The Effect of Grazing on Central Anatolian Steppe Vegetation: A Modeling Approach Using Functional Traits

Abstract

Grazing is a major ecological driver that influences vegetation dynamics globally. We investigated the long‐term effects of different grazing regimes on the vegetation structure of the Central Anatolian steppes, a region characterized by its unique convergence of biogeographical influences and historical land use. We employed the spatially explicit FATELAND model to simulate vegetation dynamics over a 50‐year period under three distinct grazing scenarios: no grazing, moderate grazing, and overgrazing. Our simulations incorporated a range of plant functional traits to predict changes across five different vegetation types in Central Anatolia, including woodland steppes and treeless steppes. The simulations revealed that moderate grazing supports the diversity and abundance of various plant functional groups, excluding resprouter trees, which flourish under no grazing conditions. In contrast, overgrazing leads to significant reductions in the abundance of perennial forbs and both spiny and non‐spiny subshrubs, often resulting in a shift toward grassland dominated by resprouter gramineae or an annual herb‐dominated grassland, depending on the initial abundance of gramineae. Our findings highlight the critical role of grazing management in maintaining biodiversity and ecological stability in steppe ecosystems. While moderate grazing can enhance plant functional group diversity, overgrazing significantly threatens the ecological integrity of the Central Anatolian steppes. In conclusion, our modeling approach reveals that the grazing regime is a major driver in shaping the vegetation structure of Central Anatolian steppes. Grazing management strategies that are adjusted to the ecological characteristics and historical context of specific regions are required to prevent degradation and promote sustainable grassland vegetation.

Full text

1 of 16
Ecology and Evolution, 2024; 14:e70499
https://doi.org/10.1002/ece3.70499
Ecology and Evolution
RESEARCH ARTICLE OPEN ACCESS
The Effect of Grazing on Central Anatolian Steppe
Vegetation: A Modeling Approach Using Functional Traits
AnılBahar1,2 | ÇağatayTavşanoğlu2
1Institute of Science, Hacettepe University, Ankara, Türkiye | 2Division of Ecology, Department of Biology, Hacettepe University, Ankara, Türkiye
Correspondence: Çağatay Tavşanoğlu (ctavsanoglu@gmail.com)
Received: 12 May 2024 | Revised: 8 October 2024 | Accepted: 14 October 2024
Funding: The authors received no specific funding for this work.
Keywords: Anatolian steppes| disturbance| grazing regimes| modeling| plant functional groups| temperate grasslands| vegetation dynamics
ABSTRACT
Grazing is a major ecological driver that influences vegetation dynamics globally. We investigated the long- term effects of dif-
ferent grazing regimes on the vegetation structure of the Central Anatolian steppes, a region characterized by its unique conver-
gence of biogeographical influences and historical land use. We employed the spatially explicit FATELAND model to simulate
vegetation dynamics over a 50- year period under three distinct grazing scenarios: no grazing, moderate grazing, and overgraz-
ing. Our simulations incorporated a range of plant functional traits to predict changes across five different vegetation types in
Central Anatolia, including woodland steppes and treeless steppes. The simulations revealed that moderate grazing supports
the diversity and abundance of various plant functional groups, excluding resprouter trees, which flourish under no grazing
conditions. In contrast, overgrazing leads to significant reductions in the abundance of perennial forbs and both spiny and non-
spiny subshrubs, often resulting in a shift toward grassland dominated by resprouter gramineae or an annual herb- dominated
grassland, depending on the initial abundance of gramineae. Our findings highlight the critical role of grazing management in
maintaining biodiversity and ecological stability in steppe ecosystems. While moderate grazing can enhance plant functional
group diversity, overgrazing significantly threatens the ecological integrity of the Central Anatolian steppes. In conclusion, our
modeling approach reveals that the grazing regime is a major driver in shaping the vegetation structure of Central Anatolian
steppes. Grazing management strategies that are adjusted to the ecological characteristics and historical context of specific re-
gions are required to prevent degradation and promote sustainable grassland vegetation.
ÖZET
Otlatma, küresel olarak vejetasyon dinamiklerini etkileyen önemli ekolojik bir faktördür. Bu çalışmada, geçmişteki arazi kul-
lanımının belirgin biçimde öne çıktığı ve biyocoğrafi olarak eşsiz bir bölgede yer alan İç Anadolu bozkırlarında, farklı otlatma
rejimlerinin uzun vadeli etkileri incelenmiştir. Vejetasyon dinamiklerini 50 yıllık bir süre boyunca üç farklı otlatma senaryosu
altında simüle etmek için mekânsal özellikli FATELAND modelini kullandık: otlatma yok, orta düzeyde otlatma ve aşırı ot-
latma. Simülasyonlarımız, ağaçlı ve ağaçsız bozkırlar dahil, İç Anadolu'daki beş farklı vejetasyon tipinde, değişiklikleri tah-
min etmek için çeşitli bitki fonksiyonel karakterlerini içermektedir. Simülasyonlar, orta düzeyde otlatmanın bitki fonksiyonel
grubu çeşitliliğini ve bolluğunu desteklediğini, ancak yeniden sürgün verebilen ağaçların yalnızca otlatma olmayan koşull-
arda arttığını göstermiştir. Buna karşılık, aşırı otlatma, çok yıllık otsu bitkilerin ve dikenli/dikensiz yarı çalıların bolluğunda
önemli seviyede azalmalara yol açmış ve genellikle, başlangıçtaki gramineae bolluğuna bağlı olarak, yeniden sürgün verebilen
gramineae'lerin hâkim olduğu çayırlara veya tek yıllık otlarla kaplı çayırlara dönüşüme neden olmuştur. Bulgularımız, bozkır
This is a n open access ar ticle under the terms of t he Creative Commons Attr ibution License, which p ermits use, dis tribution and repro duction in any medium, p rovided the orig inal work is
properly cited.
© 2024 T he Author(s). Ecology and Evolution publis hed by John Wiley & Sons L td.

2 of 16 Ecology and Evolution, 2024
ekosistemlerinde biyoçeşitliliği ve ekolojik kararlılığı sürdürmede otlatma yönetiminin kritik bir rolü olduğunu göstermektedir.
Orta düzeyde otlatma bitki fonksiyonel grup çeşitliliğini artırabilirken, aşırı otlatma İç Anadolu bozkırlarının ekolojik bütün-
lüğünü ciddi şekilde tehdit etmektedir. Sonuç olarak, modelleme yaklaşımımız, otlatma rejiminin İç Anadolu bozkırlarının
vejetasyon yapısını şekillendirmede başlıca bir faktör olduğunu ortaya koymaktadır. Belirli bölgelerin ekolojik özelliklerine ve
tarihi bağlamına uygun olarak tasarlanmış otlatma yönetimi stratejileri, arazi bozulmasını önlemek ve sürdürülebilir bir çayır
vejetasyonunu teşvik etmek için gereklidir.
1 | Introduction
Grazing has the ability to significantly alter ecosystems on a
global scale (Asner et al. 2004; Zhang etal. 2023). Many open
ecosystems, such as grasslands are maintained by disturbances,
including vertebrate herbivory (Evans et al. 2015; Dantas
etal. 2 016; Bond 2019). The intensity of grazing markedly in-
fluences the abundance of herbaceous species in savanna and
grassland ecosystems (Hayes and Holl 2003; Gebremedhn
etal.2023) and can alter the spatial heterogeneity of vegetation
(Adler, Raff, and Lauenroth2001). Moderate and low- intensity
grazing play a crucial role in sustaining high levels of biodi-
versity within grassland ecosystems (Isselstein et al. 2007;
Török etal.2 016; Joubert, Pryke, and Samways2017; Wolański
etal.2021; Davies etal.2024). Overgrazing, on the other hand,
leads to a significant decline in grassland biodiversity (Wesche
etal.2016; Rahmanian etal.2019; Munkhzul etal.2021; Zhang,
Wang, and Niu2021). Many studies on the impact of grazing
on grasslands have shown that low- intensity grazing has a
positive effect, while overgrazing negatively impacts grassland
biodiversity (Metera et al. 2010). For example, in Mongolian
temperate grasslands, increasing grazing intensity negatively af-
fects plant cover and aboveground biomass, with high- intensity
grazing leading to a decline in tall grasses and an increase in
short grasses (Zainelabdeen et al. 2020). European grasslands
also show this trend: the type of grazer affects biodiversity and
species composition; cattle grazing, compared to sheep grazing,
promotes more trait- rich vegetation with higher forb cover (Tóth
et al. 2018). A long- term field experiment in North American
semi- arid grasslands demonstrated that long- term grazing
causes slow, continuous changes in plant communities without
inducing alternative stable states, emphasizing the importance
of understanding phase shifts rather than focusing solely on
thresholds between states (Porensky et al. 2016). In savanna
grasslands, heavy cattle grazing can increase tree density by
reducing grass biomass and creating more open sites for tree
seedling establishment, which may eventually lead to the ces-
sation of grazing in woody- encroached grasslands (Mogashoa,
Dlamini, and Gxasheka2021).
Indeed, grassland vegetation undergoes dynamic changes over
time in response to disturbances such as herbivory and fire (Van
Langevelde et al. 2003; Sankaran, Ratnam, and Hanan 2004;
Baudena et al. 2015; Dantas et al. 2016; Bond 2019). As a re-
sult, it often coexists with or transitions into alternative stable
states with woodlands (Beisner, Haydon, and Cuddington2003;
Pausas and Bond 2020). Herbivory not only promotes grass
establishment but also regulates tree cover in savannas (Van
Langevelde et al.2003; Bond 2019). Grazing is one of the pri-
mary drivers of state transition in grasslands (Twidwell, Allred,
and Fuhlendorf2013), although these changes takes more than
a decade to occur (Milchunas2011; Bestelmeyer etal.2013). In
many other parts of the world, however, the effect of grazing on
grassland vegetation also depends on vegetation type and local
climate (Munkhzul etal.2021).
The temperate grasslands of Central Anatolia are renowned for
their extraordinary biodiversity, primarily due to the unique
convergence of plant species from both the Irano- Turanian and
Mediterranean phytogeographic regions (Ekim and Güner2000;
Kurt, Nilhan, and Ketenoglu 2006). The region has a long-
standing history of pastoralism, shaping its landscape for nearly
10,000 years (Hammer and Arbuckle 2017; Middleton 2018).
Archeological evidence suggests the earliest domestication of
herbivores in the region around 9000 cal BC, with earlier ev-
idence from the Eastern Mediterranean dating back to 12,000
cal BC (Payne1985; Zeder2008; Middleton2018). Additionally,
aurochs (Bos primigenius) and mouf lon (Ovis orientalis) existed
in Anatolia during the Neolithic ages (Perkins and Daly1968).
The domestication timeline reveals that sheep and goats were
domesticated before cattle, with domestic pigs being domesti-
cated much later than in surrounding regions (Arbuckle2013;
Peters etal.2013). Domestic grazing, a form of herbivory regu-
lated by human intervention, involves the consumption of vege-
tation by domesticated mammalian species (Metera etal.2010;
Rosenthal, Schrautzer, and Eichberg 2012). Herbivory, on the
other hand, involves the consumption of plants by a mix of do-
mestic and wild animals, ranging from insects and rodents to
large mammals. Currently, domestic grazing, predominantly
by cattle, sheep, and goats, is the main form of mammal her-
bivory in Central Anatolian steppes (Fırıncıoğlu, Seefeldt, and
Şahin2007; Tavşanoğlu2017). Over the last 50 years in Central
Anatolia, intensified agricultural activities led to the conversion
of grasslands to croplands, driving overgrazing in many remain-
ing rangelands (Fırıncıoğlu, Seefeldt, and Şahin2007; Ambarlı
etal.2016). Consequently, overgrazing has been a major driver
of habitat degradation in many parts of Central Anatolia
(Kürschner and Parolly2012; Koc, Schacht, and Erkovan2015;
Ambarlı etal.2016; Gökbulak etal.2018).
It is widely recognized that overgrazing is a primary fac-
tor in vegetation degradation in Central Anatolia (Ünal and
Fırıncıoğlu2007; Gökbulak et al.2018). In his region, grazing
significantly reduces the cover of forbs and grasses (Fırıncıoğlu
et al. 2009, 2010), with forbs experiencing the most substan-
tial negative impact (Fırıncıoğlu, Seefeldt, and Şahin 2007).
However, the cover of cushion- type subshrubs appears similar
in both grazed and ungrazed areas (Fırıncıoğlu et al. 2010).
Certain subshrub species, such as Thymus spp., have shown
increased cover in grazed areas (Fırıncıoğlu etal.2009). Such
experiments suggest that plant cover is more robust in ungrazed
areas than in grazed ones, plant diversity does not increase

