardless of sward disturbance intensity, showed a dis-
tinct displacement along the first GNMDS-axis towards the unsown
plots. Plots from all six treatments moved considerably over time along
the second axis, and in the same direction. At the end of the experiment,
in 2016, plots without sward disturbance and without sowing still dif-
fered from all sown treatments, with both early mown treatments (T1,
T3) being the most dissimilar from them. Table 1 summarizes the mean
cover and number of sown species for all treatments in 2010 and 2016.
Species richness on all sowing treatments has increased considerably
compared to pre-sowing conditions in 2010. With a few exceptions (e.g.
Galium album, Festuca rupicola, Poa angustifolia), the sown species
profited from early mowing. The most successful sown forbs in the
June-mown treatments were Leucanthemum ircutianum and Knautia ar-
vensis, reaching together about 40% of total vegetation cover compared
to c. 2% in the September-mown sites. The abundance of spontaneously
immigrating or recovering resident species on undisturbed and unsown
sites was quite similar for both management treatments: grasses were
most successful; especially Arrhenatherum elatius and Dactylis glomerata,
followed by Elymus repens and Poa trivialis (Table 2).
3.2. Spontaneous diversification of grass margins
In the undisturbed and unsown plots, removal of the litter layer and
mowing led to a steady increase in species number from 9 ( ± 3) in
2010 to 14 ( ± 3) in 2011, and 20 ( ± 4) in 2016 (Table 1 and 2).
However, there was little successful immigration of target forbs (part of
the OSS species group) into the unsown mowing treatments from ad-
jacent sown treatments. Although the total number of established target
forbs in the early-mown treatments was almost twice as high as in the
late-mown treatments (10 versus 6 species, Table 1), the difference in
mean number was not significant after seven years (p = 0.117) with
mean grass cover of c. 55% in 2016 for both mowing treatments.

92
A. Kirmer et al. Agriculture, Ecosystems and Environment 251 (2018) 88–98

Fig. 2. Percentage of permanent plots (n = 10) with established sown

species belonging to dry ruderal, dry grassland, mesic grassland, and
fringe communities, separated in early (light grey/orange) and late
(medium grey/blue) mowing treatments in the last observation year.
Numbers in brackets refer to the number of sown species per group.
Plots with different sward disturbance intensities (once/three times)
were pooled. (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this article.)

3.3. Development of sown species groups for the late mowing treatments for the two periods, 2011–2014 and
2015–16 (Table 6). CoGR differed significantly between the two dis-
In 2016, sward disturbance intensity before sowing was no longer turbance intensities in the period 2011–2014, being higher in the least
significant due to the growing impact of mowing time. Although the disturbed plots. Nevertheless, CoGR decreased over time (from 2011–
total number of established sown species was quite similar in both 2014 to 2015–2016) in the least disturbed plots whereas it increased in
mowing treatments (42 September-mown, 43 June-mown), the number the most disturbed plots.
of colonized plots per species belonging to the four plant communities
differed between treatments (Fig. 2). After seven years, species of fringe
4. Discussion
communities, and mesic and dry grasslands showed a significantly
higher number of species in the sown sites mown in June compared to
4.1. Diversification of species-poor grass margins by high-diversity sowing of
September-mown sown sites, but not for sown species belonging to dry wild plants
ruderal communities (Table 3).
Overall, percentage cover of only sown species (CoOSS) increased Compiling a high-diversity mixture of 49 wild plants allowed us to
significantly from 2011 until 2013 before stabilising (Fig. 3, Table 4).
pursue different objectives: diminishing the risk of failure caused by
CoOSS increased significantly more over time in the three last years of environmental perturbations, and providing feeding, nesting, and hi-
the experiment, 2014–2016, for June-mown plots compared to Sep-
bernating habitats for different animal species. Species-rich vegetation
tember-mown plots. CoOSS also differed significantly between the two hosting a variety of species traits can respond better to disturbances
sward disturbance intensities, with higher cover in the least disturbed
(Tilman et al., 2006; Proulx et al., 2010) and high diversity seed mix-
plots, but there was no significant difference over time. The number of tures are known to increase the resistance of plant communities to
only sown species (NoOSS) increased significantly from 2011 to 2013,
climate extremes (Isbell et al., 2015; Craven et al., 2016). In the rather
then stabilized at a high level before decreasing to the 2012 level in dry climate of Central Germany, prolonged droughts in late spring and
2016 (Fig. 4, Table 5). The early mown plots had significantly higher
early summer are quite common. On nutrient-rich sites, extreme
NoOSS than the late mown plots (Table 5). drought hampers nitrogen uptake by plants, thus lowering the compe-
tition pressure of nutrient-demanding species (Bloor and Bardgett,
3.4. Development of grasses 2012; Vogel et al., 2012), and benefitting the establishment of less
competitive dry grassland species from the seed mixture.
Mowing time significantly affected cover of grasses (CoGr) over the We found that the number and cover of successfully established
years (Fig. 5, Table 6), decreasing for the early mowing but increasing sown wildflowers is still high even after seven years in all sowing
treatments (T1–T4). This is in contrast to other authors (Warren et al.,
Table 3
2002; Bokenstrand et al., 2004; de Cauwer et al., 2005; Smith et al.,
Effect of mowing time on the occurrence of seeded species belonging to dry ruderal, dry 2010; Fritch et al., 2011), who reported a steady decline in species
grassland, mesic grassland, and fringe communities in plots of treatments T1–T4 in the numbers after sowing irrespective of mowing time or location,
last observation year (2016). Model parameters estimates are derived from the minimal
adequate GLM model with (i) binomial or (ii) quasi-binomial error structure. Reference
levels: mowing time: June. Since none of the parameters was significant for species of dry 4.2. Spontaneous development of regularly mown grass margins
ruderal communities, the minimal adequate model is the null model.
We found a distinct increase in resident grass cover and the emer-
Plant Fixed effects Estimate SE t-value p
community
gence of annual weeds and perennial ruderals in the year following
litter removal. But even after seven years with regular management, the
Dry ruderal (i) Intercept −0.2162 0.1248 −1.733 0.083 vegetation is still dominated by a few competitive grasses. During the
Dry grassland Intercept 0.277 0.185 observation period, Elymus repens was considerable reduced by mowing
(ii) mowing time −0.841 0.266 −3.16 0.005
(September vs June)
with biomass removal, although other grasses, such as Poa angustifolia
Mesic grassland Intercept 1.551 0.240 and Arrhenatherum elatius, benefited from the mowing, especially on the
(i) mowing time −1.282 0.303 −4.24 < 0.001 sites mown in September. This is in accordance with Marshall (1990),
(September vs June) who described that vegetative growth of E. repens is hampered by other
Fringe (i) Intercept 1.237 0.268
grasses, in particular by A. elatius on productive sites.
mowing time −0.726 0.354 −2.05 0.040
(September vs June) Despite growing diaspore pressure due to the increasing population
size of target species on adjacent sown sites, immigration of target forbs

