RESEARCH ARTICLE

Legacy effects of past land use on current biodiversity in alow-intensity farming landscape in Transylvania (Romania)

Patrick D. Culbert . Ine Dorresteijn . Jacqueline Loos .

Murray K. Clayton . Joern Fischer . Tobias Kuemmerle

Received: 1 October 2015 / Accepted: 29 August 2016

� Springer Science+Business Media Dordrecht 2016

Abstract

Context Ecological impacts of past land use can

persist for centuries. While present-day land use is

relatively easy to quantify, characterizing historical

land uses and their legacies on biodiversity remains

challenging. Southern Transylvania in Romania is a

biodiversity-rich area which has undergone major

political and socio-economic changes, from the Aus-

tro-Hungarian Empire to two World Wars, communist

dictatorship, capitalist democracy, and EU acces-

sion—all leading to widespread land-use changes.

Objectives We investigated whether present-day

community composition of birds, plants, and butter-

flies was associated with historical land use.

Methods We surveyed birds, plants, and butterflies

at 150 sites and classified those sites as forest, arable

land, or managed grassland for six epochs using

historical maps from the 1870s, 1930s, and 1970s,

satellite imagery from 1985 to 2000, and field visits in

2012. Sites were labelled permanent if they had the

same land use at all epochs and non-permanent

otherwise. We used clustering and PERMANOVA

based on community similarity to test for associations

between community composition and land-use

history.

Results We found significant differences (p = 0.030)

in bird communities between permanent and non-

permanent forest sites, and permanent and non-perma-

nent grassland sites (p = 0.051). No significant associ-

ations were found among plants or butterflies and land-

use history.

Conclusions Bird communities were associated

with historical land use, though plants and butter-

flies were not. Historical land-use change in our

study area was likely not sufficiently intense to

cross relevant ecological thresholds that would lead

to legacy effects in present-day plant and butterfly

communities.

Keywords Agricultural intensification � Birds �Butterflies � Farmland abandonment � Historical

ecology � Land-use change � Legacy effects � Plants

Electronic supplementary material The online version ofthis article (doi:10.1007/s10980-016-0441-3) contains supple-mentary material, which is available to authorized users.

P. D. Culbert (&) � T. Kuemmerle

Geography Department, Humboldt-Universitat zu Berlin,

Unter den Linden 6, 10099 Berlin, Germany

e-mail: patrick.culbert@ubc.ca

I. Dorresteijn � J. Fischer

Faculty of Sustainability, Leuphana University Luneburg,

Rotenbleicher Weg 67, 21335 Luneburg, Germany

J. Loos

Agroecology, Georg August University, Grisebachstrasse

6, 37077 Gottingen, Germany

M. K. Clayton

Department of Statistics, University of Wisconsin-

Madison, 1300 University Avenue, Madison, WI 53706,

USA

123

Landscape Ecol

DOI 10.1007/s10980-016-0441-3

Introduction

Agroecosystems provide important habitat for many

species (Donald et al. 2001; Stoate et al. 2009). Land-

use change is a key driver of biodiversity loss in

agroecosystems, (Foley et al. 2005), with both aban-

donment and intensification threatening species’ per-

sistence (Strohbach et al. 2015). However, present-day

land use is not the only important factor influencing

species richness and community composition. Past

land use, particularly agricultural, may result in

ecosystem changes that persist for decades to centuries

(Dupouey et al. 2002; Foster et al. 2003). Yet, few

studies have successfully reconstructed historical land

use and related it to current biodiversity, and most

existing studies have focused on plants only (e.g.,

Dupouey et al. 2002; Lindborg and Eriksson 2004;

Cousins and Lindborg 2008), while the responses of

other taxa remain poorly understood.

Time delayed effects of past land use can occur due

to a range of factors (Lindborg and Eriksson 2004;

Dullinger et al. 2013; Essl et al. 2015). The local

extinction of species, especially long-lived species,

may be substantially delayed following habitat loss or

degradation (Helm et al. 2006; Kuussaari et al. 2009).

Where such an extinction debt exists, contemporary

community composition may better reflect historical

conditions. For example, in a pan-European study of

remnant grassland fragments, present-day richness of

vascular plants was more closely associated with

historical land use patterns, indicating an extinction

debt over a 40 year time frame (Krauss et al. 2010).

However, short-lived specialist butterfly species

showed no extinction debt over this time frame

(Krauss et al. 2010). Conversely, certain forcing

events, such as conversion of an arable field to semi-

natural grassland, may make an area suitable to a new

group of species, though an immigration lag results in

an immigration credit, a temporary deficit in species

while the area is recolonized (Jackson and Sax 2010).

Aside from these time lag effects, other factors may

contribute to a persistent legacy of historical land use

(Foster et al. 2003). Differences in soil characteristics

and plant species composition (Dupouey et al. 2002)

as well as phosphorus levels, pH and seed bank

richness (Plue et al. 2008) were measureable up to

1600–2000 years after cultivation. Past landscape

composition and configuration may also influence

present-day communities. In remnant semi-natural

grasslands in Sweden, herbaceous plant richness at a

given location had a small but significant positive

association with the amount of surrounding grassland

in 1800 (Reitalu et al. 2012), and remnant grassland

species diversity is strongly related to habitat connec-

tivity from 50 to 100 years ago (Lindborg and

Eriksson 2004).

Such legacy effects due to past land use should be

particularly widespread in regions with a long land-

use history. Agriculture has been a dominant driver of

land-use change for centuries in Europe (Ellis and

Ramankutty 2008; Jepsen et al. 2015). Over the past

50 years, Europe as a whole has experienced intensi-

fication of agricultural practices, including increased

applications of fertilizer and pesticides, reduced crop

diversity, increasing mechanization, increasing field

size, and the removal of hedgerows and edge habitat

(Chamberlain et al. 2000; Benton et al. 2003;

Tscharntke et al. 2005; Stoate et al. 2009; Geiger

et al. 2010). This intensification has been associated

with a sharp decline in farmland biodiversity (Cham-

berlain et al. 2000; Donald et al. 2001; Kleijn et al.

2009; Stoate et al. 2009). Along with intensification,

Europe is also experiencing large-scale abandonment

of agricultural land, largely in less profitable locations

(Ceausu et al. 2015; Estel et al. 2015). While

abandonment may benefit some species, it threatens

species dependent on low-intensity agricultural man-

agement (Plieninger et al. 2014; Queiroz et al. 2014).

Eastern Europe is an especially important area in

which to study the relationship between biodiversity

and agricultural land-use change, for two reasons.

First, the level of agricultural intensification has not

yet reached that of Western Europe, and the tradi-

tional farming landscapes that are still widespread in

Eastern Europe continue to support high biodiversity

(Tscharntke et al. 2005; Fischer et al. 2012; Sutcliffe

et al. 2014). However, following EU accession and

EU Common Agricultural Policy, these landscapes

are increasingly threatened by agricultural intensifi-

cation and land abandonment, which may have

severe negative consequences on many species

(Donald et al. 2001, 2002; Loos et al. 2014a, 2015;

Dorresteijn et al. 2015b). Second, Eastern Europe has

experienced several dramatic changes in government

and political systems in the past two centuries,

driving pronounced land-use changes (Lerman et al.