3 of 16
with grazing, and grasslands are transitioned to subshrub-
dominated sites following extensive overgrazing in Central
Anatolia (Fırıncıoğlu, Seefeldt, and Şahin 2007; Fırıncıoğlu
etal.2010). On the other hand, an experiment that included ar-
tificial disturbances to vegetation and soil suggests that Central
Anatolian vegetation has resilience to disturbance, at least on
a small scale (Özüdoğru, Özüdoğru, and Tavşanoğlu 2021).
The diversity of belowground organs in herbaceous plants of
Anatolian steppes (Ülgen and Tavşanoğlu2024) also highlights
the resilience of this vegetation to various disturbances. Indeed,
many species resprout after being clipped at ground level
(Tavşanoğlu et al., unpublished data). In addition to grazing,
seeds of Central Anatolian steppe plants can resist low- intensity
heat shocks, suggesting that these plants are also resilient to
surface fire regimes (Tavşanoğlu, Çatav, and Özüdoğru 2015)
which were frequent during the early- to mid- Holocene (Turner
etal.2010). Although there are studies providing a general over-
view of the effects of grazing on Central Anatolian vegetation
(Gintzburger, Le Houérou, and Saïdi2006; Ambarlı etal.2016;
Tavşanoğlu2017; Gökbulak etal.2018) and a few studies men-
tioned above that provide experimental data from controlled
experiments (Fırıncıoğlu, Seefeldt, and Şahin2007; Fırıncıoğlu
etal.2009, 2010; Özüdoğru, Özüdoğru, and Tavşanoğlu2021),
the absence of long- term studies hinders our ability to predict the
dynamics of grazing–vegetation relationships in the steppe vege-
tation of Central Anatolia under global environmental changes.
Although modeling studies may enhance our knowledge of
long- term dynamics in such cases, no modeling study has yet
been conducted to specifically examine the impact of grazing on
shaping the steppe vegetation of Central Anatolia. Considering
the current acceleration of agricultural land abandonment and
changes in domestic grazing regimes (Tavşanoğlu2017), unrav-
eling long- term vegetation dynamics is crucial to fully under-
stand the post- abandonment recovery processes, conservation,
and management of Anatolian steppe vegetation in the near
future. Modeling the long- term vegetation dynamics of the un-
derstudied Central Anatolian vegetation in response to various
grazing regimes would also enhance our understanding of how
grazing shapes temperate grassland ecosystems.
Vegetation ecology fundamentally revolves around patterns
and processes that vary significantly across spatial and tem-
poral scales (Wiegleb 1989). A pivotal aspect of this field is
the use of disturbance- based vegetation models in which are
instrumental in testing hypotheses about vegetation changes
under disturbances of various frequency and intensity (Adler,
Raff, and Lauenroth 2001; Pausas 2006; Seidl et al. 2011). In
disturbance- prone environments, models that incorporate
both disturbance and response mechanisms become essential
for thoroughly understanding vegetation mechanics (Lavorel
etal. 1997; Millington et al. 2009). Disturbance- based models
offer advantages over dynamic global vegetation models, gap
models, and resource- based models by effectively simulating the
direct impacts of disturbances and ongoing dynamics in ecosys-
tems (He and Mladenoff1999; Guiot and Cramer2016; Baudena
etal.2020; Holdo and Nippert2023). With their detailed focus
on plant functional traits and responses to disturbances, these
models provide more realistic predictions, especially in eco-
systems frequently exposed to disturbance, compared to dy-
namic global vegetation models and gap models (Pausas 1999;
Bugmann2001; Risch, Heiri, and Bugmann2005). For instance,
the FATELAN D model, which has been rigorously tested in f ire-
prone Mediterranean ecosystems (Pausas2006; Pausas, Lloret,
and Vila 2006; Pausas and Lloret 2007; Bahar 2018), effec-
tively represents vegetation mechanics in these environments
(Millington etal.2009). Such models allow us to predict long-
term vegetation dynamics under various disturbance regimes,
including grazing intensity and provide valuable insights for
future ecological forecasting for conservation and management.
In this study, we aim to unravel the long- term dynamics of
Central Anatolian steppes using the disturbance- based model-
ing approach utilizing plant functional traits. Given that graz-
ing is the primary disturbance factor in the Central Anatolian
steppe ecosystem, and considering the variety of vegetation
types in the region, our study included five distinct vegetation
types subjected to var ying grazing regimes. The overarching ob-
jective is to comprehensively understand the influence of graz-
ing on the structural development of vegetation. Additionally,
our study aims to examine the possible occurrence of different
vegetation states under the differential pressures imposed by
various grazing regimes. Based on our knowledge from other
grazing- meditated grassland ecosystems, we hypothesized that
moderate grazing simulation would support the stability of the
studied vegetation by maintaining growth form diversity, while
simulations of no grazing and overgrazing would have nega-
tive effects, potentially leading to vegetation state changes in
the latter scenarios. We also expected different plant functional
groups to respond differently to various grazing regime simu-
lations. To test these hypotheses, we used the spatially explicit
FATELAND model to predict 50 years of vegetation dynamics
under alternative grazing regimes by examining the dynamics
of plant functional groups with varying traits related to growth,
reproduction, and response to grazing.
2 | Methods
2.1 | Study Area
The study area is located in Central Anatolia, Türkiye,
within the Irano- Anatolian biodiversity hotspot (Mittermeier
etal.2005; Şekercioğlu etal.2011), and covers two ecoregions,
namely Central Anatolian Steppes and Central Anatolian
Woodlands and Steppes (Figure1). Most of Central Anatolia
is a plain with an elevation range between 750 and 1250 m,
but there are also several volcanic mountains exceeding
2000 m. This region has a semi- arid climate characterized
by cold winters and warm, dry summers, with annual pre-
cipitation ranging from 300 to 650 mm, and annual average
temperature varies between 7°C and 13°C across the region
(Bayer- Altın 2023). The diversity in bedrock types includes
volcanic (andesite, basalt, tuff, and agglomerate), rhyolite,
ignimbrite, radiolarite, flysch, marly, serpentine, calcareous,
gypsum, limestone, and rocky slopes. Correspondingly, vari-
ous soil types occur on these geological materials. This spatial
variation in bedrock and soil types significantly influences
the diversity of vegetation types in the area. The vegetation
types in the Central Anatolian steppe are primarily catego-
rized as grasslands dominated by grassy plants, dry steppes
dominated by chamaephyte shrubs, forest steppes, and sa-
line steppes (Kurt, Nilhan, and Ketenoglu 2006; Kürschner