93
A. Kirmer et al. Agriculture, Ecosystems and Environment 251 (2018) 88–98

Fig. 3. Development of cover of the only sown species group on sown sites with different sward disturbance intensity (once/three time) before sowing and different mowing time (once in
June or September) between 2011 and 2016.

Table 4
Effects of mowing time (two-level factor), sward disturbance intensities (two-level
factor), and year (six-level factor) on the cover of only sown species (CoOSS). Model
parameters estimates derived from the minimal adequate model. Reference levels:
mowing time: June, sward disturbance intensities = once, and year = 2011.
Table 5
Effects of mowing time (two-level factor), sward disturbance intensities (two-level
factor), and year (six-level factor) on number of only sown species (NoOSS). Model
parameters estimates derived from the minimal adequate model. Reference levels:
mowing time: June, sward disturbance intensities = once, and year = 2011.

Fixed effects on cover of only sown Estimate SE t-value p Fixed effects on number of only sown Estimate SE z-value p
species species

Intercept −2.687 0.158 Intercept 2.851 0.060

Year (2012 vs 2011) 1.182 0.186 6.37 < 0.001 Year (2012 and 2016 vs 2011) 0.239 0.065 3.66 < 0.001
Year (2013 vs 2011) 1.934 0.186 10.43 < 0.001 Year (2013–2015 vs 2011) 0.372 0.061 6.08 < 0.001
Year (2014, 2015, and 2016 vs 2011) 2.765 0.152 18.25 < 0.001 Mowing time (September vs June) −0.114 0.061 6.08 0.011
Mowing time (September vs June) −0.209 0.207 3.37 0.004
Sward disturbance (3 vs 1) 0.402 0.178 −4.20 < 0.001
Year x Mowing time (2012 [September] vs 0.258 0.262 0.98 0.328 into unsown treatments was negligible in all years, independent on
2011 [June])
mowing time. Over longer time frames, mowing in June combined with
Year x Mowing time (2013 [September] vs −0.303 0.262 −1.15 0.252
2011 [June]) the ongoing loss of nutrients by removal of biomass might diminish the
Year x Mowing time (2014–2016 −1.000 0.214 −4.67 < 0.001 competitive pressure of grasses, increasing the chance for adjacent
[September] vs 2011 [June]) target forbs to colonise the sites. However, Smith et al. (2010) as well as
Valkó et al. (2016) reported that, even after 13 years, only few target
species were able to colonise unsown sites due to high competition from
grasses. In addition, Valkó et al. (2016) stated that establishment gaps
in combination with sowing of target species were necessary to

Fig. 4. Development of species number of the only sown species group on sown sites with different sward disturbance intensity (once/three times) before sowing and different mowing
time (once in June or September) between 2011 and 2016.

94
A. Kirmer et al. Agriculture, Ecosystems and Environment 251 (2018) 88–98

Fig. 5. Development of grass cover on sown sites with different sward disturbance intensity (once/three times) before sowing and different mowing time (once in June or September)
between 2011 and 2016.