2004; Hostert et al. 2011; Munteanu et al. 2014).

Landscape Ecol

123

These changes can be grouped into several time

periods (Munteanu et al. 2014; Jepsen et al. 2015).

During the 1800s (the Austro-Hungarian Empire),

there was a general trend away from feudalism to land

ownership by peasants, along with the introduction of

new crops and technologies such as the steel plow

(Jepsen et al. 2015). After World War I, agricultural

practices further intensified with new machinery, the

introduction of natural mineral fertilizer, and a shift

from subsistence agriculture to market-oriented pro-

duction (Jepsen et al. 2015). After World War II, under

Soviet influence, most countries in Eastern Europe

experienced dramatic agricultural intensification

through collectivization of agricultural land with an

increase in farm and field sizes, further mechanization,

and increased use of agrochemicals (Jepsen et al.

2015). After the fall of Eastern European socialism in

1989–1991, state farms dissolved and privatized,

resulting in widespread agricultural abandonment

with remaining agricultural production occurring in

large, private, industrial farms or small subsistence

farms (Lerman et al. 2004; Hartvigsen 2014; Estel

et al. 2015). Most recently, as Eastern European

countries have acceded to the European Union, there

has been a mix of agricultural abandonment in

marginal areas, intensification, and some re-cultiva-

tion of formerly abandoned land (Estel et al. 2015).

While land-use change in this region has been

primarily agricultural, forested areas have also under-

gone changes including harvesting, clearing of forest

land, and the planting of timber plantations. Consid-

ering the large magnitude and timespan of land-use

change in Eastern Europe, in order to best develop

biodiversity conservation strategies, it is important to

understand how current biodiversity and species

abundance patterns relate to both present and histor-

ical land use.

Focusing on Transylvania in central Romania as a

case study, the overarching goal of our study was to

determine whether there was a relationship between

present day bird, plant, and butterfly community

composition and the legacy of land use from the

Austro-Hungarian Empire (founded in 1867) to the

present. We hypothesized the following:

1. Present-day land use is the strongest driver of bird,

plant, and butterfly community composition.

2. Historical land use also affects bird, plant, and

butterfly community composition. Specifically,

sites with a single, permanent land use will have

different community compositions than sites that

have changed land use at least once.

3. The association between community composition

and historical land use will be strongest for plants,

because they are most directly affected by phys-

ical changes resulting from land use (e.g., nitrogen

enrichment of soil from fertilizer application, or

changes in physical soil structure from plowing).

Butterfly and bird communities will be less

strongly associated with land-use history because

they are further removed from these primary

effects, although they are still affected by sec-

ondary effects, such as differing vegetation struc-

ture between primary and secondary forests. In

addition, butterflies, and to a greater extent, birds,

have better dispersal ability than plants and may

therefore be able to more quickly recolonize an

area after it has returned to a species’ preferred

land-use type (e.g., re-colonization by grassland

birds of an arable field that transitions to

grassland).

Methods

Study area

The study area covered 7440 km2 in southern Tran-

sylvania, Romania (Fig. 1). The study area included

village catchments within a roughly 50 km radius of

Sighisoara. Mountainous villages were excluded,

leaving a study area characterized by undulating

terrain and a mosaic of agricultural land, grasslands,

and forest. The cities of Sighisoara and Medias were

excluded from consideration. The area is in the

foothills of the Carpathian Mountains with elevation

ranging from 230 to 1100 m above sea level. Romania

in general and Transylvania in particular have very

high biodiversity (Fischer et al. 2012; Wilson et al.

2012; Hanspach et al. 2014). Several species that are

rare or declining in the rest of Europe are common in

Transylvania, such as the corncrake Crex crex (Dor-

resteijn et al. 2015b), European brown bear Ursus

arctos (Roellig et al. 2014), yellow-bellied toad

Bombina variegata (Hartel et al. 2010) and several

butterfly species, such as Lycaena dispar, Phengaris

arion, and Euphydryas aurinia (Loos et al. 2014a).

Landscape Ecol

123

The land use and culture of the region still show

heavy influence from the Saxons, who arrived in the

area in the twelfth century (Akeroyd and Page 2006;

Fischer et al. 2012). The Saxon communities were

highly organized, with communal ownership of forest

and pasture, and private ownership of arable land

(Sutcliffe et al. 2013; Hartel et al. 2014). Following

World War II, roughly 30,000 Saxons were deported,

and other waves of Saxon emigration occurred prior

to, and immediately after the 1989 revolution

(Gundisch 1998). Today, residents of the study area

are primarily Romanian, Hungarian, or Roma, but

Saxon legacies remain, such as communal ownership

of wood pastures (Hartel et al. 2014). Present-day

agricultural practices in the study area are generally of

low intensity, with small-scale cultivation, mixed use

of livestock, low chemical inputs, low mechanization

(horse-drawn plows are common), and production

Fig. 1 Study area

composed of non-

mountainous village

catchments within roughly

50 km of Sighisoara,

Romania. The village

catchments randomly

selected for study are

outlined in black. The cities

of Sighisoara and Medias

were excluded from

consideration

Landscape Ecol

123

mainly for subsistence or local markets (Fischer et al.

2012; Hanspach et al. 2014). Land cover is a mosaic of

forest (28 %), managed grassland (24 %), and arable

land (37 %) (EEA Corine land cover 2006). Agricul-

tural field sizes are small and patches of semi-natural

vegetation are common throughout the landscape

(Loos et al. 2014a).

Study site selection

The study area was delineated into 448 villages and

their surrounding land. In order to obtain a represen-

tative sample of the landscape, villages were stratified

according to terrain ruggedness and protection status

under the EU Birds and Habitats Directives, and 30

villages were randomly selected, covering all combi-

nations of ruggedness and protection status (see Loos

et al. 2014a for further detail). These villages were

selected for use in multiple studies, (e.g., Hanspach

et al. 2014; Loos et al. 2014a; Dorresteijn et al. 2015a),

though ruggedness and protection status were not

considered as covariates in our study. Within each

selected village, each hectare was assigned a woody

vegetation cover category (low: 0–5 %, medium:

5–15 %, high: [15 %) derived from a classification

of 10 m resolution SPOT imagery, as well as a spectral

heterogeneity class (low, medium, or high, defined as

the lower, middle, or upper one-third quantiles of the

standard deviation calculated from SPOT 2.5 m

resolution panchromatic data). Study sites were ran-

domly selected from each woody vegetation—hetero-

geneity combination, such that 60 grassland and 59

arable sites were selected, with an average of four sites

in each of the 30 village catchments (see Loos et al.

2014a for further detail). In addition, 30 forested study

sites were randomly selected, with an average of one

site per village catchment.