4 of 16 Ecology and Evolution, 2024
and Parolly 2012). The diversity of different growth forms,
long- standing disturbance effect, and climatic variability also
create a trait diversity among plants of Anatolian grasslands,
including the Central Anatolian steppe (Ülgen 2019; Ülgen
and Tavşanoğlu2024).
2.2 | The Model and Simulation Scenarios
We employed the spatially explicit, grid- based FATELAND
model (Pausas and Ramos 2006; accessible at https:// www.
uv. es/ jgpau sas/ lass. htm) to simulate long- term vegetation dy-
namics in our studied ecosystem. FATELAND incorporates
disturbances and responses along with competition between
plant species or functional groups. The model operates on
annual time steps, simulating plant cohorts transitioning
through discrete stages: propagules, seedlings, immature,
and mature plants. Each grid cell can contain multiple spe-
cies (Pausas and Ramos 2006), with abundance measured
on a qualitative scale, categorized as absent, low, medium,
or high within each cell. Survival within the model is rep-
resented by a matrix that accounts for survival probabilities
across various life stages under different resource levels (low,
medium, and high). The model also incorporates the effect of
stratum (or height) on species dynamics, where taller plants
have a competitive advantage in accessing resources, influ-
encing their survival rates and overall abundance (Pausas and
Ramos2006). Survival is represented by a matrix of nine ele-
ments, where each element indicates whether survival occurs
(1 for yes and 0 for no) across three life stages—germinants
(seedlings), immatures, and matures—evaluated under three
different resource levels: low, medium, and high. The default
matrix assumes universal survival across all stages and re-
source levels. FATELAND is primarily a deterministic model,
with the exception of its dispersal module, which introduces
stochastic elements (Pausas and Ramos2006). These mechan-
ics allow the model to capture the complexity of plant com-
munity dynamics over time, reflecting the impact of resource
availability and disturbances on population structure and dis-
tribution across the landscape.
We created three grazing regime scenarios: grazing exclusion
(no grazing), moderate- intensity grazing, and overgrazing.
Disturbance events were arranged to occur annually. The “no
grazing” scenario assumes the exclusion of grazing activity from
the landscape for the entire simulation period. The moderate-
intensity grazing scenario represents a sustainable grazing re-
gime, operating within the carrying capacity of rangelands. In
contrast, the overgrazing scenario simulates heavy pressure
on the vegetation, with livestock density approximately four
times higher than the carrying capacity, reflecting conditions
observed in many parts of Central Anatolia (Gintzburger, Le
Houérou, and Saïdi2006). Each scenario was run over a 50- year
period to capture both intermediate- and long- term dynamics in
the vegetation.
2.3 | Creation of Initial Landscapes in the Model
To assess distinct initial landscape structures for modeling
purposes, we created a dataset containing abundance data
for plant species from phytosociological studies conducted in
the region. The dataset comprises studies conducted in steppe
and woodland- steppe vegetation in Central Anatolia and its
FIGUR E  | The map of the studied ecoregions; Central Anatolian steppes (marked in yellow) and Central Anatolian woodlands and steppes
(marked in green), and the locations of vegetation data gathered from phytosociological studies used to create initial landscapes for the model. The
ecoregion map is based on Olson etal.(2001).

5 of 16
surroundings from 1961 to 2020 (Figure1). These studies cover
various vegetation types (i.e., alliances) commonly found in
Central Anatolia and reflect the abundance and cover of plant
species in the region. Additionally, the dataset includes family,
genus and species names, growth forms, locality names, and co-
ordinates where each study was conducted, and the reference
(or cited original reference) for each study (Appensix S2). As the
cover/abundance of plant species was measured in the field by
following the Braun- Blanquet method (Braun- Blanquet1932) in
these studies, we converted these data to percentages for model
use. In total, data from 668 relevés, 58 alliances, and 13 phy-
tosociological studies across the Central Anatolian plain were
compiled to create this dataset (Figure1, AppendixS2).
We then transformed the species- level abundance data into a
new dataset by combining individual species trait data based on
growth form, using plant trait databases such as TRY (Kattge
etal. 2020) and BROT (Tavşanoğlu and Pausas 2018), as well
as the published flora of Türkiye (Davis1965–1985) and online
flora sources such as World Flora Online (Borsch et al.2020),
TÜBİVES (Bakış, Babaç, and Uslu2011), and numerous other
online flora and herbarium websites (TableS1). This data com-
pilation approach allowed for a detailed categorization of species
according to their specific growth forms and traits. As a result,
we identified five growth forms: tree, subshrub, perennial forb,
perennial gramineae, and annual herb. We distinguished be-
tween perennial forbs and gramineae due to their known differ-
ential response to grazing (Fulbright et al. 2021; Gebremedhn
etal.2023). Furthermore, we divided the subshrub category into
non- spinous and spiny subshrubs by including the spinescence
trait, as spinescence offers varying degrees of resistance to graz-
ing. This categorization was based on the presence of spines,
thorns, or prickles on the leaves or stems of subshrubs (based on
Davis1965–1985; Ülgen2019; Kattge etal.2020), placing them
in the spiny subshrubs category. Note that trees in our model
represent resprouting species such as oaks (Quercus spp.) and
Pyrus elaeagnifolia as characterized by Central Anatolian wood-
land steppe (Kenar and Kikvidze2019).
We determined the plant abundances in each functional group
for each relevé by utilizing the maximum values of these
functional groups' abundances (Appendix S3). Based on the
abundance of growth forms in each type of vegetation, we
performed a cluster analysis to consolidate 58 vegetation al-
liances into distinct vegetation types. For this, we employed
the Elbow method, a technique used to determine the opti-
mal number of clusters in the k- means clustering algorithm
(FigureS1; Jain2010; Shi etal.2021). This analysis resulted in
the classification of all vegetation alliances under five distinct
vegetation types: (1) tree- dominated woodland steppe (here-
after “Landscape 1”), (2) woodland steppe with less abundant
trees (hereafter, “Landscape 2”), (3) herbaceous- dominated
steppe with a high abundance of non- spiny subshrubs (hereaf-
ter, “Landscape 3”), (4) non- spiny subshrub- dominated steppe
with low total abundance (hereafter, “Landscape 4”), and (5)
spiny shrub- dominated steppe (hereafter, “Landscape 5”).
Although both Landscape 1 and Landscape 2 can be classified
as woodland steppe, the main difference between these two
vegetation types was the total abundance of trees, with mean
tree abundance being approximately as 50% in Landscape 1%
and 10% in Landscape 2.
Finally, we created an initial landscape consisting of 10,000 cells
(100 × 100 cells) in the FATELAND model for each landscape
representing different vegetation types. We assumed that each
cell measures 10 m × 10 m; therefore, each landscape covers a
total area of 1000 m × 10 00 m.
2.4 | Traits and Response to Grazing
We used several functional traits to inform about the life his-
tory of each functional group while simulating vegetation dy-
namics in the FATELAND model (Tables1–3). In the model,
most traits and response parameters to disturbances were
included as semiquantitative data (e.g., low- medium- high)
or Boolean data (no/yes), while some quantitative traits were
also presented. The model's reliance on the functional traits
of various species or functional groups is informed by plant
trait databases and field experience in the Central Anatolian
steppes. Priority is given to the most abundant species in these
databases, with traits and parameters detailed in Table1. We
randomly distributed all plants (at seed, immature, and ma-
ture stages) across the landscapes to establish their initial
abundance. The FATELAND model initiates its simulation
by defining the lifespan of functional groups through mature
age and maximum age traits, thereby establishing their long-
term life histories (Tables 1 and 3). Seed dispersal distance
and rate are determined based on their dispersal capabilities
(Tables 1 and 3), which are crucial for incorporating spatial
dynamics. The seed germination process is characterized by
the fecundity trait, while the transformation of these seeds
into immature individuals is governed by the germination
trait (Tables1 and 2). Thus, seed germination and the early
development stages set the stage for how functional traits in-
fluence vegetation dynamics. Once seeds transform into im-
mature individuals, the FATELAND model further examines
how these growing plants interact with their environment.
This progression from seed to mature plant is critical as it
underscores the transition phases that directly impact plant
survival and distribution (Pausas and Ramos 2006). Plant
height is another crucial trait in the model that significantly
influences the grazing response of functional groups. Taller
growth forms, such as trees, have more resistance to grazing,
particularly after they grow tall enough to escape the grazing
zone. Therefore, the grazing responses of functional groups
vary based on their maturity (immature versus mature life
stages; Tables 1 and 3). This distinction, determined by the
designated age for each group, indicates that mature individu-
als tend to be more resistant to grazing, while their immature
counterparts are often more susceptible. Similarly, larger (or
adult) plants have an advantage in resource acquisition, both
belowground (root capacity) and aboveground (sunlight cap-
ture), compared to smaller (or immature) ones. This results in
larger and adult plants having a higher resilience capacity to
grazing than smaller or younger individuals (Tables1 and 3).
Consequently, both age class and height are represented in the
disturbance (i.e., grazing) response of each functional group
through the parameters “age limit,” “killresp,” and “respage”
(Tables 1 and 3). These parameters and traits help model the
differential impact of grazing on plants based on their size and
maturity, reflecting more complex vegetation dynamics under
various grazing pressures.