Table 6 precipitation in the study region compared to our site. In our dry cli-
Effects of mowing time (two-level factor), sward disturbance intensities (two-level mate, resident species recovery after sward disturbance was slower,
factor), and year (six-level factor) on cover of grasses (CoGr). Model parameters estimates
leaving more gaps for establishment and spread of sown target forbs.
derived from the minimal adequate. Reference levels: mowing time: June, sward dis-
turbance intensities = once, and year = 2011–2014. The highest disturbance intensity is caused by turf stripping, that
removes vegetation, nutrients, and the soil seed bank. Consequently,
Fixed effects on cover of grasses Estimate SE t-value p some authors (Hölzel and Otte, 2003; Edwards et al., 2007; Pywell
et al., 2007) reported higher establishment success of sown species on
Intercept −0.747 0.146
Year (2015–2016 vs 2011–2014) −0.774 0.158 −4.91 < 0.001 sites with turf stripping compared to power harrowing. Our approach
Mowing time (September vs June) 0.044 0.169 0.26 0.797 with rototiller and grubber is more destructive than harrowing, but not
Sward disturbance (3 vs 1) −0.422 0.169 −2.50 0.020 comparable to turf stripping. In regions with higher precipitation, turf
Year x Mowing time (2015–2016 1.247 0.182 6.85 < 0.001 stripping can be beneficial on nutrient-rich former arable land, espe-
[September] vs 2011–2014 [June])
cially if nutrient-poor vegetation is the restoration target (Temperton
Year x Sward disturbance (2015–2016 [3] 0.530 0.182 2.91 0.004
vs 2011–2014 [1]) et al., 2012).
Sward removal with herbicide spraying is not recommendable for
site preparation. Rodriguez and Jacobo (2013) found a dramatic de-
overcome microsite and propagule limitation. Even when the diaspore cline in seed densities of perennial grasses, legumes and nonlegume
pressure is high due to adjacent populations of target species, successful forbs in the seed bank of a temperate grasslands after glyphosate
immigration depends on the availability of establishment niches and treatments. Hofmann and Isselstein (2004) also described lower emer-
therefore, a thorough disturbance of the grass sward is a prerequisite for gence rates of seedlings from sown target species after sward removal
successful diversification. with Glyphosate compared to sward disturbance by harrowing.

4.3. Effect of sward disturbance intensity before sowing on the establishment 4.4. Effect of mowing time in maintaining species-richness and species
of target species composition of restored sown field margins

In the first years following sward disturbance and sowing, we found
higher establishment rates of sown species in treatments with higher
disturbance intensity, although differences converged in the seventh
year. In contrast, Schmiede et al. (2012) found no effect of disturbance
intensity (ploughing vs harrowing) on the establishment rate of in-
troduced floodplain grassland species within three years after the ex-
periment started, but significantly lower cover of resident grasses in
ploughing treatments that benefited sown target forbs. This might be
caused by higher water availability in floodplain grasslands and higher
Unlike the pre-seeding disturbance regime, the vegetation of the
sown field margins reacted astonishingly fast to different mowing
times. After the establishment phase, when all sown treatments were
mown simultaneously, it took only two years until the effect of early
and late mowing was evident, with these differences increasing with
time. On sown sites mown in September, the cover of sown species
decreased steadily to the benefit of competitive grasses. We also found
that late mowing had no short-term effect on plant species richness, but
it strongly affected plant species composition. This is similar to the