Community composition data

Abundance of breeding birds was sampled three times at

30 forest, 58 grassland, and 53 arable sites between mid-

April and mid-June, 2012, using a 10 min point count

including visual and auditory observations from the plot

center. Point counts were conducted in suitable weather

conditions by one of four experienced observers between

5:30 and 11:00 a.m., and all birds heard or seen within the

1 ha site (i.e. within a 56 m circle) were recorded (see

Dorresteijn 2015 for further detail). Plants were surveyed

in eight 1 m2 plots in each of 23 forest, 57 grassland, and

59 arable study sites between May 26 and August 26,

2012. Within the 1 ha circular study sites, plots were

distributed every 45� at stratified random distances from

the plot center, with possible distances alternating

between\40 m and[40 m in order to equally sample

the inner and outer 0.5 ha of each site. All vascular plants

were identified and recorded by percent cover (see Loos

et al. 2015 for further detail). Abundance of butterflies

and burnet moths (which are ecologically similar to

butterflies) was sampled in 59 arable, 60 grassland, and

15 forest sites by walking four 50 m transects in the

cardinal directions, starting 6 m from the plot center.

Butterflies observed within 2.5 m of the transect line and

5 m in front of the observer were identified and recorded.

Surveys were repeated four times in each site between

May and August 2012 by four different trained observers

(see Loos et al. 2014a for further detail).

Land-use history

We used historical maps, classified and raw satellite

imagery, and field visits to classify each study site as

arable, managed grassland, or forest for six epochs:

1870, 1930s, 1970s, 1985, 2000, and 2012s. These

epochs represent five distinct periods: The Austro-

Hungarian Empire (1870), interwar period (1930s),

Socialism (1970s, 1985), transition to Capitalism

(2000), and accession to the EU (2012). Historical

maps were used to classify the first three epochs. The

initial epoch was classified from maps produced by the

2nd and 3rd Military Mapping Surveys of the Austrian

Empire (Fig. 2a, b). Both surveys mapped Transylva-

nia around 1870, and thus we consulted both map sets

simultaneously. Unfortunately, it was not possible to

distinguish arable land from grassland with these

maps, so at this epoch, each study site was classified as

forest or non-forest. For the 1930s and 1970s epochs,

each study site was classified based on topographic

maps from those time periods (Fig. 2c, d). Land cover

for 1985 and 2000 was based on an existing classifi-

cation derived from all available Landsat imagery in

the time window considered (Griffiths et al. 2013),

with the classification of each study site then visually

verified with raw Landsat imagery (and corrected

when necessary). Finally, the present land cover of

each site was verified during field work in 2012.

After classifying land use at each epoch, study sites

were assigned to one of six land-use history classes:

Landscape Ecol

123

permanent forest, non-permanent forest, permanent

grassland, non-permanent grassland, permanent ara-

ble, and non-permanent arable. We used a strict

definition of permanent. For example, a site that was

forest in all epochs would be considered permanent

forest, while a site that was forest in 2012, but non-

forest in any earlier epoch, would be non-permanent

forest. Because the 1870 epoch was classified as forest

or non-forest, sites that were non-forest in 1870 and

grassland in all other epochs were considered perma-

nent grassland, and likewise, sites that were non-forest

in 1870 and arable in all subsequent epochs were

labelled permanent arable.

Statistical approach

We calculated species richness for each taxon as the

total (i.e. pooled) number of species observed across

Fig. 2 Example of historical maps used for land-cover classification in area of Drauseni, Romania. a 2nd Austrian Military Survey, ca.

1870. b 3rd Austrian Military Survey, ca. 1870. c Interwar period, ca. 1930s. d Socialist period, ca. 1970s

Landscape Ecol

123

all point counts, transects, or plots at each study site,

and for each present-day land use, we tested for

difference in richness between permanent and non-

permanent sites using two-tailed, two-sample t-tests.

We also produced tornado plots for each combination

of taxon and present-day land use to visually compare

the abundance of each species between permanent and

non-permanent sites.

To quantify the similarity of plant, butterfly, and bird

communities among permanent and non-permanent

study sites, we used the Yue-Clayton community

similarity index (Yue and Clayton 2005). The Yue-

Clayton community similarity index is a measure of

similarity of two communities ranging from 0 (least

similar) to 1 (most similar). It is related to the Jaccard

index, however it considers species proportions of both

shared and non-shared species (Yue and Clayton 2005).

For each taxon, we calculated the similarity index for all

pairs of sites using the observed total abundances of

each species. We converted the measures from similar-

ity to dissimilarity by subtracting each from 1.

We performed Ward clustering to group the study

sites based on the dissimilarity matrices for each taxon

and present-day land use and displayed the results as

dendrograms. We annotated the dendrograms with

land-use history, woody vegetation level, and hetero-

geneity level and visually examined them for indica-

tions of which factors were driving the observed

clustering. Because dendrograms are not amenable to

quantitative interpretation, we also used PERMA-

NOVA (permutational multivariate analysis of vari-

ance), a statistical method that partitions sources of

variation within distance (i.e. dissimilarity) matrices

(Anderson 2001). We tested for differences in com-

munity composition based on present-day land use and

land-use history. For arable and grassland sites, we

also tested for an effect of land-use history after

accounting for differences in woody vegetation and

heterogeneity, because these are considered major

drivers of contemporary farmland biodiversity (Dor-

resteijn et al. 2015a).

Results

Land-use history

Classification of study sites by land-use history, based

on land use during the six epochs considered, showed

25 permanent and five non-permanent forest sites, 26

permanent and 34 non-permanent grassland sites, and

44 permanent and 15 non-permanent arable sites.

There were slightly more sites classified as forest in

early epochs (Fig. 3). Through time, there was a

noticeable increase of grassland sites and a decrease of

arable sites, with both leveling off in later epochs.

Richness and community composition

In forest sites, species richness of all three taxa was

higher in permanent than non-permanent sites, but the

opposite was true in grassland and arable sites

(Table 1). However, the differences in species rich-

ness were only significant for plants in permanent

versus non-permanent arable sites (p = 0.030, two-

tailed two-sample t test).

Plant communities showed very high dissimilar-

ity among sites, while bird communities had high

dissimilarity within arable and grassland sites and

much lower dissimilarity in forest sites (Supplement

S1). Butterfly communities showed less dissimilar-

ity among study sites than birds or plants (Supple-

ment S1). Examination of tornado plots gave no

evidence of the existence of indicator species that

were strongly associated with permanent or non-

permanent sites (Supplement S2), and there were no

apparent generalizable patterns (e.g., related to

functional groups) regarding differences in commu-

nity composition.

Clustering and PERMANOVA

Clustering indicated a relationship between forest

land-use history and bird communities—in the den-

drogram displaying the clustering of forest sites by

similarity of bird community composition, all non-

permanent sites were in the same cluster at the initial

split (Fig. 4). For the other combinations of taxa and

present-day land use, dendrograms showed some

evidence of clustering related to current land use,

heterogeneity, and woody vegetation, but no cluster-

ing related to land-use history was visually evident

(Supplement S3).