6 of 16 Ecology and Evolution, 2024
2.5 | Model Outputs and Data Analysis
We run each scenario once due to the deterministic nature of
the FATELAND model. We obtained three main outputs after
running the model for 50 years of simulation. The first out-
put includes the abundance of each functional group under
each scenario and landscape for each year during the 50- year
simulation (Appendix S4). We summarized these outputs to
depict the trends of abundance changes over the 50 years. The
second output is the final abundance of each functional group
after 50 years of simulation under each scenario (AppendixS4),
visualized as mean plots for each scenario and landscape. The
third is a visual representation of the vegetation structure from a
bird's- eye view, illustrating the structure of the landscape at the
TABLE  | Parameters and traits used in the FATEL AND model (sensu Pausas and Ramos2006). The main parameters besides disturbance- and
germination- related ones, along with their descriptions and categories, are given. Traits w ithout categories are quantitative ones.
Parameters Description Categories
Main parameters
Maxab The maximum number of species in each cell in
the landscape (1: low; 2: medium; and 3: high)
1–2- 3
Mature age Age at which it can produce seeds or shoots —
Max age Lifespan —
Size The size of immature plants relative to mature
plants in terms of height (1: small proportion;
2: half; 3: most; and 4: same height)
1–2–3–4
Stratum Maximum canopy layer (height) attainable by adult
individuals (1–5; from ground level to the high canopy)
1–2–3–4–5
S Disp Short- distance dispersal capacity No- Low- Med- High
M Disp Medium- distance dispersal capacity No- Low- Med- High
H Disp Long- distance dispersal capacity No- Low- Med- High
K Disp Rate of decrease in dispersal curve
from medium distance onward
—
Limit Spatial extent of short, medium, and long
dispersal distances in meters
—
Fecund Number of seeds opened at a randomly
selected distance each year
—
Disturbance parameters
Age limit Age group affected by intervention (e.g.,
separate for mature and immatures)
—
Seed broken Seedling emergence rate after intervention No- Low- Med- High- All
Propkill Ratio of seeds and propagules killed
during the inter vention
No- Few- Half- Most- All
Killresp Survival and regeneration capacity of
individuals after intervention
No- Few- Half- Most- All
Respage Functional age of shoots after intervention —
Germination parameters
Germination rate Germination ability under low, medium,
and high resource availability
No- Low- Med- High- All
Survival of germinants Survival ability of seedlings under low,
medium, and high resource availability
No - Ye s
Survival of immatures Survival ability of non- adult individuals under
low, medium, and high resource availability
No - Ye s
Survival of matures Survival ability of adult individuals under low,
medium, and high resource availability
No - Ye s

7 of 16
initial stage (before the model run) and at the final stage (after
50 years of the model run). These outputs enable us to interpret
the different responses of various vegetation types in the Central
Anatolian steppes (represented by initial landscapes) under var-
ious grazing regimes.
We collected raw data for a 50- year simulation using the
FATELAND model and conducted subsequent data cleaning,
plotting, and analyses in the R environment (R Core Team2023;
Supplementary R code). We identified the optimal number of
clusters using the Elbow method and PCA graphs using the
“cluster” package (Maechler etal.2023).
3 | Results
Each landscape representing different vegetation types
showed different trajectories over 50 years of simulation of
vegetation dynamics under various grazing regimes. These
long- term dynamics represented vegetation state shifts in
some of the simulated landscapes under specific grazing re-
gimes, while some resulted in no considerable change in vege-
tation structure (Figures2–4).
In woodland steppes of Landscapes 1 and 2, the long- term
abundance of trees showed similar trends in response to spe-
cific grazing regimes with no change in moderate grazing
and overgrazing regimes and a significant increase in the no-
grazing scenario (Figures3 and 5). However, other functional
groups in these two landscapes that differ in the initial abun-
dance of these groups showed different trends in response to
various grazing regimes. In Landscape 1, the abundance of
these groups notably increased in response to moderate graz-
ing, but conversely, when grazing was excluded, the increase
in their abundance was suppressed by trees in the long- term,
with the exception of gramineae, which showed a slow but
steady increase in abundance (Figure 5). A similar increase
was also observed in Graminae, spiny subshrub, and perennial
forb groups in Landscape 2 under moderate grazing regime
but relatively in a less extent due to high initial abundances
of these groups in this landscape (Figure 5). Notably, the
abundance of non- spiny subshrubs stabilized after a decade of
decline in Landscape 2, but at the final stage had similar abun-
dance value as in Landscape 1 under moderate grazing regime
(Figures3 and 5). In the no- grazing scenario, the abundance
of all functional groups except trees decreased in the long term
in Landscape 2, too (Figures3 and 5). Overgrazing resulted in
differential responses in various functional groups in wood-
land steppes. Either in the case of a low (Landscape 1) or high
(Landscape 2) initial abundance of subshrubs (both spiny and
non- spiny ones) and perennial forbs, these functional groups
were completely lost after 50 years of overgrazing simulation
(Figure s 3 and 5). On the other hand, trees and gramineae
kept their initial abundance and were not affected by over-
grazing at all, and annual herbs showed a slight increase in
their abundance over time (Figures3 and 5).
Different functional groups showed different trends under
tested grazing regimes in treeless steppe vegetations of
Landscapes 3, 4, and 5. In all these landscapes, peren-
nial forbs, gramineae, and spiny subshrubs increased their
abundance over time under the moderate grazing regime
(Figure 6), resulting in their high abundance at the end of
the 50 years of simulation (Figures4 and 6). Under moderate
TABLE  | Germination and seedling survival response of functional groups included in the study for different grazing regimes. N and Y mean
“No” and “Yes,” respectively, and represent the seedl ing survival possibility in binar y form under a particula r resource amount for each grow th form.
Trait Low Medium High Trait Low Medium High
Non- spiny subshrub Resource Perennial gramineae Resource
Germination rate None Low Medium Germination rate Low Low Low
Survival of germinants N N Y Survival of germinants N Y Y
Survival of immatures N Y Y Survival of immatures Y Y Y
Survival of matures Y Y Y Survival of matures Y Y Y
Spiny subshrub Resource Annuals Resource
Germination rate 0 None 1 Low 2
Medium
Germination rate 1 Low 1 Low 1 Low
Survival of germinants N N Y Survival of germinants N N Y
Survival of immatures N Y Y Survival of immatures N N Y
Survival of matures Y Y Y Survival of matures N Y Y
Perennial forb Resource Resprouter tree Resource
Germination rate 1 Low 1 Low 1 Low Germination rate 1 Low 2 Medium 3 High
Survival of germinants N Y Y Survival of germinants N Y Y
Survival of immatures Y Y Y Survival of immatures Y Y Y
Survival of matures Y Y Y Survival of matures Y Y Y

8 of 16 Ecology and Evolution, 2024
grazing, non- spiny subshrubs showed an increase or decrease
depending on their initial abundance, ended up with a simi-
lar abundance in all three landscapes within 20 years of sim-
ulation, and their abundance remained stable until the end
of the simulation (Figure 6). Grazing exclusion promoted the
expansion of both spiny and non- spiny subshrubs as their
abundance significantly increases over time. Spiny subshrubs
showed the same trends between no grazing and moderate
grazing scenarios, but a considerable difference between these
two scenarios was observed in non- spiny shrubs, which were
suppressed by moderate grazing but favored by grazing exclu-
sion (Figure6). In contrast, the abundance of perennial forbs
and gramineae were suppressed in no grazing scenario in
comparison with moderate grazing regime (Figures4 and 6).
As in landscapes representing woodland steppes, overgrazing
led to the complete removal of spiny and non- spiny subshrubs
and perennial forbs, while annual herbs and gramineae main-
tained their initial abundance over 50 years under the over-
grazing scenario in treeless steppe vegetation (Figures4 and
6). In treeless steppe vegetation types, annual herbs showed
the same trend under all three grazing regimes. In each graz-
ing regime, the abundance of annual herbs showed a slow but
steady increase when their initial abundance was relatively
lower (Landscapes 4 and 5) and remained stable through
time in the case of higher initial abundance (Landscape 3)
(Figure6).
4 | Discussion
Our model results covering 50 years of simulations suggest
drastic changes in vegetation structure under different graz-
ing regimes in Central Anatolian steppes. Moderate grazing
regimes promote the growth form diversity and cover of var-
ious functional groups except resprouter trees. In contrast,
overgrazing and no grazing scenarios change the vegetation
state in the long term. Woodland steppes tend to transform
into closed woodlands under no grazing regime, while steppes
without trees into dense scrubland. Overgrazing results in the
total loss of subshrubs (both spiny and non- spiny) and peren-
nial forbs in the long term, while trees (if they exist) are not
affected by overgrazing and gramineae and annual herbs pro-
moted by overgrazing. As a result, overgrazing does not cause
a vegetation state change in woodland steppes, at least within
the 50- year simulation period, but steppes without trees trans-
form into grassland dominated by resprouter gramineae if
their initial abundance is not very low. Resprouter gramineae
appear to be crucial for the resilience of the plant commu-
nity under overgrazing pressure, as their initial abundance
remained stable across all landscape in the overgrazing sce-
nario. On the other hand, if resprouter gramineae are absent
or have low abundance in the initial vegetation, the landscape
transforms into an annual herb- dominated grassland under
the overgrazing scenario.
TABLE  | Main and disturbance parameters that are assigned to functional groups in the FATEL AND model in the study.
Traits
Non- spiny
subshrub
Spiny
subshrub
Forb
(perennial)
Gramineae
(perennial) Annuals
Resprouter
tree
Main traits
Max abundance 1 1 1 1 1 1
Mature age 2 2 1 2 1 10
Max age 30 30 5 5 2 300
Size 1 1 2 2 4 1
Stratum 1, 2 1, 2 1, 1 1, 1 1, 1 1, 3
S Disp High High High High High High
M Disp Low Low Med Low Low High
H Disp Low Low Low Low No High
K Disp 2, 10 2, 10 3, 20 3, 10 2 , 10 2, 10
Limit (m) 5, 10, 50 5, 10, 50 5, 30, 100 5, 10, 50 5, 10, 20 10, 50, 10 0
Fecund 3 3 3 3 5 2
Disturbance traits
Age limit 2, 5 2, 5 2 1 1 3, 10
Seed broken 0 0 0 0 0 0
Propkill 0 0 0 0 0 0
Killresp (3, 3, 1)
(0, 1, 2)
(3, 2, 0)
(0, 1, 1)
(3, 1)
(0, 2)
(3, 0)
(0, 3)
(1, 3)
(0, 0)
(3, 2, 0)
(0, 1, 0)
Respage (0, 1, 2) (0, 1, 2) (1, 2) (1, 2) (−1,- 1) (1, 5, 10)