95
A. Kirmer et al.
findings of Chaudron et al. (2016) as well as Klimeš et al. (2013) who
also reported that delaying the first mowing to late summer for three or
four subsequent years had no impact on species richness of road verges
or grasslands. Comparable to our results, Klimeš et al. (2013) detected a
distinct change in species abundance on productive grassland sites.
After seven years, the sown species still persisted on our sown treat-
ments mown in September, although with decreasing abundance. A
meta-analysis of European grasslands (Humbert et al., 2012) found that
delaying mowing either from spring to fall, or from early to late
summer had a negative effect on plant species richness. In our study,
further monitoring is necessary to reveal the point when the sown
target species will finally disappear from our September-mown field
margins.
Early mowing of sown treatments supported not only mesic grass-
land species, but also species typical of dry grasslands as well as fringe
communities. Coverage of sown forbs was still high after seven years,
providing copious pollen and nectar resources. Large numbers of the
flowering plant species preferred by pollinators enhances their abun-
dances in the wider landscape (Carrié et al., 2012; Jönsson et al., 2015;
Hatt et al., 2017). Carvell et al. (2007) also found that field margins
sown with mixtures containing nectar- and pollen-producing plants
were more effective in supporting bumblebees than margins sown with
a grass mixture. Establishing a flower-rich vegetation in an im-
poverished agricultural landscape considerably increased the number of
butterflies, burnet moths and wild bees on our species-rich field mar-
gins compared to species-poor grass margins in the same region (own
unpubl. data). As suggested by several authors (de Snoo and van der
Poll, 1999; Marshall and Moonen, 2002; Damgaard et al., 2014; Hahn
et al., 2014), we recommend as being mandatory the establishment of
buffer zones between field margins and adjacent fields to reduce her-
bicide and pesticide drift into field margin sites.
5. Conclusions
Our results, based on continuous monitoring over seven years,
proved that sowing a high-diversity mixture of wild plants in combi-
nation with initial sward disturbance by harrowing and specific man-
agement is a very successful tool for diversifying species-poor grass
margins. By selecting and sowing a high-diversity seed mixture of re-
gional wild ecotypes, the emerging vegetation will strongly support
pollinators such as butterflies and wild bees, thus enhancing the bio-
logical diversity of species-poor, highly productive farmland. Even after
seven years, immigration success of adjacent target forbs into un-
disturbed and unsown grass margins was very low. Mowing once a year
in either early or late summer with biomass removal also increased
species numbers on undisturbed and unsown plots, although mostly
ruderal species emerged. Successful diversification will likely require
active introduction of target species after a thorough disturbance of the
existing grass sward. We strongly recommend adapting species com-
position to the restoration site. Based on our experience on nutrient-rich
sites in a region with low precipitation and prolonged summer drought,
a mixture should contain at least 35 species (see Table 1). Less com-
petitive species, e.g. Campanula patula, Dianthus carthusianorum, Fili-
pendula vulgare, Plantago media, can be added on especially dry sites. In
any case, the distribution areal of species must be taken into account.
And only a small number of seeds for nutrient-fixing legumes should be
used to avoid their dominance.
Mowing time proved to be highly important for restoration success
in the long run. Whereas mowing of sown field margins in late summer
led to considerably higher grass cover and decreased the cover of sown
target species, mowing in June increased their abundance.
Consequently, restored field margins on nutrient-rich sites should be
mown in early summer to maintain plant species richness over time.
Biodiversity conservation on farmland is critical to CAP policy ob-
jectives of halting the loss of biodiversity. Therefore, it would be im-
portant for all European countries to provide competent guidance for
96
Agriculture, Ecosystems and Environment 251 (2018) 88–98
restoring biodiversity on farmland and to give farmers higher incentives
to restore and maintain species-rich field margins in the next CAP
funding period.
Acknowledgements
This study was funded by the German Ministry of Research (grant
no. 323011000033) and the German Federal Environmental
Foundation (grant no. 31006-33/2). Matthias Stolle gave helpful advice