Community composition for all three taxa was

significantly associated with present-day land use, as

confirmed by PERMANOVA results (Table 2). Fur-

thermore, in non-forest sites, community composition

of birds was significantly associated with woody

Landscape Ecol

123

vegetation level, but this was not the case for plants or

butterflies (Table 3). The level of heterogeneity was

only significantly associated with bird communities in

arable sites (Table 4). With regard to land-use history,

a significant difference in bird community

composition was found between permanent and non-

permanent forest sites (p = 0.030), although the

sample size was low for non-permanent forest

(n = 5), and a borderline significant difference

between permanent and non-permanent grassland

Fig. 3 The number of study

sites classified by each land-

use category in each epoch.

The number of sites

classified as forest increased

slightly in the earlier epochs

but were otherwise stable.

Agricultural study sites

showed a steady increase in

the number of grassland

sites at the expense of arable

sites, with land-use

stabilizing in the most recent

epochs

Table 1 Mean and

standard error of species

richness by land use and

taxon

H0 lPerm = lNon-Perm, HA

lPerm = lNon-Perm

p-values are from a two-

sample two-tailed t-test.

Asterisks denote

significance level

(* p\ 0.1, ** p\ 0.05,

*** p\ 0.01)

Taxon Land use Mean species richness p-value

Permanent Non-permanent

Birds Forest 11.0 ± 0.6 (n = 25) 8.8 ± 0.4 (n = 5) 0.123

Birds Grassland 5.4 ± 0.5 (n = 24) 5.5 ± 0.5 (n = 34) 0.831

Birds Arable 5.3 ± 0.6 (n = 42) 6.3 ± 1.2 (n = 15) 0.400

Butterflies Forest 4.8 ± 0.8 (n = 13) 4.5 ± 0.5 (n = 2) 0.900

Butterflies Grassland 20.4 ± 1.4 (n = 26) 21.1 ± 1.5 (n = 34) 0.749

Butterflies Arable 18.1 ± 1.1 (n = 44) 21.7 ± 1.9 (n = 15) 0.111

Plants Forest 16.4 ± 1.9 (n = 19) 14.3 ± 2.9 (n = 4) 0.634

Plants Grassland 55.8 ± 3.1 (n = 23) 61.1 ± 2.2 (n = 32) 0.157

Plants Arable 31.3 ± 2.5 (n = 40) 41.8 ± 3.7 (n = 15) 0.030**

Landscape Ecol

123

(p = 0.051) (Table 5). Land-use history was not

significant for other taxa and land uses.

Because grassland sites had diverse characteristics,

we also used PERMANOVA to verify that the effect

of land-use history remained significant after

accounting for the level of heterogeneity and woody

vegetation at each site. In grassland sites, we found no

significant association (p = 0.847) between bird com-

munities and heterogeneity (Table 4). Woody vegeta-

tion was significantly associated (p = 0.00001) with

Fig. 4 Dendrogram showing Ward clustering based on bird community similarity in forest sites. Study site labels show land use in each

epoch (N non-forest, F forest). The initial split clusters all non-permanent forest sites in the same branch of the tree

Table 2 PERMANOVA results comparing community composition among present-day land use classes

Taxon Current land cover Comparison groups R2 Pr ([F)

Plants All Arable, forest, grassland 0.155 0.00001***

Plants All Forest, non-forest 0.055 0.00001***

Plants Non-forest Arable, grassland 0.128 0.00001***

Birds All Arable, forest, grassland 0.146 0.00001***

Birds All Forest, non-forest 0.128 0.00001***

Birds Non-forest Arable, grassland 0.024 0.00198***

Butterflies All Arable, grassland, forest 0.149 0.00001***

Butterflies All Forest, non-forest 0.126 0.00001***

Butterflies Non-forest Arable, grassland 0.030 0.00411***

Bold indicates the significance of values in Pr ([F) column

For all three taxa, community composition was significantly associated with present-day land use

Asterisks visually indicate the level of significance (* p\ 0.1, ** p\ 0.05, *** p\ 0.01)

Landscape Ecol

123

Table 3 PERMANOVA results comparing community composition among different levels of woody vegetation in non-forest study

sites

Taxon Current land cover Comparison groups R2 Pr ([F)

Plants Grassland Woody vegetation (low, med., high) 0.012 0.79773

Plants Arable Woody vegetation (low, med., high) 0.021 0.25636

Plants Non-forest Woody vegetation (low, med., high) 0.009 0.48851

Birds Grassland Woody vegetation (low, med., high) 0.070 0.00001***

Birds Arable Woody vegetation (low, med., high) 0.124 0.00001***

Birds Non-forest Woody vegetation (low, med., high) 0.080 0.00001***

Butterflies Grassland Woody vegetation (low, med., high) 0.016 0.41370

Butterflies Arable Woody vegetation (low, med., high) 0.014 0.56659

Butterflies Non-forest Woody vegetation (low, med., high) 0.007 0.51049

Bold indicates the significance of values in Pr ([F) column

A significant association exists for birds but not plants or butterflies

Asterisks visually indicate the level of significance (* p\ 0.1, ** p\ 0.05, *** p\ 0.01)

Table 4 PERMANOVA results comparing community composition among different levels of heterogeneity

Taxon Current land cover Comparison groups R2 Pr ([F)

Plants Grassland Heterogeneity (low, med., high) 0.016 0.57511

Plants Arable Heterogeneity (low, med., high) 0.016 0.59402

Plants Non-forest Heterogeneity (low, med., high) 0.007 0.80306

Birds Grassland Heterogeneity (low, med., high) 0.012 0.84707

Birds Arable Heterogeneity (low, med., high) 0.040 0.02016**

Birds Non-forest Heterogeneity (low, med., high) 0.014 0.07347

Butterflies Grassland Heterogeneity (low, med., high) 0.019 0.31252

Butterflies Arable Heterogeneity (low, med., high) 0.013 0.63036

Butterflies Non-forest Heterogeneity (low, med., high) 0.008 0.43984

Bold indicates the significance of values in Pr ([F) column

Only for birds in arable sites was the association significant

Asterisks visually indicate the level of significance (* p\ 0.1, ** p\ 0.05, *** p\ 0.01)

Table 5 PERMANOVA

results comparing

community composition

between sites with

permanent and non-

permanent land-use history

Bold indicates the

significance of values in

Pr ([F) column

Asterisks visually indicate

the level of significance

(* p\ 0.1, ** p\ 0.05,

*** p\ 0.01)

Taxon Current land cover Comparison groups R2 Pr ([F)

Plants Forest Perm. forest, non-perm. forest 0.029 0.921

Plants Grassland Perm. grassland, non-perm. grassland 0.024 0.208

Plants Arable Perm. arable, non-perm. arable 0.024 0.143

Birds Forest Perm. forest, non-perm. forest 0.068 0.030**

Birds Grassland Perm. grassland, non-perm. grassland 0.029 0.051*

Birds Arable Perm. arable, non-perm. arable 0.020 0.400

Butterflies Forest Perm. forest, non-perm. forest 0.055 0.676

Butterflies Grassland Perm. grassland, non-perm. grassland 0.009 0.762

Butterflies Arable Perm. arable, non-perm. arable 0.011 0.709

Landscape Ecol

123

bird community composition in grasslands, and after

accounting for this relationship, the association

between bird community composition and land-use

history was more significant than without accounting

for woody vegetation (p = 0.033) (Table 6).