9 of 16
FIGUR E  | The illustration of the initial and at 50- year final landscapes for each vegetation type under various grazing regimes. The simulated
landscape consists of 10,000 cells (100 × 100 cells) covering hypothetically 1000 m × 1000 m total area in size, each representing a 10 m × 10 m area.
The color in the cel ls indicates the dominanc e of a specific funct ional gr oup in each cel l: green for trees, red for s ubshrubs, purple for spiny sub shrubs,
orange for perennial forbs, blue for perennial gramineae, and gray for annual herbs. The white color represents empty (no vegetation) cells.
FIGUR E  | Final abundances of f unctional groups in the wood land steppe vegetation ty pes (Landscapes 1 and 2) after a 50 - yea r simulation under
various grazing regimes.

10 of 16 Ecology and Evolution, 2024
Evidence from open ecosystems such as grasslands, savannas,
and woodland steppes suggests that a lack of long- term graz-
ing leads to a shift toward more closed woodland vegetation
(Bernardi et al. 2019; Bond 2019; Wolański et al. 2021). Our
model results support this conclusion for Central Anatolian
woodland steppes, showing that a lack of grazing causes a shift
from open woodland steppe to closed woodland, regardless of
the initial abundance of trees (whether semi- open or open wood-
land). In contrast, our models also indicate that moderate graz-
ing promotes the abundance and diversity of functional groups
by maintaining the initial state of the vegetation. Enhanced di-
versity through moderate grazing has been observed in many
herbivory- mediated grassland ecosystems worldwide (Török
etal.2016; Joubert, Pryke, and Samways2017; Davies etal.2024).
In savannas, moderate grazing fosters biodiversity by reducing
the dominance of certain grass species, allowing less competitive
species to thrive (Sankaran, Ratnam, and Hanan 2008), while
overgrazing leads to a marked decline in perennial forbs and
some subshrubs. Conversely, North American prairies, which
have coevolved with native ungulates like bison, show a differ-
ent pattern. In these ecosystems, grazing helps maintain grass-
land structure by preventing woody plant encroachment (Knapp
etal. 1999). In the Mongolian steppes, overgrazing results in a
considerable loss of biomass and diversity, especially in dry and
high mountain steppes, while species richness increases under
moderate grazing in more mesic steppes (Munkhzul etal.2021).
In many cases, the positive relationship between plant diversity
and livestock grazing can be attributed to a long- term history of
large mammal herbivory in these ecosystems. For example, spi-
nescence is a plant trait that serves as structural anti- herbivore
defense (Atkinson et al. 2024), and the origin of this trait can
be traced back through geological history, when large herbivore
mammals evolved and grazed these ecosystems (Lauenroth1998;
Charles- Dominique etal.2016). In our models, spinescence also
emerged as a significant trait, where dominance of spiny or
non- spiny shrubs shifted notably under moderate grazing sce-
narios. The dominance of spiny shrubs in grazed steppe areas
in Central Anatolia (Vural and Adıgüzel2006; Kürschner and
Parol ly 2012; Tavşanoğlu 2017; Ülgen 2019) and other regions
(Lauenroth1998; Rahmanian etal.2019; Atkinson etal.2024) is
frequently observed. The presence of clumps of grazing- resistant
spiny plants in steppe areas may have additional ecological im-
portance, as they can provide refuges for many grazing- sensitive
species (Rebollo et al. 2002). Our model results, showing the
dominance of spiny shrubs over non- spiny ones under moderate
grazing, suggest that the presence of spiny shrubs in a location
in Central Anatolia can be used as a proxy for moderate grazing
pressure, while loss of both spiny and non- spiny shrubs can indi-
cate overgrazing. However, it is important to note that past land
use may also shape the current vegetation structure, possibly
in combination with various grazing regimes. Therefore, in the
Central Anatolian steppes, the response of vegetation to grazing
is complex, likely influenced by the region's harsh climate, past
land use, and historical grazing patterns.
In our study, overgrazing emerged as a critical driver of vege-
tation change in Central Anatolia, significantly reducing the
abundance of all growth forms across all landscapes, except
for trees in woodland steppes, which can contribute to bio-
diversity loss. The Eurasian steppe, specifically in Mongolia
FIGUR E  | Final abundances of functional groups in the steppe (treeless) vegetation types (Landscapes 3, 4, and 5) after a 50- year simulation
under various grazing regimes.

11 of 16
and Tibet regions, characterized by a harsher climate, exhib-
its even more drastic effects of grazing with rapid vegetation
degradation and soil erosion when subjected to overgrazing
(Wesche etal.2 016; Munkhzul etal.2021). Our results high-
light the necessity for adapted grazing management strategies
that take into account the distinct ecological characteristics
of the Central Anatolian steppes to prevent habitat degrada-
tion and encourage plant diversity. Consequently, the dynam-
ics of woodland steppes highlight the resilience of trees and
certain grass species under grazing pressures, which contrasts
sharply with the vulnerability of perennial forbs and sub-
shrubs. This differential response underlines the importance
of understanding species- specific traits and life histories
when developing conservation strategies. For instance, the
adaptability of gramineae and annual herbs to quickly regen-
erate makes them resilient under varying grazing intensities,
suggesting that management practices need to be species and
context- specific to enhance ecosystem resilience and prevent
biodiversity loss. This phenomenon aligns with the concept
of disturbance- mediated biodiversity, where moderate levels
of disturbance, such as grazing, can prevent any single spe-
cies from dominating an ecosystem, thereby maintaining
higher species richness and ecological integrity in the plant
community (Milchunas, Sala, and Lauenroth1988; Milchunas
and Lauenroth 1993). Moreover, rotational grazing systems
can help quickly regenerating functional groups, increasing
rangeland stability and resilience, especially in areas where
grazing pressure on vegetation is expected to be high (Briske
etal.2008; de Otálora etal.2021; Jordon etal.2022). In this
way, overgrazing can be prevented or reduced by avoiding
practices that return livestock to the same area too soon and
by controlling livestock numbers to ensure they do not exceed
the rangeland's carrying capacity.
Under the overgrazing scenario in our models, annuals were
positively or not affected as their cover increased with time espe-
cially if their abundance was low in the initial landscape. Similar
results have been obtained from many studies with increasing
disturbance frequency or intensity; as in increased grazing in-
tensity transform plant communities to annual- dominated ones
in grasslands in Britain (Pakeman 2004), in high disturbance
frequency in an artificial disturbance experiment in the Central
Anatolian steppe (Özüdoğru, Özüdoğru, and Tavşanoğlu2021),
under heavy- plowing treatment in regenerating pine forests
in an eastern Mediterranean ecosystem (Ürker, Tavşanoğlu,
and Gürkan 2018). A meta- analysis also suggests that, on a
global scale, grazing favors annual plants over perennials (Díaz
etal.2007). The life cycle and life history traits of annuals seem
to be the main reasons for their resilience to high- intensity dis-
turbances, in our case overgrazing, as they are seeder species that
are able to establish their seedlings. In overgrazing scenarios, the
increase in the abundance of annuals could also be attributed to
the total vanishing of other competitor functional groups, namely
perennial forbs and subshrubs. However, contrasting results sug-
gest a decrease in the proportion of annuals under overgrazing
regimes in semi- arid grasslands (Rahmanian et al. 2019). Our
FIGUR E  | The long- term abundance trend of functional groups in woodland steppes under varying grazing reg imes over 50 years. Above panel:
Landscape 1, below panel: Landscape 2. Different colors represent the abundance of different functional groups over a 50- year period. The names
of six functional groups (tree, non- spiny subshrub, spiny subshrub, perennial gramineae, perennial forb, and annuals) are written on the curves.