concerning the selection of species for the seed mixture. We also thank
the agricultural faculty, especially Dieter Orzessek and Stefan Gille, for
their support. We are grateful to several colleagues and students, in
particular Nele Adert, Heiner Hensen, Mark Pfau, Stefan Schreiter, and
Lea Schubert, for their help with the field work, and to Keith Edwards
and Nancy Shaw for language corrections.
References
Alignier, A., Baudry, J., 2015. Changes in management practices over time explain most
variation in vegetation of field margins in Brittany, France. Agric. Ecosyst. Environ.
211, 164–172. http://dx.doi.org/10.1016/j.agee.2015.06.008.
Amiaud, B., Touzard, B., Bonis, A., Bouzillé, J.-B., 2008. After grazing exclusion, is there
any modification of strategy for two guerrilla species: Elymus repens (L.) Gould and
Agrostis stolonifera (L.)? Plant Ecol. 197, 107–117. http://dx.doi.org/10.1007/
s11258-007-9364-z.
Auestad, I., Rydgren, K., Austad, I., 2016. Near-natural methods promote restoration of
species-rich grassland vegetation-revisiting a road verge trial after 9 years: near-
natural grassland restoration in road verges. Restor. Ecol. 24, 381–389. http://dx.doi.
org/10.1111/rec.12319.
Badenhausser, I., Cordeau, S., 2012. Sown grass strip—a stable habitat for grasshoppers
(Orthoptera: acrididae) in dynamic agricultural landscapes. Agric. Ecosyst. Environ.
159, 105–111. http://dx.doi.org/10.1016/j.agee.2012.06.017.
Baessler, C., Klotz, S., 2006. Effects of changes in agricultural land-use on landscape
structure and arable weed vegetation over the last 50 years. Agric. Ecosyst. Environ.
115, 43–50. http://dx.doi.org/10.1016/j.agee.2005.12.007.
Bakker, J.P., Berendse, F., 1999. Constraints in the restoration of ecological diversity in
grassland and heathland communities. Trends Ecol. Evol. 14, 63–68. http://dx.doi.
org/10.1016/S0169-5347(98)01544-4.
Bates, D., Maechler, M., Bolker, B., Walker, S., 2015. Lme4: Linear Mixed-Effects Models
Using Eigen and S4 Classes. R Package Version 1.1-9. R Foundation for Statistical
Computing.
Blake, R.J., Woodcock, B.A., Westbury, D.B., Sutton, P., Potts, S.G., 2013. Novel man-
agement to enhance spider biodiversity in existing grass buffer strips. Agric. For.
Entomol. 15, 77–85. http://dx.doi.org/10.1111/j. 1461–9563.2012.00593.x.
Bloor, J.M.G., Bardgett, R.D., 2012. Stability of above-ground and below-ground pro-
cesses to extreme drought in model grassland ecosystems: interactions with plant
species diversity and soil nitrogen availability. Perspect. Plant Ecol. Evol. Syst. 14,
193–204. http://dx.doi.org/10.1016/j.ppees.2011.12.001.
Bokenstrand, A., Lagerlöf, J., Torstensson, P.R., 2004. Establishment of vegetation in
broadened field boundaries in agricultural landscapes. Agric. Ecosyst. Environ. 101,
21–29. http://dx.doi.org/10.1016/S0167-8809(03)00275-5.
Carrié, R.J.G., George, D.R., Wäckers, F.L., 2012. Selection of floral resources to optimise
conservation of agriculturally-functional insect groups. J. Insect Conserv. 16,
635–640. http://dx.doi.org/10.1007/s10841-012-9508-x.
Carvell, C., Meek, W.R., Pywell, R.F., Goulson, D., Nowakowski, M., 2007. Comparing the
efficacy of agri-environment schemes to enhance bumble bee abundance and di-
versity on arable field margins. J. Appl. Ecol. 44, 29–40. http://dx.doi.org/10.1111/
j.1365-2664.2006.01249.x.
Chaudron, C., Chauvel, B., Isselin-Nondedeu, F., 2016. Effects of late mowing on plant
species richness and seed rain in road verges and adjacent arable fields. Agric.
Ecosyst. Environ. 232, 218–226. http://dx.doi.org/10.1016/j.agee.2016.03.047.
Craven, D., Isbell, F., Manning, P., Connolly, J., Bruelheide, H., Ebeling, A., Roscher, C.,
van Ruijven, J., Weigelt, A., Wilsey, B., Beierkuhnlein, C., de Luca, E., Griffin, J.N.,
Hautier, Y., Hector, A., Jentsch, A., Kreyling, J., Lanta, V., Loreau, M., Meyer, S.T.,
Mori, A.S., Naeem, S., Palmborg, C., Polley, H.W., Reich, P.B., Schmid, B., Siebenkäs,
A., Seabloom, E., Thakur, M.P., Tilman, D., Vogel, A., Eisenhauer, N., 2016. Plant
diversity effects on grassland productivity are robust to both nutrient enrichment and
drought. Philos. Trans. R. Soc. B Biol. Sci. 371. http://dx.doi.org/10.1098/rstb.2015.
0277.
Crawley, M.J., 2013. The R Book, 2nd ed. Wiley Chichester.
Damgaard, C., Strandberg, B., Mathiassen, S.K., Kudsk, P., 2014. The effect of glyphosate
on the growth and competitive effect of perennial grass species in semi-natural
grasslands. J. Environ. Sci. Health Part B 49, 897–908. http://dx.doi.org/10.1080/
03601234.2014.951571.
de Snoo, G.R., van der Poll, R.J., 1999. Effect of herbicide drift on adjacent boundary
vegetation. Agric. Ecosyst. Environ. 73, 1–6. http://dx.doi.org/10.1016/S0167-