Discussion

Understanding how past land use influences contem-

porary biodiversity is important for designing effec-

tive conservation policies. Our study sites showed a

steady transition from arable to managed grassland,

stabilizing between 1985 and 2012. Even through

periods of dramatic political and social change, land

use in the study area has changed remarkably grad-

ually over nearly two centuries. Southern Transylva-

nia thus appears to have been less affected by

collectivization and restitution than other parts of

Romania (Kuemmerle et al. 2008). Yet, even with less

dramatic land-use change than expected, we found

evidence of legacy effects, with statistically significant

differences in bird community composition between

permanent and non-permanent forest and between

permanent and non-permanent grasslands.

We found mixed support for our hypotheses. Our

first hypothesis, that present-day land use was the

strongest driver of community composition, was true in

nearly all cases, with a significant relationship between

community composition and land use and a stronger

association (higher R2 value) than with land-use

history, woody vegetation, or heterogeneity. The sole

exception was birds in grassland and arable sites,

where the association between community composi-

tion with woody vegetation (R2 = 0.08, p = 0.00001)

was stronger than the association with land use

(R2 = 0.024, p = 0.002). We found only partial

confirmation of our second hypothesis, that

community composition of birds, plants, and butter-

flies would be associated with land-use history. Bird

communities had a significant association with land-

use history in forest and grassland sites, but not in

arable sites. Surprisingly, no association was found

between land-use history and plant or butterfly com-

munities. This also directly contradicted our third

hypothesis, that the association between land-use

history and community composition would be stron-

gest for plants and weaker for birds and butterflies.

The unexpected lack of a relationship between

land-use history and plant or butterfly community

composition may be related to post-war agricultural

practices in the study area. Agricultural practices in

the study area presently utilize low levels of agro-

chemical inputs, low mechanization, and high labor-

intensity (Fischer et al. 2012; Hanspach et al. 2014).

Agricultural intensification may have been relatively

low in comparison to other areas of Romania, or

Europe in general; moreover, after the collapse of

communism, agricultural practices may have returned

to previous low-intensity levels (Fischer et al. 2012;

Mikulcak et al. 2013). For example, the average

agricultural use of inorganic fertilizer was only 16 kg/

ha in 2013 in the four counties including our study area

(INS—Institutul National de Statistica (Romanian

National Institute of Statistics) 2015), while the

average use in 2009 was 28 kg/ha in Romania and

76 kg/ha for the entire European Union (Eurostat

2012). Likewise, a pan-EU study of agricultural

intensity (Levers et al. 2015) found relatively low

agricultural land-use intensity in our study area (and

Romania as a whole) compared to the rest of the EU.

Low-intensity agricultural fields are more likely to

return to a natural state after abandonment than high-

intensity fields, with a threshold effect between

intensity level and lasting impacts on the biotic

community (Cramer et al. 2008). Perhaps in our study

Table 6 Sequential PERMANOVA table testing association of bird community composition in grassland sites with land-use history,

given woody vegetation

Df Sum of squares Mean square F R2 Pr ([F)

Woody vegetation class 1 1.53 1.53 4.25 0.070 0.00001***

Land-use history 1 0.64 0.64 1.77 0.029 0.033**

Residuals 55 19.83 0.36 0.901

Total 57 22.00 1.000

The effect of land-use history is slightly more significant after accounting for woody vegetation class

Asterisks denote significance level (* p\ 0.1, ** p\ 0.05, *** p\ 0.01)

Landscape Ecol

123

area, the level of intensity in arable fields throughout

socialism was low enough that no thresholds were

exceeded, and fields that transitioned from arable to

grassland were thus able to transition back to a state

similar to that of permanent grasslands.

Plant communities have been shown to be associ-

ated with past land use elsewhere (Dupouey et al.

2002; Lindborg and Eriksson 2004; Standish et al.

2008). In a study site in France, Dupouey et al. (2002)

found that 200 years of agriculture during Roman

times changed soil chemical and physical structure in a

manner still measurable today, which was reflected by

current vegetation. In semi-natural grasslands in

Sweden, present-day plant species diversity was more

strongly associated with landscape configuration

50–100 years ago than with present-day configuration

(Lindborg and Eriksson 2004). In a review of land-use

legacy studies, Foster et al. (2003) concluded that the

major drivers of agricultural land-use legacy effects

relate to changes in the soil (e.g., pH, carbon,

nitrogen). Similarly, Cramer et al. (2008) speculated

that recovery of former agricultural fields was strongly

related to the intensity and duration of agricultural

activity. Evidently, agricultural intensity in our study

area remained low enough such that soil characteris-

tics did not exceed critical ecological thresholds.

After agricultural abandonment (or in our case, the

conversion from arable to grassland) the availability of

seed banks and required dispersal distances influence

recovery of native vegetation (Cramer et al. 2008).

Remnant patches of native vegetation in an agricultural

matrix function as sources of grassland plant species

when former arable fields are converted to managed

grassland, with the influence diminishing with increas-

ing distance (Cousins and Lindborg 2008). Our study

area is notable for small field sizes and a heterogeneous

landscape of forest, grassland, and arable land (Loos

et al. 2014a), potentially reducing the dispersal dis-

tances required for re-colonization of native vegetation

during conversion from arable land to grassland.

Similarly, abundant field margins and roadside vege-

tation (Loos et al. 2014a) may have provided refugia for

plant species extirpated from arable fields. The bulk of

the conversion from arable to grassland in our study

area occurred by our 1985 epoch (Fig. 3), so, assuming

soil characteristics were not dramatically altered by

agriculture, it is possible that the three decades between

that time point and the biodiversity sampling was

sufficient time for these communities in non-permanent

sites to recover to a state similar to that in permanent

sites (i.e., to overcome an immigration credit).

Butterfly community composition is strongly asso-

ciated with plant community composition (Futuyma

1976), and in our study, roughly one-third of species

observed are specialists utilizing a single plant species

for food (Hanspach et al. 2015). Because no significant

differences in plant communities were found between

permanent and non-permanent sites, it is unsurprising

that no significant differences in butterfly communi-

ties were found. Butterfly communities were also more

homogeneous overall than bird or plant communities

(Supplement S1.1), and a previous study using these

butterfly data found significant overlap in community

composition between arable and grassland sites (Loos

et al. 2014a). With limited differences in butterfly

communities among different land uses, it is not

unexpected that no strong community differences

were found between permanent and non-permanent

sites in the same land-use class.

We were surprised to find significant associations

between land-use history and bird community compo-

sition since we did not find associations for plants or

butterflies. The strongest association was between

birds and land-use history in forest sites (R2 = 0.068,

p = 0.030). Though we did not measure forest tree

species composition or stand structure at our study

sites, these attributes are influenced by land-use history

(Rhemtulla et al. 2009), and vegetation structure in

particular can influence bird communities (Culbert

et al. 2013). Similarly, because trees grow slowly (e.g.,

compared to grassland plants), a forest site that was

formerly arable land or grassland may retain, for a long

period of time, floristic and structural attributes that

differ from those in permanent forest. In the Carpathian

Mountains, areas of non-permanent forest were more

likely to be logged in recent times than permanent

forest, even after accounting for other covariates (e.g.,

elevation, distance to roads) (Munteanu et al. 2015).