12 of 16 Ecology and Evolution, 2024
results also suggest that if gramineae exists in moderate abun-
dance in the initial landscape, annuals and perennial gramineae
can coexist under the overgrazing pressure. In ecosystems where
herbivory pressure is managed effectively, grazing can contrib-
ute to biodiversity conservation by creating niches for a variety of
plant species, thus promoting a more diverse and resilient ecosys-
tem. For instance, in the savannas of Africa, controlled grazing
has been shown to reduce the dominance of aggressive grass spe-
cies, allowing for the proliferation of forbs and other grass spe-
cies, which contributes to overall biodiversity (Fuhlendorf and
Engle 2001; Scott- Shaw and Morris2015). Similarly, research in
North American prairies has demonstrated that when executed
with conservation goals in mind, grazing supports the mainte-
nance of plant diversity by mimicking natural herbivory patterns
that existed prior to extensive human intervention (Knapp
etal.1999). However, the relationship between grazing and bio-
diversity is not straightforward and depends heavily on the graz-
ing intensity and the specific ecological context. Overgrazing, as
observed in some sections of the Central Anatolian steppes, leads
to significant degradation of plant communities, reducing plant
cover and biodiversity (Fırıncıoğlu, Seefeldt, and Şahin 2007;
Kürschner and Parolly2012; Ambarlı et al. 2016; Rahmanian
etal.2019). This negative outcome underscores the need for im-
plementing grazing regimes that consider ecological thresholds
and are tailored to the carrying capacity of the landscape.
Results from the FATELAND model are based on simula-
tions assuming the current climatic conditions. Therefore, we
FIGUR E  | The long- term abundance trend of functional groups in steppe vegetation types without trees under varying grazing regimes over
50 years. Above panel: Landscape 3; middle panel: Landscape 4; and below panel: Landscape 5. Different colors represent the abundance of different
functional groups over a 50- year period. The names of five functional groups (non- spiny subshrub, spiny subshrub, perennial gramineae, perennial
forb, and annuals) are written on the curves.

13 of 16
should note here that long- term vegetation dynamics may re-
sult in different directions if the response of species to chang-
ing climate is also considered. The ongoing climatic change
may increase uncertainties about the fate of vegetation dy-
namics under various grazing regimes, especially in the long
term. While climate plays a role in shaping woodland commu-
nities in Central Anatolia (Kenar and Kikvidze2019), the five
landscapes used in our models also ref lect the land use history
and past grazing regimes, not just climate. Thus, our results
should be interpreted in the context of different past land uses
under relatively stable climatic conditions. We also did not
include fire as a disturbance factor in our models, focusing
solely on grazing regimes. This decision was based on the lack
of wildfires in Central Anatolian grasslands over a long pe-
riod due to the absence of continuous fuel, resulting from mil-
lennia of domestic grazing and agricultural activity, as well
as limited knowledge of fire response in Central Anatolian
steppe plants. However, it is worth to noting that the Holocene
fire regimes of the region could re- emerge due to land aban-
donment and climate change in the future (Tavşanoğlu2017).
Therefore, the response of Central Anatolian vegetation inter-
actions between grazing and wildfire should be a focus of fu-
ture research. Since our models are restricted to a 50 - year time
period, our finding of no vegetation state change in woodland
steppes under overgrazing may differ over longer periods, ex-
ceeding the lifespan of existing trees in these ecosystems. In
such cases, where no new seedling establishment is expected
under overgrazing, the long- term outcome could be the even-
tual loss of trees and a shift in vegetation state under centuries
of overgrazing.
In conclusion, our findings indicate the significant impact of
grazing on the vegetation dynamics of Central Anatolian steppes
and suggest the importance of adaptive management strategies
that consider both the ecological characteristics of the region.
Sustainable grazing practices, adjusted to the unique conditions
of the Central Anatolian steppes, are essential to sustain Central
Anatolian steppe vegetation states by preserving its plant func-
tional diversity.
Author Contributions
Anıl Bahar: data curation (lead), formal analysis (lead), methodology
(equal), writing – original draft (lead), writing – review and editing
(equa l). Çağatay Tavanoğlu: conceptualization (lead), formal anal-
ysis (supporting), methodology (equal), supervision (lead), writing – re-
view and editing (equal).
Acknowledgments
We are grateful to the two anonymous reviewers for their constructive
comments on this study. We also thank Ali Kavgacı and Özge Erişöz
Kasap for their comments on the study. Anıl Bahar acknowledges the
financial support of the 100/2000 doctoral scholarship by the Council
of Higher Education of Turkey. This study is a part of the Ph.D. dis-
sertation of the first author (Institute of Science, Hacettepe University,
Tü rki ye).
Conflicts of Interest
The authors declare no conflicts of interest.
Data Availability Statement
The vegetation dataset, abundance- cover values of growth forms for
each landscape, raw data of model outcomes, and R codes for analy-
ses and producing graphs are available as supplementary information.
Model inputs are presented in tables within the manuscript.
References
Adler, P., D. Raff, and W. Lauenroth. 2 001. “The Eff ect of Grazing on the
Spatial Heterogeneity of Vegetation.” Oecologia 128: 465–479.
Ambarlı, D., U. S. Zeydanlı, Ö. Balkız, etal. 2016. “An Overview of
Biodiversity and Conservation Status of Steppes of the Anatolian
Biogeographical Region.” Biodiversity and Conservation 25: 2491–2519.
Arbuckle, B. S. 2013. “The Late Adoption of Cattle and Pig Husbandry
in Neolithic Central Turkey.” Journal of Archaeological Science 40, no.
4: 1805–1815.
Asner, G. P., A. J. Elmore, L. P. Olander, R. E. Martin, and A. T. Harris.
2004. “Grazing Systems, Ecosystem Responses, and Global Change.”
Annual Review of Environment and Resources 29, no. 1: 261–299.
Atkinson, J., R. Gallagher, S. Czyżewski, et al. 2024. “Integrating
Functional Traits Into Trophic Rewilding Science.” Journal of Ecolog y
112, no. 5: 936–953.
Bahar, A. 2 018. Modelling of Fire Fr equency and Vegetation Cover Effects
on Mediterranean Vegetation Dynamics. M.Sc. Thesis, Ankara, Türkiye:
Hacettepe University. [in Turkish w ith English summary].
Bakış, Y., M. T. Babaç, and E. Uslu. 2011. “Updates and Improvements
of Turkish Plants Data Service (TÜBİVES).” In 6th International
Symposium on Health Informatics and Bioinformatics (HIBIT), Izmir,
Turkey: IEEE, 136–140.
Baudena, M., S. C. Dekker, P. M. van Bodegom, et al. 2015. “Forests,
Savannas, and Grasslands: Bridging the Knowledge Gap Between
Ecology and Dynamic Global Vegetation Models.” Biogeosciences 12:
1833–184 8.
Baudena, M., V. M. Santana, M. J. Baeza, etal. 2020. “Increased Aridity
Drives Post- Fire Recovery of Mediterranean Forests Towards Open
Shrublands.” New Phytologist 225, no. 4: 150 0–1515.
Bayer- Altın, T. 2023. “Long- Term Trends and Variability in Air
Temperatures of Central Anatolia Region, Turkey, for the Period 1960 –
2021.” Acta Geophysica 71, no. 4: 2031–2047.
Beisner, B. E., D. Haydon, and K. L. Cuddington. 2003. “Alternative
Stable States in Ecology.” Frontiers in Ecology and the Environment 1:
376–382.
Bernardi, R. E., M. Buddeberg, M. Arim, and M. Holmgren. 2019.
“Forests Expand as Livestock Pressure Declines in Subtropical South
America .” Ecology and Society 24, no. 2: 19.
Bestelmeyer, B. T., M. C. D uniway, D. K. James, L. M. Burkett, and K. M.
Havstad. 2013. “A Test of Critical Thresholds and Their Indicators in a
Desertification- Prone Ecosystem: More Resilience Than We Thought.”
Ecology Letters 16, no. 3: 339–345.
Bond, W. J. 2019. Open Ecosystems: Ecology and Evolution Beyond the
Forest Edge. Oxford: Oxford University Press.
Borsch, T., W. Berendsohn, E. Dalcin, etal. 2020. “World Flora Online:
Placing Taxonomists at the Heart of a Definitive and Comprehensive
Global Resource on the World's Plants.” Tax on 69: 1311–1341.
Braun- Blanquet, J. 1932. Plant Sociology: The Study of Plant
Communities. New York, NY: McGraw- Hill.
Briske, D. D., J. D. Derner, J. R. Brown, etal. 2008. “Rotational Grazing
on Rangelands: Reconciliation of Perception and Experimental
Evidence.” Rangeland Ecology & Management 61, no. 1: 3–17.