8809(99)00008-0.
De Cauwer, B., Reheul, D., D’hooghe, K., Nijs, I., Milbau, A., 2005. Evolution of the ve-
getation of mown field margins over their first 3 years. Agric. Ecosyst. Environ. 109,
87–96. http://dx.doi.org/10.1016/j.agee.2005.02.012.
A. Kirmer et al.
Edwards, A.R., Mortimer, S.R., Lawson, C.S., Westbury, D.B., Harris, S.J., Woodcock, B.A.,
Brown, V.K., 2007. Hay strewing, brush harvesting of seed and soil disturbance as
tools for the enhancement of botanical diversity in grasslands. Biol. Conserv. 134,
372–382. http://dx.doi.org/10.1016/j.biocon.2006.08.025.
Fritch, R.A., Sheridan, H., Finn, J.A., Kirwan, L., hUallacháin, D.Ó., 2011. Methods of
enhancing botanical diversity within field margins of intensively managed grassland:
a 7-year field experiment. J. Appl. Ecol. 48, 551–560. http://dx.doi.org/10.1111/j.
1365-2664.2010.01951.x.
Hölzel, N., Otte, A., 2003. Restoration of a species-rich flood meadow by topsoil removal
and diaspore transfer with plant material. Appl. Veg. Sci. 6, 131–140. http://dx.doi.
org/10.1111/j.1654-109X.2003. tb00573. x.
Hackett, M., Lawrence, A., 2014. Multifunctional Role of Field Margins in Arable
Farming, report for European crop protection association (ECPA) by Cambridge en-
vironmental assessments. ADAS UK Ltd.
Hahn, M., Lenhardt, P.P., Brühl, C.A., 2014. Characterization of field margins in in-
tensified agro-ecosystems—why narrow margins should matter in terrestrial pesticide
risk assessment and management. Integr. Environ. Assess. Manag. 10, 456–462.
http://dx.doi.org/10.1002/ieam.1535.
Hansson, M., Fogelfors, H., 1998. Management of permanent set-aside on arable land in
Sweden. J. Appl. Ecol. 35, 758–771. http://dx.doi.org/10.1046/j.1365-2664.1998.
355350. x.
Hart, K., 2015. Green Direct Payments: Implementation Choices of Nine Member States
and Their Environmental Implications. IEEP, London, UK.
Hatt, S., Lopes, T., Boeraeve, F., Chen, J., Francis, F., 2017. Pest regulation and support of
natural enemies in agriculture: experimental evidence of within field wildflower
strips. Ecol. Eng. 98, 240–245. http://dx.doi.org/10.1016/j.ecoleng.2016.10.080.
Hill, M.O., Gauch, H.G., 1980. Detrended correspondence analysis: an improved ordi-
nation technique. Vegetatio 42, 47–58. http://dx.doi.org/10.1007/BF00048870.
Hofmann, M., Isselstein, J., 2004. Seedling recruitment on agriculturally improved mesic
grassland: the influence of disturbance and management schemes. Appl. Veg. Sci. 7,
193–200. http://dx.doi.org/10.1111/j.1654-109X.2004. tb00610. x.
Hovd, H., Skogen, A., 2005. Plant species in arable field margins and road verges of
central Norway. Agric. Ecosyst. Environ. 110, 257–265. http://dx.doi.org/10.1016/j.
agee.2005.04.013.
Humbert, J.-Y., Pellet, J., Buri, P., Arlettaz, R., 2012. Does delaying the first mowing date
benefit biodiversity in meadowland? Environ. Evid. 1, 9. http://dx.doi.org/10.1186/
2047-2382-1-9.
Isbell, F., Craven, D., Connolly, J., Loreau, M., Schmid, B., Beierkuhnlein, C., Bezemer,
T.M., Bonin, C., Bruelheide, H., de Luca, E., Ebeling, A., Griffin, J.N., Guo, Q.,
Hautier, Y., Hector, A., Jentsch, A., Kreyling, J., Lanta, V., Manning, P., Meyer, S.T.,
Mori, A.S., Naeem, S., Niklaus, P.A., Polley, H.W., Reich, P.B., Roscher, C., Seabloom,
E.W., Smith, M.D., Thakur, M.P., Tilman, D., Tracy, B.F., van der Putten, W.H., van
Ruijven, J., Weigelt, A., Weisser, W.W., Wilsey, B., Eisenhauer, N., 2015. Biodiversity
increases the resistance of ecosystem productivity to climate extremes. Nature 526,
574–577. http://dx.doi.org/10.1038/nature15374.
Jönsson, A.M., Ekroos, J., Dänhardt, J., Andersson, G.K.S., Olsson, O., Smith, H.G., 2015.
Sown flower strips in southern Sweden increase abundances of wild bees and ho-
verflies in the wider landscape. Biol. Conserv. 184, 51–58. http://dx.doi.org/10.
1016/j.biocon.2014.12.027.
John, H., Dullau, S., Baasch, A., Tischew, S., 2016. Re-introduction of target species into
degraded lowland hay meadows: how to manage the crucial first year? Ecol. Eng. 86,
223–230. http://dx.doi.org/10.1016/j.ecoleng.2015.11.001.
Kiehl, K., Kirmer, A., Donath, T.W., Rasran, L., Hölzel, N., 2010. Species introduction in
restoration projects–evaluation of different techniques for the establishment of semi-
natural grasslands in Central and Northwestern Europe. Basic Appl. Ecol. 11,
285–299. http://dx.doi.org/10.1016/j.baae.2009.12.004.
Kiehl, K., Kirmer, A., Jeschke, D., Tischew, S., 2014. Restoration of Species-rich Field
Margins and Fringe Communities by Seeding of Native Seed Mixtures, In: Guidelines
for Native Seed Production and Grassland Restoration. Cambridge Scholars
Publishing, pp. 246–275.
Kleijn, D., Joenje, W., Le Coeur, D., Marshall, E.J.P., 1998. Similarities in vegetation