This could lead to perpetuating differences in forest

structure between non-permanent and permanent

forests. In the case of grasslands, large scattered trees

are an important habitat component for birds (Hartel

et al. 2013). It is possible that grasslands that were once

arable fields would be less likely to include these large

old trees than permanent grasslands.

Aside from ecological explanations for our unex-

pected results, there are potential sources of error in the

analysis. First, sample sizes of sites with non-

Landscape Ecol

123

permanent land-use histories were low for forest

(n = 5) and arable sites (n = 15), potentially limiting

statistical power. Historical land-use classification

may have been influenced by mapping errors during

the surveying and production of the original maps or

during our visual interpretation of land use from these

maps. Forest boundaries in historical maps were easily

identifiable and closely aligned with contemporary

boundaries. However, grassland and arable land were

more difficult to distinguish in the historical maps,

possibly resulting in misclassification. To examine this

possibility, we re-ran the PERMANOVA analysis for

arable and grassland sites with a lax definition of

permanent land-use (i.e., the same land use in at least

five of six epochs). The association between bird

community composition and land-use history in

grassland sites remained significant (p = 0.020 vs.

p = 0.051 for the strict definition), and the association

between plant communities and land-use history in

arable land was closer to significant (p = 0.083 vs.

p = 0.209). There were otherwise no notable changes

in results. It is also possible that sites classified as

permanent may have transitioned to a different land use

and then back to the original one during the period in

between our epochs, although this appears unlikely.

Another potential error source is the biodiversity

sampling. Although our sampling plan was carefully

determined following a pilot study that evaluated

different levels of survey effort (Loos et al. 2014b),

bias due to differences in detectability among species is

still possible in mobile species such as birds (Boulinier

et al. 1998) and butterflies, as well as for sedentary

species like plants (Chen et al. 2009). Lastly, historical

landscape configuration (in addition to composition) is

known to influence biotic communities (Lindborg and

Eriksson 2004; Reitalu et al. 2012). While we were

confident in assigning land use to our study sites from

the historical maps, identifying the precise boundaries

among cover types from the historical maps proved

difficult, and we do not believe the accuracy would

have been adequate for making inference about

landscape measures. It is therefore possible that we

missed effects of landscape pattern in the analysis.

Conclusion

Through the use of historical maps, satellite imagery,

and field visits, we reconstructed historical land use for

150 field sites in southern Transylvania at six epochs

spanning 142 years, and related these land-use histo-

ries to present-day plant, bird, and butterfly commu-

nity compositions. Bird communities in forest and

managed grassland sites were significantly associated

with land-use history, though we found no significant

associations between land-use history and plant or

butterfly community composition. Our land-use his-

tory classification showed that the study area under-

went gradual land-use changes rather than dramatic

transitions, and apparently critical ecological thresh-

olds were not transgressed, even during Socialism. We

suggest that similar analyses be carried out in other

areas of Romania or Eastern Europe that experienced

more dramatic agricultural intensification during the

post-war agricultural collectivization period, and de-

intensification or abandonment in the post-socialist

period since the early 1990s. Many of these areas

harbor high biodiversity that is under threat due to both

agricultural intensification and abandonment.

Although previous studies mainly focused on the

legacy effects of historical land use on plants, our

study showed that land-use legacy can also be

significant for other taxa (e.g., birds). Conservation

planning efforts in particular must be aware that

present-day ecosystems are not solely driven by

present-day conditions, but also historical conditions.

Studies of plant and animal distributions and commu-

nity composition, particularly in areas that have

undergone land-use change, will benefit from consid-

eration of land-use legacies.

Acknowledgments We thank Anne-Catherine Klein, Cosmin

Moga, Elek Telek, Jorg Steiner, Joszef Pal Frink, Kimberley

Pope, Laura Sutcliffe, Laurie Jackson, Paul Kirkland, Pavel Dan

Turtureanu and Remi Bigonneau for their tremendous efforts in

the field. Catalina Munteanu provide helpful translation

assistance for the historical maps. JF was funded through a

Sofja Kovalevskaja Award by the Alexander von Humboldt

Foundation. PC and TK gratefully acknowledge funding by the

European Commission (projects VOLANTE, No. 265104 and

HERCULES, No. 603447) and by the Einstein Foundation,

Berlin.

References

Akeroyd JR, Page N (2006) The Saxon villages of Southern

Transylvania: conserving biodiversity in a historic land-

scape. In: Gafta DD, Akeroyd DJ (eds) Nature conserva-

tion. Springer, Berlin, pp 199–210

Anderson MJ (2001) A new method for non-parametric multi-

variate analysis of variance. Austral Ecol 26:32–46

Landscape Ecol

123

Benton TG, Vickery JA, Wilson JD (2003) Farmland biodi-

versity: is habitat heterogeneity the key? Trends Ecol Evol

18:182–188

Boulinier T, Nichols JD, Sauer JR, Hines JE, Pollock KH (1998)

Estimating species richness: the importance of hetero-

geneity in species detectability. Ecology 79:1018–1028

Ceausu S, Hofmann M, Navarro LM, Carver S, Verburg PH,

Pereira HM (2015) Mapping opportunities and challenges

for rewilding in Europe. Conserv Biol 29:1017–1027

Chamberlain DE, Fuller RJ, Bunce RGH, Duckworth JC,

Shrubb M (2000) Changes in the abundance of farmland

birds in relation to the timing of agricultural intensification

in England and Wales. J Appl Ecol 37:771–788

Chen G, Kery M, Zhang J, Ma K (2009) Factors affecting

detection probability in plant distribution studies. J Ecol

97:1383–1389

Cousins SAO, Lindborg R (2008) Remnant grassland habitats as

source communities for plant diversification in agricultural

landscapes. Biol Conserv 141:233–240

Cramer VA, Hobbs RJ, Standish RJ (2008) What’s new about

old fields? Land abandonment and ecosystem assembly.