14 of 16 Ecology and Evolution, 2024
Bugmann, H. 2001. “A Review of Forest Gap Models.” Climatic Change
51, no. 3: 259–305.
Charles- Dominique, T., T. J. Davies, G. P. Hempson, etal. 2016. “Spiny
Plants, Mammal Browsers, and the Origin of African Savannas.” PNAS
113, no. 38: E5572–E5579.
Dantas, V. D. L., M. Hirota, R. S. Oliveira, and J. G. Pausas. 2016.
“Disturbance Maintains Alternative Biome States.” Ecology Letters 19,
no. 1: 12–19.
Davies, K. W., C. S. Boyd, J. D. Bates, L. N. Svejcar, and L. M. Porensky.
2024. “Ecological Benefits of Strategically Applied Livestock Grazing in
Sagebrush Communities.” Ecosphere 15, no. 5: e4859.
Davis, P. H., ed. 1965 –1985. Flora of Turkey and the East Aegean Islands.
Vol. 1–9. Edinburgh: Edinburgh University Press.
de Otálora, X. D., L. Epelde, J. Arranz, C. Garbisu, R. Ruiz, and N.
Mandaluniz. 2021. “Regenerative Rotational Grazing Management
of Dairy Sheep Increases Springtime Grass Production and Topsoil
Carbon Storage.” Ecological Indicators 125, no. 107: 484.
Díaz, S., S. Lavorel, S . U. E. McInt yre, etal. 2007. “Plant Trait Responses
to Grazing–A Global Synthesis.” Global Change Biology 13, no. 2:
313–341.
Ekim, T., and A. Güner. 2000. “The Floristic Richness of Turkey.”
Curtis's Botanical Magazine 17, no. 2: 48–59.
Evans, D. M., N. Villar, N. A. Littlewood, etal. 2015. “The Cascading
Impacts of Livestock Grazing in Upland Ecosystems: A 10- Year
Ex peri ment.” Ecosphere 6, no. 3: 1–15.
Fırı ncıoğlu, H. K., N. A dıgüzel, B. B ani, and B. Şah in. 2010. “Assessment
of Grazing Effect on Two Sub- Shrubs (Astragalus schottianus and
Thymus sipyleus) Dominated Mountain bozoğlan Grasslands in the
Semi- Arid Central- Southern Anatolian Region of Turkey.” Arid Land
Research and Management 24, no. 4: 282–300.
Fırıncıoğlu, H. K., S. S. Seefeldt, and B. Şahin. 2007. “The
Effects of Long- Term Grazing Exclosures on Range Plants in the
Central Anatolian Region of Turkey.” Environmental Management 39:
326–33 7.
Fırıncıoğlu, H. K., S. S. Seefeldt, B. Şahin, and M. Vural. 20 09.
“Assessment of Grazing Effect on Sheep Fescue (Festuca valesiaca)
Dominated Steppe Rangelands, in the Semi- Arid Central Anatolian
Region of Turkey.” Journal of Arid Environments 73, no. 12: 1149–1157.
Fuhlendorf, S. D., and D. M. Engle. 2001. “Restoring Heterogeneity on
Rangelands: Ecosystem Management Based on Evolutionary Grazing
Pat tern s.” Bioscience 51, no. 8: 625–632.
Fulbright, T. E., D. J. Drabek, J. A. Ortega- S, etal. 2021. “Forb Standing
Crop Response to Grazing and Precipitation.” Rangeland Ecolog y &
Management 79: 175 –185.
Gebremedhn, H. H., O. Ndiaye, S. Mensah, etal. 2023. “Grazing Effects
on Vegetation Dynamics in the Savannah Ecosystems of the Sahel.”
Ecological Processes 12: 54.
Gintzburger, G., H. N. Le Houérou, and S. Saïdi. 2006. “Near East–
West Asia Arid and Semiarid Rangelands.” Science et Changements
Planétaires/Sécheresse 17, no. 1: 152–168.
Gökbulak, F., B. U. Erdoğan, H. T. Yildirim, and M. S. Özçelik. 2018.
“Causes of Land Degradation and Rehabilitation Efforts of Rangelands
in Turkey.” Forest 68, no. 2: 106–113.
Guiot, J., and W. Cramer. 2016. “Climate Change: The 2015 Paris
Agreement Thresholds and Mediterranean Basin Ecosystems.” Science
354, no. 6311: 465–468.
Hammer, E. L., and B. S. Arbuckle. 2017. “10,000 Years of Pastoralism
in Anatolia: A Review of Evidence for Variability in Pastoral Lifeways.”
Nomadic Peoples 21, no. 2: 214–267.
Hayes, G. F., and K. D. Holl. 2003. “Cattle Grazing Impacts on Annual
Forbs and Vegetation Composition of Mesic Grasslands in California.”
Conservation Biology 17, no. 6: 1694–1702.
He, H. S., and D. J. Mladenoff. 1999. “The Effects of Seed Dispersal on
the Simulation of Long- Term Forest Landscape Change.” Ecosystems 2:
308–319.
Holdo, R. M., and J. B. Nippert. 2023. “Linking Resource- and
Disturbance- Based Models to Explain Tree–Grass Coexistence in
Savannas.” New Phytologist 237, no. 6: 1966–1979.
Isselstein, J., B. A. Griffith, P. Pradel, and S. Venerus. 20 07. “Effects of
Livestock Breed and Grazing Intensity on Biodiversity and Production
in Grazing Systems.” Grass and Forage Science 62, no. 2: 145–158.
Jain, A. K. 2010. “Data Clustering: 50 Years Beyond K- Means.” Pattern
Recognition Letters 31, no. 8: 651–666.
Jordon, M. W., K. J. Willis, P. C. Bürkner, and G. Petrokofsky. 2022.
“Rotational Gr azing and Multispecie s Herbal Leys Increa se Productivit y
in Temperate Pastoral Systems–A Meta- Analysis.” Agriculture,
Ecosystems & Environment 337, no. 108: 75.
Joubert, L., J. S. Pryke, and M. J. Samways. 2017. “Moderate Grazing
Sustai ns Plant Diversity in Af romontane Grassland.” Applied Vegetat ion
Science 20, no. 3: 340–351.
Kattge, J., G. Bönisch, S. Díaz, etal. 2020. “TRY Plant Trait Database—
Enhanced Coverage and Open Access.” Global Change Biology 26:
119–188.
Kenar, N., and Z. Kikvidze. 2019. “Climatic Drivers of Woody Species
Distribution in the Central Anatolian Forest- Steppe.” Journal of Arid
Environments 169: 34 –41.
Knapp, A. K., J. M. Blair, J. M. Briggs, et al. 1999. “The Keystone Role
of Bison in North American Tallgrass Prairie: Bison Increase Habitat
Heterogeneity and Alter a Broad Array of Plant, Community, and
Ecosystem Processes.” Bioscience 49, no. 1: 39–50.
Koc, A., W. H. Schacht, and H. I. Erkovan. 2015. “The History and
Current Direction of Rangeland Management in Turkey.” Rangelands
37, no. 1: 39–46.
Kürschner, H., and G. Parolly. 2012. “The Central Anatolian Steppe.”
In Eurasian Steppes. Ecological Problems and Livelihoods in a Changing
World, 149–171. Dordrecht: Springer Netherlands.
Kurt, L., G. N. Nilhan, and O. Ketenoglu. 2006. “Synoptic View of the
Steppe Vegetation of Central Anatolia (Turkey).” Asian Journal of Plant
Sciences 5: 733–739.
Lauenroth, W. K. 1998. “Guanacos, Spiny Shrubs and the Evolutionar y
History of Grazing in the Patagonian Steppe.” Ecología Austral 8, no.
2: 211–215.
Lavorel, S., S. McIntyre, J. Landsberg, and T. D. A. Forbes. 1997. “Plant
Functional Classifications: From General Groups to Specific Groups
Based on Response to Disturbance.” Trends in Ecology & Evolution 12,
no. 12: 474–478.
Maechler, M., P. Rousseeuw, A. Struyf, M. Hubert, and K. Hornik. 2023.
“cluster: Cluster Analysis Basics and Extensions.” https:// CRAN. R-
proje ct. org/ packa ge= cluster.
Metera, E., T. Sakowski, K. Słoniewski, and B. Romanowicz. 2010.
“Grazing as a Tool to Maintain Biodiversity of Grassland—A Review.”
Animal Science Papers and Reports 28: 315–334.
Middleton, C. 2018. “The Beginning of Herding and Animal
Management: The Early Development of Caprine Herding on the Konya
Plain, Central Anatolia.” Anatolian Studies 68: 1–31.
Milchunas, D. G. 2011. “Plant Community Dynamics in Shortgrass
Steppe With Grazing Relaxation and Imposition by Large and Small
Herbivores.” Israel Journal of Ecology & Evolution 57, no. 1–2: 23– 41.