development of newly established herbaceous strips along contrasting European field
boundaries. Agric. Ecosyst. Environ. 68, 13–26. http://dx.doi.org/10.1016/S0167-
8809(97)00098-4.
Klimeš, L., Hájek, M., Mudrák, O., Dančák, M., Preislerová, Z., Hájková, P., Jongepierová,
I., Klimešová, J., 2013. Effects of changes in management on resistance and resilience
in three grassland communities. Appl. Veg. Sci. 16, 640–649. http://dx.doi.org/10.
1111/avsc.12032.
Marshall, E.J., Moonen, A., 2002. Field margins in northern Europe: their functions and
interactions with agriculture. Agric. Ecosyst. Environ. 89, 5–21. http://dx.doi.org/
10.1016/S0167-8809(01)00315-2.
Marshall, E.J.P., 1990. Interference between sown grasses and the growth of rhizome of
Elymus repens (couch grass). Agric. Ecosyst. Environ. 33, 11–22. http://dx.doi.org/
10.1016/0167-8809(90)90140-9.
Meek, B., Loxton, D., Sparks, T., Pywell, R., Pickett, H., Nowakowski, M., 2002. The effect
of arable field margin composition on invertebrate biodiversity. Biol. Conserv. 106,
259–271. http://dx.doi.org/10.1016/S0006-3207(01)00252-X.
Meyer, S., Wesche, K., Krause, B., Leuschner, C., 2013. Dramatic losses of specialist arable
plants in Central Germany since the 1950/60 s–a cross-regional analysis. Divers.
Distrib. 19, 1175–1187. http://dx.doi.org/10.1111/ddi.12102.
Meyer, S., Bergmeier, E., Becker, T., Wesche, K., Krause, B., Leuschner, C., 2015.
Detecting long-term losses at the plant community level–arable fields in Germany
revisited. Appl. Veg. Sci. 18, 432–442. http://dx.doi.org/10.1111/avsc.12168.
Minchin, P.R., 1987. An evaluation of the relative robustness of techniques for ecological
ordination. In: Prentice, I.C., van der Maarel, E. (Eds.), Theory and Models in
Vegetation Science. SpringerNetherlands, Dordrecht, pp. 89–107.
97
Agriculture, Ecosystems and Environment 251 (2018) 88–98
Nathan, R., Schurr, F.M., Spiegel, O., Steinitz, O., Trakhtenbrot, A., Tsoar, A., 2008.
Mechanisms of long-distance seed dispersal. Trends Ecol. Evol. 23, 638–647. http://
dx.doi.org/10.1016/j.tree.2008.08.003.
Økland, R.H., 1986. Rescaling of ecological gradients. I. calculation of ecological distance
between vegetation stands by means of their floristic composition. Nord. J. Bot.
651–660. http://dx.doi.org/10.1111/j.1756-1051.1986. tb00464.x.
Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B., Simpson,
G.L., Solymos, P., Stevens, M.H.H., Wagner, H., 2016. Vegan: Community Ecology
Package. R Package Version 2. 3-3.
Öster, M., Ask, K., Römermann, C., Tackenberg, O., Eriksson, O., 2009. Plant colonization
of ex-arable fields from adjacent species-rich grasslands: the importance of dispersal
vs. recruitment ability. Agric. Ecosyst. Environ. 130, 93–99. http://dx.doi.org/10.
1016/j.agee.2008.12.005.
Pe’er, G., Zinngrebe, Y., Hauck, J., Schindler, S., Dittrich, A., Zingg, S., Tscharntke, T.,
Oppermann, R., Sutcliffe, L.M.E., Sirami, C., Schmidt, J., Hoyer, C., Schleyer, C.,
Lakner, S., 2016. Adding some green to the greening: improving the EU’s ecological
focus areas for biodiversity and farmers: evaluation of EU’s ecological focus areas.
Conserv. Lett. http://dx.doi.org/10.1111/conl.12333.
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.
Proulx, R., Wirth, C., Voigt, W., Weigelt, A., Roscher, C., Attinger, S., Baade, J., Barnard,
R.L., Buchmann, N., Buscot, F., Eisenhauer, N., Fischer, M., Gleixner, G., Halle, S.,
Hildebrandt, A., Kowalski, E., Kuu, A., Lange, M., Milcu, A., Niklaus, P.A., Oelmann,
Y., Rosenkranz, S., Sabais, A., Scherber, C., Scherer-Lorenzen, M., Scheu, S., Schulze,
E.-D., Schumacher, J., Schwichtenberg, G., Soussana, J.-F., Temperton, V.M.,
Weisser, W.W., Wilcke, W., Schmid, B., 2010. Diversity promotes temporal stability
across levels of ecosystem organization in experimental grasslands. PLoS One 5,
e13382. http://dx.doi.org/10.1371/journal.pone.0013382.
Pywell, R.F., Bullock, J.M., Tallowin, J.B., Walker, K.J., Warman, E.A., Masters, G., 2007.
Enhancing diversity of species-poor grasslands: an experimental assessment of mul-
tiple constraints. J. Appl. Ecol. 44, 81–94. http://dx.doi.org/10.1111/j. 1365-2664.
2006.01260.x.
R Development Core Team, 2015. R: a Language and Environment for Statistical
Computing. R Foundation for Statistical Computing, Vienna, Austria.
Rayburn, A.P., Laca, E.A., 2013. Strip-seeding for grassland restoration: past successes
and future potential. Ecol. Restor. 31, 147–153. http://dx.doi.org/10.3368/er.31.2.
147.
Robinson, R.A., Sutherland, W.J., 2002. Post-war changes in arable farming and biodi-
versity in Great Britain. J. Appl. Ecol. 39, 157–176. http://dx.doi.org/10.1046/j.
1365-2664.2002.00695. x.