Trends Ecol Evol 23:104–112

Culbert PD, Radeloff VC, Flather CH, Kellndorfer JM, Ritten-

house CD, Pidgeon AM (2013) The influence of vertical

and horizontal habitat structure on nationwide patterns of

avian biodiversity. Auk 130:656–665

Donald PF, Green RE, Heath MF (2001) Agricultural intensi-

fication and the collapse of Europe’s farmland bird popu-

lations. Proc R Soc Lond B 268:25–29

Donald PF, Pisano G, Rayment MD, Pain DJ (2002) The

Common Agricultural Policy, EU enlargement and the

conservation of Europe’s farmland birds. Agric Ecosyst

Environ 89:167–182

Dorresteijn I (2015) Biodiversity conservation in traditional

farming landscapes: the future of birds and large carnivores

in Transylvania. Ph.D. dissertation, Leuphana University,

Lueneburg

Dorresteijn I, Loos J, Hanspach J, Fischer J (2015a) Socioeco-

logical drivers facilitating biodiversity conservation in

traditional farming landscapes. Ecosyst Health Sustain

1:1–9

Dorresteijn I, Teixeira L, von Wehrden H, Loos J, Hanspach J,

Stein JAR, Fischer J (2015b) Impact of land cover

homogenization on the Corncrake (Crex crex) in traditional

farmland. Landscape Ecol 30:1483–1495

Dullinger S, Essl F, Rabitsch W, Erb K-H, Gingrich S, Haberl H,

Hulber K, Jarosık V, Krausmann F, Kuhn I, Pergl J, Pysek

P, Hulme PE (2013) Europe’s other debt crisis caused by

the long legacy of future extinctions. Proc Natl Acad Sci

USA 110:7342–7347

Dupouey JL, Dambrine E, Laffite JD, Moares C (2002) Irre-

versible impact of past land use on forest soils and biodi-

versity. Ecology 83:2978–2984

Ellis EC, Ramankutty N (2008) Putting people in the map:

anthropogenic biomes of the world. Front Ecol Environ

6:439–447

Essl F, Dullinger S, Rabitsch W, Hulme PE, Pysek P, Wilson

JRU, Richardson DM (2015) Historical legacies accumu-

late to shape future biodiversity in an era of rapid global

change. Divers Distrib 21:534–547

Estel S, Kuemmerle T, Alcantara C, Levers C, Prishchepov A,

Hostert P (2015) Mapping farmland abandonment and

recultivation across Europe using MODIS NDVI time

series. Remote Sens Environ 163:312–325

Eurostat (2012) Agri-environmental indicator—mineral fer-

tiliser consumption—Statistics Explained. http://ec.

europa.eu/eurostat/statistics-explained/index.php/Agri-

environmental_indicator_-_mineral_fer t i l iser_

consumption. Accessed 27 March 2016

Fischer J, Hartel T, Kuemmerle T (2012) Conservation policy in

traditional farming landscapes. Conserv Lett 5:167–175

Foley JA, DeFries R, Asner GP, Barford C, Bonan G, Carpenter

SR, Chapin FS, Coe MT, Daily GC, Gibbs HK, Helkowski

JH, Holloway T, Howard EA, Kucharik CJ, Monfreda C,

Patz JA, Prentice IC, Ramankutty N, Snyder PK (2005)

Global consequences of land use. Science 309:570–574

Foster D, Swanson F, Aber J, Burke I, Brokaw N, Tilman D,

Knapp A (2003) The importance of land-use legacies to

ecology and conservation. BioScience 53:77–88

Futuyma DJ (1976) Food plant specialization and environ-

mental predictability in Lepidoptera. Am Nat 110:285–292

Geiger F, Bengtsson J, Berendse F, Weisser WW, Emmerson M,

Morales MB, Ceryngier P, Liira J, Tscharntke T, Winqvist C,

Eggers S, Bommarco R, Part T, Bretagnolle V, Plantegenest

M, Clement LW, Dennis C, Palmer C, Onate JJ, Guerrero I,

Hawro V, Aavik T, Thies C, Flohre A, Hanke S, Fischer C,

Goedhart PW, Inchausti P (2010) Persistent negative effects

of pesticides on biodiversity and biological control potential

on European farmland. Basic Appl Ecol 11:97–105

Griffiths P, Muller D, Kuemmerle T, Hostert P (2013) Agri-

cultural land change in the Carpathian ecoregion after the

breakdown of socialism and expansion of the European

Union. Environ Res Lett 8:45024

Gundisch K (1998) Transylvania and the Transylvanian Saxons. In:

SibiWeb. http://sibiweb.de/geschi/history_of_transylvania_

and_the_transylvanian_saxons.php. Accessed 14 Apr 2015

Hanspach J, Hartel T, Milcu AI, Mikulcak F, Dorresteijn I, Loos

J, von Wehrden H, Kuemmerle T, Abson D, Kovacs-

Hostyanszki A, Baldi A, Fischer J (2014) A holistic

approach to studying social-ecological systems and its

application to southern Transylvania. Ecol Soc 19(4):32

Hanspach J, Loos J, Dorresteijn I, von Wehrden H, Moga CI,

David A (2015) Functional diversity and trait composition

of butterfly and bird communities in farmlands of Central

Romania. Ecosyst Health Sustain 1:1–8

Hartel T, Dorresteijn I, Klein C, Mathe O, Moga CI, Ollerer K,

Roellig M, von Wehrden H, Fischer J (2013) Wood-pas-

tures in a traditional rural region of Eastern Europe:

Characteristics, management and status. Biol Conserv

166:267–275

Hartel T, Fischer J, Campeanu C, Milcu AI, Hanspach J, Fazey I

(2014) The importance of ecosystem services for rural

inhabitants in a changing cultural landscape in Romania.

Ecol Soc 19(2):42

Hartel T, Schweiger O, Ollerer K, Cogalniceanu D, Arntzen JW

(2010) Amphibian distribution in a traditionally managed

rural landscape of Eastern Europe: probing the effect of

landscape composition. Biol Conserv 143:1118–1124

Hartvigsen M (2014) Land reform and land fragmentation in

Central and Eastern Europe. Land Use Policy 36:330–341

Landscape Ecol

123

Helm A, Hanski I, Partel M (2006) Slow response of plant

species richness to habitat loss and fragmentation. Ecol

Lett 9:72–77

Hostert P, Kuemmerle T, Prishchepov A et al (2011) Rapid land

use change after socio-economic disturbances: the collapse

of the Soviet Union versus Chernobyl. Environ Res Lett

6:45201

INS—Institutul National de Statistica (Romanian National

Institute of Statistics) (2015) http://statistici.insse.ro/shop/

?page=tempo2&lang=ro&context=45. Accessed 27 March

2016

Jackson ST, Sax DF (2010) Balancing biodiversity in a changing

environment: extinction debt, immigration credit and spe-

cies turnover. Trends Ecol Evol 25:153–160

Jepsen MR, Kuemmerle T, Muller D, Erb K, Verburg PH,

Haberl H, Vesterager JP, Andric M, Antrop M, Austrheim

G, Bjorn I, Bondeau A, Burgi M, Bryson J, Caspar G,

Cassar LF, Conrad E, Chromy P, Daugirdas V, Van Eet-

velde V, Elena-Rossello R, Gimmi U, Izakovicova Z,

Jancak V, Jansson U, Kladnik D, Kozak J, Konkoly-Gyuro

E, Krausmann F, Mander U, McDonagh J, Parn J,

Niedertscheider M, Nikodemus O, Ostapowicz K, Perez-

Soba M, Pinto-Correia T, Ribokas G, Rounsevell M,

Schistou D, Schmit C, Terkenli TS, Tretvik AM, Trzepacz

P, Vadineanu A, Walz A, Zhllima E, Reenberg A (2015)

Transitions in European land-management regimes

between 1800 and 2010. Land Use Policy 49:53–64

Kleijn D, Kohler F, Baldi A, Batary P, Concepcion E, Clough Y,

Diaz M, Gabriel D, Holzschuh A, Knop E, Kovacs A,

Marshall EJ, Tscharntke T, Verhulst J (2009) On the

relationship between farmland biodiversity and land-use

intensity in Europe. Proc R Soc B 276:903–909

Krauss J, Bommarco R, Guardiola M, Heikkinen RK, Helm A,

Kuussaari M, Lindborg R, Ockinger E, Partel M, Pino J,

Poyry J, Raatikainen KM, 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

Kuemmerle T, Muller D, Griffiths P, Rusu M (2008) Land use

change in Southern Romania after the collapse of social-

ism. Reg Environ Change 9:1–12

Kuussaari M, Bommarco R, Heikkinen RK, Helm A, Krauss J,

Lindborg R, Ockinger E, Partel M, Pino J, Roda F, Ste-

fanescu C, Teder T, Zobel M, Steffan-Dewenter I (2009)

Extinction debt: a challenge for biodiversity conservation.