15 of 16
Milchunas, D. G., and W. K. Lauenroth. 1993. “Quantitative Effects of
Grazing on Vegetation and Soils Over a Global Range of Environments:
Ecological Archives M063- 001.” Ecological Monographs 63, no. 4:
327–366.
Milchunas, D. G., O. E . Sala, and W. K. Lauenroth. 1988. “A Generalized
Model of the Effects of Grazing by Large Herbivores on Grassland
Community Structure.” American Naturalist 132, no. 1: 87–106.
Millington, J. D., J. Wainwright, G. L. Perry, R. Romero- Calcerrada,
and B. D. Malamud. 2009. “Modelling Mediterranean Landscape
Succession- Disturbance Dynamics: A Landscape Fire- Succession
Model.” Environmental Modelling & Software 24, no. 10: 1196 –1208.
Mittermeier, R. A., G. P. Robles, M. Hoffman, et al. 2005. Hotspots
Revisited: Earth's Biologically Richest and Most Endangered Terrestrial
Ecoregions. Washington, DC: Conservation International.
Mogashoa, R., P. Dlamini, and M. Gxasheka. 2021. “Grass Species
Richness Decreases Along a Woody Plant Encroachment Gradient in
a Semi- Arid Savanna Grassland, South Africa.” Landscape Ecology 36:
617– 636.
Munkhzul, O., K. Oyundelger, N. Narantuya, et al. 2021. “Grazing
Effects on Mongolian Steppe Vegetation—A Systemat ic Review of Local
Literature.” Frontiers in Ecolog y and Evolution 9, no. 703: 220.
Olson, D. M., E. Dinerstein, E. D. Wikramanayake, et al. 2001.
“Terrestrial Ecoregions of the World: A New Map of Life on Earth.”
Bioscience 51, no. 11: 933–938.
Özüdoğru, Ö., B. Özüdoğru, and Ç. Tavşanoğlu. 2021. “Recovery of a
Plant Community in the Central A natolian Steppe After Small- Scale
Disturbances.” Folia Geobotanica 56: 2 41–25 4.
Pakeman, R. J. 200 4. “Consistency of Plant Species and Trait Responses
to Grazing Along a Productivity Gradient: A Multi- Site Analysis.”
Journal of Ecology 92, no. 5: 893–905.
Pausas, J. G. 1999. “Mediterranean Vegetation Dynamics: Modelling
Problems and Functional Types.” Plant Ecology 140, no. 1: 27–39.
Pausas, J. G. 2006. “Simulating Mediterranean Landscape Pattern and
Vegetation Dyna mics Under Different Fire Regimes.” Plant Ecology 187:
249 –259.
Pausas, J. G., and W. J. Bond. 2020. “Alternative Biome States in
Terrestrial Ecosystems.” Trends in Plant Science 25, no. 3: 250 –263.
Pausas, J. G., and F. Lloret. 2007. “Spatial and Temporal Patterns of
Plant Functional Types Under Simulated Fire Regimes.” International
Journal of Wildland Fire 16, no. 4: 484–492.
Pausas, J. G., F. Lloret, and M. Vila. 2006. “Simulating the Effects of
Different Disturbance Regimes on Cortaderia selloana Invasion.”
Biological Conservation 128, no. 1: 128–135.
Pausas, J. G. , and J. I. Ramos. 20 06. “Landsc ape Analysis a nd Simulation
Shell (LASS).” Environmental Modelling & Software 21, no. 5: 629–639.
Payne, S. 1985. “Animal bones from Aşıklı Hüyük.” Anatolian Studies
35: 109–122.
Perkins, D., and P. Daly. 1968. “A Hunters' Village in Neolithic Turkey.”
Scientific American 219, no. 5: 96–109.
Peters, J., H. Buitenhuis, G. Grupe, K. Schmidt, and N. Pöllath.
2013. “The Long and Winding Road: Ungulate Exploitation and
Domestication in Early Neolithic Anatolia (10000–7000 Cal BC).” In
Origins and Spread of Domestic Animals in Southwest Asia and Europe,
83–114. New York: Routledge.
Porensky, L. M., K. E. Mueller, D. J. Augustine, and J. D. Derner. 2016.
“Thresholds and Gradients in a Semi- A rid Grassland: Long- Term
Grazing T reatments Induce Slow, Continuous and Reversible Vegetation
Change.” Journal of Applied Ecology 53, no. 4: 1013–1022.
R Core Team. 2023. R: A Language and Environment for Statistical
Computing. Vienna, Austria: R Foundation for Statistical Computing.
Rahmanian, S., M. Hejda, H. Ejtehadi, M. Farzam, F. Memariani, and
P. Pyšek. 2019. “Effects of Livestock Grazing on Soil, Plant Functional
Diversity, and Ecological Traits Vary Between Regions With Different
Climates in Northeastern Iran.” Ecology and Evolution 9, no. 14:
8225–8237.
Rebollo, S., D. G. Milchunas, I. Noy- Meir, and P. L. Chapman. 2002.
“The Role of a Spiny Plant Refuge in Structuring Grazed Shortgrass
Steppe Plant Communities.” Oikos 98, no. 1: 53–64.
Risch, A. C., C. Heiri, and H. Bugmann. 2005. “Simulating Structural
Forest Patterns With a Forest Gap Model: A Model Evaluation.”
Ecological Modelling 181, no. 2–3: 161–172.
Rosenthal, G., J. Schrautzer, and C. Eichberg. 2012. “Low- Intensity
Grazing With Domestic Herbivores: A Tool for Maintaining
and Restoring Plant Diversity in Temperate Europe.” Tuexenia 32:
167–205.
Sankaran, M., J. Ratnam, and N. Hanan. 2008. “Woody Cover in
African Savannas: The Role of Resources, Fire and Herbivory.” Global
Ecology and Biogeography 17, no. 2: 236–245.
Sankaran, M., J. Ratnam, and N. P. Hanan. 200 4. “Tree–Grass
Coexistence in Savannas Revisited—Insights From an Examination of
Assumptions and Mechanisms Invoked in Existing Models.” Ecology
Letters 7: 480–490.
Scott- Shaw, R., and C. D. Morris. 2015. “Grazing Depletes Forb Species
Diversity in the Mesic Grasslands of KwaZulu- Natal, South Africa.”
African Journal of Range & Forage Science 32, no. 1: 21–31.
Seidl, R., P. M. Fernandes, T. F. Fonseca, etal. 2011. “Modelling Natural
Disturbances in Forest Ecosystems: A Review.” Ecological Modelling
222, no. 4: 903–92 4.
Şekercioğlu, Ç. H ., S. Anderson, E. A kçay, eta l. 2011. “Turkey's Global ly
Important Biodiversity in Crisis.” Biological Conservation 144, no. 12:
2752–2769.
Shi, C., B . Wei, S. Wei, W. Wang, H. Liu, and J. Liu. 20 21. “A Q uantitative
Discriminant Method of Elbow Point for the Optimal Number of
Clusters in the Clustering Algorithm.” EURASIP Journal on Wireless
Communications and Networking 2021:31.
Tavşanoğlu, Ç. 2017. “Disturbance Regimes Proceeding in Anatolian
Steppe Ecosystems.” Kebikeç 43: 259–288. [in Turkish with English
abstract].
Tavşanoğlu, Ç., Ş. S. Çatav, and B. Özüdoğru. 2015. “Fire- Related
Germination and Early Seedling Growth in 21 Herbaceous
Species in Central Anat olian Steppe.” Journal of Ari d Environments 122:
109–116.
Tavşanoğ lu, Ç., and J. G. Pausas. 2018. “A Functional Trait Database for
Mediterranean Basin Plants.” Scientific Data 5, no. 180: 135.
Török, P., O. Valkó, B. Deák, A. Kelemen, E. Tóth, and B. Tóthmérész.
2016. “Managing for Species Composition or Diversity? Pastoral and
Free Grazing Systems in Alkali Steppes.” Agriculture, Ecosystems &
Environment 234: 23–30.
Tóth, E., B. Deák, O. Valkó, etal. 2018. “Livestock Type Is More Crucial
Than Grazing Intensity: Traditional Cattle and Sheep Grazing in Short-
Grass Steppes.” Land Degradation & Development 29, no. 2: 231–239.
Turner, R., N. Roberts, W. J. Eastwood, E. Jenkins, and A. Rosen. 2010.
“Fire, Climate and the Origins of Agriculture: Micro- Charcoal Records
of Biomass Burning During the Last Glacial–Interglacial Transition in
Southwest Asia.” Journal of Quaternary Science 25: 371–386 .
Twidwell, D., B. W. Allred, and S. D. Fuhlendorf. 2013. “National-
Scale Assessment of Ecological Content in the world's Largest Land
Management Framework.” Ecosphere 4, no. 8: 1–27.
Ülgen, C. 2019. Functional Trait Patterns of Plants of Anatolian Steppes.
M.Sc. Thesis, Ankara, Türkiye: Hacettepe University. [in Turkish with
English summary].

16 of 16 Ecology and Evolution, 2024
Ülgen, C., and Ç. Tavşanoğlu. 2024. “A Taxonomic Snapshot of
Belowground Or gans in Plants of Anat olian St eppes.” Folia Geobotanica
58: 231–2 43.
Ünal, S., and H. K. Fırıncıoğlu. 2007. “Variations in Some Wild
Medicago Populations of the Central Anatolian Highlands of Turkey.”
Tarla Bitkileri Merkez Araştırma Enstitüsü Dergisi 16, no. 1–2: 39–50.
Ürker, O., Ç. Tavşanoğlu, and B. Gürkan. 2018. “Post- Fire Recovery of
the Plant Community in Pinus Brutia Forests: Active Versus Indirect
Restoration Techniques After Salvage Logging.” iForest - Biogeosciences
and Forestry 11: 635– 642.
Van Langevelde, F., C. A. D. M. Van De Vijver, L. Kumar, et al. 2003.
“Effects of Fire and Herbivory on the Stability of Savanna Ecosystems.”
Ecology 84: 337–350.
Vural, M., and N. Adıgüzel. 2006. “Bozkırlar [Steppes].” In Türkiye'nin
Önemli Doğa Alanları [Turkey's Important Nature Areas], edited by G.
Eken, M. Bozdoğan, S. İsfendiyaroğlu, D. T. Kılıç, and Y. Lise, 28–30.
Ankara: Doğa Derneği.
Wesche, K., D. Ambarlı, J. Kamp, P. Török, J. Treiber, and J. Dengler.
2016. “The Palaearctic Steppe Biome: A New Synthesis.” Biodiversity
and Conservation 25, no. 12: 2197–2231.
Wiegleb, G. 1989. “Explanation and Prediction in Vegetation Science.”
Veget atio 83: 17–34.
Wolański, P., A. Bobiec, B . Ortyl , eta l. 2021. “ The Import ance of Livest ock
Grazing at Woodland- Grassland Interface in the Conservation of Rich
Oakwood Plant Communities in Temperate Europe.” Biodiversit y and
Conservation 30: 741–760.
Zainelabdeen, Y. M., R. Yan, X. Xin, Y. Yan, A. I. Ahmed, L. Hou, and
Y. Zhang. 2020. “The Impact of Grazing on the Grass Composition in
Temperate Grassland.” Agronomy 10, no. 9: 1230.
Zeder, M. A. 2008. “Domestication and Early Agriculture in the
Mediterranean Basin: Origins, Diffusion, and Impact.” Proceedings of
the National Academy of Sciences 105, no. 33: 11597–11604.
Zhang, M., M. Delgado- Baquerizo, G. Li, et al. 2023. “Experimental
Impacts of Grazing on Grassland Biodiversity and Function Are
Explained by Aridity.” Nature Communications 14: 5040.
Zhang, R., J. Wang, and S. Niu. 2021. “Toward a Sustainable Grazing
Management Based on Biodiversity and Ecosystem Multifunctionality
in Drylands.” Current Opinion in Environmental Sustainability 48:
36– 43.
Supporting Information
Additional supporting information can be found online in the
Supporting Information section.