Rodriguez, A.M., Jacobo, E.J., 2013. Glyphosate effects on seed bank and vegetation
composition of temperate grasslands. Appl. Veg. Sci. 16, 51–62. http://dx.doi.org/10.
1111/j.1654-109X.2012.01213.x.
Rydgren, K., Nordbakken, J.F., Austad, I., Auestad, I., Heegaard, E., 2010. Recreating
semi-natural grasslands: a comparison of four methods. Ecol. Eng. 36, 1672–1679.
http://dx.doi.org/10.1016/j.ecoleng.2010.07.005.
Schmiede, R., Otte, A., Donath, T.W., 2012. Enhancing plant biodiversity in species-poor
grassland through plant material transfer–the impact of sward disturbance. Appl.
Veg. Sci. 15, 290–298. http://dx.doi.org/10.1111/j.1654-109X.2011.01168.x.
Simmering, D., Waldhardt, R., Otte, A., 2013. Erfassung und analyse der pflanze-
nartenvielfalt in der „Normallandschaft“ – ein beispiel aus mittelhessen. Berichte
Reinhold-Tüxen-Ges. 25, 73–94.
Smith, H., Feber, R.E., Morecroft, M.D., Taylor, M.E., Macdonald, D.W., 2010. Short-term
successional change does not predict long-term conservation value of managed arable
field margins. Biol. Conserv. 143, 813–822. http://dx.doi.org/10.1016/j.biocon.
2009.12.025.
Stoate, C., Báldi, A., Beja, P., Boatman, N.D., Herzon, I., van Doorn, A., de Snoo, G.R.,
Rakosy, L., Ramwell, C., 2009. Ecological impacts of early 21 st century agricultural
change in Europe–a review. J. Environ. Manag. 91, 22–46. http://dx.doi.org/10.
1016/j.jenvman.2009.07.005.
Sutcliffe, L.M.E., Batáry, P., Kormann, U., Báldi, A., Dicks, L.V., Herzon, I., Kleijn, D.,
Tryjanowski, P., Apostolova, I., Arlettaz, R., Aunins, A., Aviron, S., Baležentienė, L.,
Fischer, C., Halada, L., Hartel, T., Helm, A., Hristov, I., Jelaska, S.D., Kaligarič, M.,
Kamp, J., Klimek, S., Koorberg, P., Kostiuková, J., Kovács-Hostyánszki, A.,
Kuemmerle, T., Leuschner, C., Lindborg, R., Loos, J., Maccherini, S., Marja, R., Máthé,
O., Paulini, I., Proença, V., Rey-Benayas, J., Sans, F.X., Seifert, C., Stalenga, J.,
Timaeus, J., Török, P., van Swaay, C., Viik, E., Tscharntke, T., 2015. Harnessing the
biodiversity value of Central and Eastern European farmland. Divers. Distrib. 21,
722–730. http://dx.doi.org/10.1111/ddi.12288.
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.
Temperton, V.M., Märtin, L.L.A., Röder, D., Lücke, A., Kiehl, K., 2012. Effects of four
different restoration treatments on the natural abundance of 15N stable isotopes in
plants. Front. Plant Sci. 3, 70. http://dx.doi.org/10.3389/fpls.2012.00070.
Tilman, D., Reich, P.B., Knops, J.M.H., 2006. Biodiversity and ecosystem stability in a
decade-long grassland experiment. Nature 441, 629–632. http://dx.doi.org/10.
1038/nature04742.
Tischew, S., Youtie, B., Kirmer, A., Shaw, N., 2011. Farming for restoration: building
bridges for native seeds. Ecol. Restor. 29, 219–222.
Tscharntke, T., Klein, A.M., Kruess, A., Steffan-Dewenter, I., Thies, C., 2005. Landscape
perspectives on agricultural intensification and biodiversity–ecosystem service
management. Ecol. Lett. 8, 857–874. http://dx.doi.org/10.1111/j.1461-0248.2005.
00782.x.
Valkó, O., Deák, B., Török, P., Kirmer, A., Tischew, S., Kelemen, A., Tóth, K., Miglécz, T.,
A. Kirmer et al.
Radócz, S., Sonkoly, J., Tóth, E., Kiss, R., Kapocsi, I., Tóthmérész, B., 2016. High-
diversity sowing in establishment gaps: a promising new tool for enhancing grassland
biodiversity. Tuexenia 36, 359–378. http://dx.doi.org/10.14471/2016.36.020.
Vickery, J.A., Feber, R.E., Fuller, R.J., 2009. Arable field margins managed for biodi-
versity conservation: a review of food resource provision for farmland birds. Agric.
Ecosyst. Environ. 133, 1–13. http://dx.doi.org/10.1016/j.agee.2009.05.012.
Vogel, A., Scherer-Lorenzen, M., Weigelt, A., 2012. Grassland resistance and resilience
after drought depends on management intensity and species richness. PLoS One 7,
e36992. http://dx.doi.org/10.1371/journal.pone.0036992.
Warren, J., Christal, A., Wilson, F., 2002. Effects of sowing and management on
98
Agriculture, Ecosystems and Environment 251 (2018) 88–98
vegetation succession during grassland habitat restoration. Agric. Ecosyst. Environ.
93, 393–402. http://dx.doi.org/10.1016/S0167-8809(01)00341-3.
Warton, D.I., Hui, F.K.C., 2011. The arcsine is asinine: the analysis of proportions in
ecology. Ecology 92, 3–10. http://dx.doi.org/10.1890/10-0340.1.
White, E.V., Roy, D.B., 2015. A contemporary decennial examination of changing agri-
cultural field sizes using landsat time series data. Geogr. Environ. 2, 33–54. http://dx.
doi.org/10.1002/geo2.4.
van Son, T.C., Halvorsen, R., 2014. Multiple parallel ordinations: the importance of
choice of ordination method and weighting of species abundance data. Sommerfeltia
37, 1–37. http://dx.doi.org/10.2478/som-2014-0001.