Trends Ecol Evol 24:564–571

Lerman Z, Csaki C, Feder G (2004) Evolving farm structures

and land use patterns in former socialist countries. Q J Int

Agric 43:309–336

Levers C, Muller D, Erb K, Haberl H, Jepsen MR, Metzger MJ,

Meyfroidt P, Plieninger T, Plutzar C, Sturck J, Verburg PH,

Verkerk PJ, Kuemmerle T (2015) Archetypical patterns

and trajectories of land systems in Europe. Reg Environ

Change doi:10.1007/s10113-015-0907-x

Lindborg R, Eriksson O (2004) Historical landscape connec-

tivity affects present plant species diversity. Ecology

85:1840–1845

Loos J, Dorresteijn I, Hanspach J, Fust P, Rakosy L, Fischer J

(2014a) Low-intensity agricultural landscapes in transyl-

vania support high butterfly diversity: implications for

conservation. PloS One 9:e103256

Loos J, Hanspach J, von Wehrden H, Beldean M, Moga CI,

Fischer J (2014b) Developing robust field survey protocols

in landscape ecology: a case study on birds, plants and

butterflies. Biodivers Conserv 24(1):33–46

Loos J, Turtureanu PD, von Wehrden H, Hanspach J, Dor-

resteijn I, Frink JP, Fischer J (2015) Plant diversity in a

changing agricultural landscape mosaic in Southern Tran-

sylvania (Romania). Agric Ecosyst Environ 199:350–357

Mikulcak F, Newig J, Milcu AI, Hartel T, Fischer J (2013)

Integrating rural development and biodiversity conserva-

tion in Central Romania. Environ Conserv 40:129–137

Munteanu C, Kuemmerle T, Boltiziar M, Butsic V, Gimmi U,

Halada L, Kaim D, Kiraly G, Konkoly-Gyuro E, Kozak J,

Lieskovsky J, Mojses M, Muller D, Ostafin K, Ostapowicz

K, Shandra O, Stych P, Walker S, Radeloff VC (2014)

Forest and agricultural land change in the Carpathian

region—a meta-analysis of long-term patterns and drivers

of change. Land Use Policy 38:685–697

Munteanu C, Kuemmerle T, Keuler NS, Muller D, Balazs P,

Dobosz M, Griffiths P, Halada L, Kaim D, Kiraly G,

Konkoly-Gyuro E, Kozak J, Lieskovsky J, Ostafin K,

Ostapowicz K, Shandra O, Radeloff VC (2015) Legacies of

19th century land use shape contemporary forest cover.

Glob Environ Change 34:83–94

Plieninger T, Hui C, Gaertner M, Huntsinger L (2014) The

impact of land abandonment on species richness and

abundance in the mediterranean basin: a meta-analysis.

PloS One 9:e98355

Plue J, Hermy M, Verheyen K, Thuillier P, Saguez R, Decocq G

(2008) Persistent changes in forest vegetation and seed

bank 1600 years after human occupation. Landscape Ecol

23:673–688

Queiroz C, Beilin R, Folke C, Lindborg R (2014) Farmland

abandonment: threat or opportunity for biodiversity con-

servation? A global review. Front Ecol Environ

12:288–296

Reitalu T, Purschke O, Johansson LJ, Hall K, Sykes MT,

Prentice HC (2012) Responses of grassland species rich-

ness to local and landscape factors depend on spatial scale

and habitat specialization. J Veg Sci 23:41–51

Rhemtulla JM, Mladenoff DJ, Clayton MK (2009) Legacies of

historical land use on regional forest composition and

structure in Wisconsin, USA (mid-1800s-1930s-2000s).

Ecol Appl 19:1061–1078

Roellig M, Dorresteijn I, von Wehrden H, Hartel T, Fischer J

(2014) Brown bear activity in traditional wood-pastures in

Southern Transylvania, Romania. Ursus 25:43–52

Standish RJ, Cramer VA, Hobbs RJ (2008) Land-use legacy and

the persistence of invasive Avena barbata on abandoned

farmland. J Appl Ecol 45:1576–1583

Stoate C, Baldi A, Beja P, Boatman ND, Herzon I, van Doorn A,

de Snoo GR, Rakosy L, Ramwell C (2009) Ecological

impacts of early 21st century agricultural change in Eur-

ope—a review. J Environ Manage 91:22–46

Strohbach MW, Kohler ML, Dauber J, Klimek S (2015) High

nature value farming: from indication to conservation. Ecol

Indic 57:557–563

Sutcliffe LME, Batary P, Kormann U, Baldi A, Dicks LV,

Herzon I, Kleijn D, Tryjanowski P, Apostolova I, Arlettaz

R, Aunins A, Aviron S, Balezentien _e L, Fischer C, Halada

L, Hartel T, Helm A, Hristov I, Jelaska SD, Kaligaric M,

Landscape Ecol

123

Kamp J, Klimek S, Koorberg P, Kostiukova J, Kovacs-

Hostyanszki A, Kuemmerle T, Leuschner C, Lindborg R,

Loos J, Maccherini S, Marja R, Mathe O, Paulini I, Proenca

V, Rey-Benayas J, Sans FX, Seifert C, Stalenga J, Timaeus

J, Torok P, van Swaay C, Viik E, Tscharntke T (2014)

Harnessing the biodiversity value of Central and Eastern

European farmland. Divers Distrib 21:722–730

Sutcliffe LME, Paulini I, Jones G, Marggraf R, Page N (2013)

Pastoral commons use in Romania and the role of the

Common Agricultural Policy. Int J Commons 7:58–72

Tscharntke T, Klein AM, Kruess A, Steffan-Dewenter I, Thies C

(2005) Landscape perspectives on agricultural intensifica-

tion and biodiversity—ecosystem service management.

Ecol Lett 8:857–874

Wilson JB, Peet RK, Dengler J, Partel M (2012) Plant species

richness: the world records. J Veg Sci 23:796–802

Yue JC, Clayton MK (2005) A similarity measure based on

species proportions. Commun Stat-Theory Methods

34:2123–2131

Landscape Ecol

123