Content uploaded by Solana Tabeni
Author content
All content in this area was uploaded by Solana Tabeni on Sep 26, 2017
Content may be subject to copyright.
Journal of Vegetation Science
&&
(2014)
Grazing effects on biological soil crusts and their
interaction with shrubs and grasses in an arid
rangeland
Solana Tabeni, Irene A. Garibotti, Clara Pissolito & Julieta N. Aranibar
Keywords
Cryptogams; Degradation; Livestock; Monte
Desert; Spatial distribution; Species response
curves; Thresholds; Watering points
Abbreviation
BSC = Biologicalsoil crust
Nomenclature
Brummitt & Powell (1992)
Received 31 October 2013
Accepted 14 April 2014
Co-ordinating Editor: Rasmus Ejrnæs
Tabeni, S. (Corresponding author,
stabeni@mendoza-conicet.gob.ar): Instituto
Argentino de Investigaciones de las Zonas
Aridas (IADIZA), Centro Cient
ıfico Tecnol
ogico
(CCT) CONICET, MENDOZA, Av. A. RuizLeal s/n,
CP 5500, C.C.507 Mendoza, Argentina
Garibotti, I.A. (ireneg@mendoza-
conicet.gob.ar), Pissolito, C.
(cpissolito@mendoza-conicet.gob.ar) &
Aranibar, J.N. (jaranibar@mendoza-
conicet.gob.ar): Instituto Argentino de
Nivolog
ıa, Glaciolog
ıa y Cienc ias Ambientales
(IANIGLA), CCT-CONICET, MENDOZA, Av. A.
Ruiz Leal s/n, CP 5500, C. C. 503 Mendoza,
Argentina
Aranibar, J.N. : Instituto de Ciencias B
asicas
(ICB), Universidad Nacional de Cuyo, Padre
Contreras 1300, CP 5502 Mendoza, Argentina
Abstract
Question: Biological soil crust (BSC) communities can be used in the identifica-
tion and monitoring of degradation. A key question is how landscape-scale live-
stock disturbance and other local-scale factors influence BSC communities. We
hypothesize that at the landscape scale, increased grazing pressure would lead to
decreasing cover of BSC, but at the local scale biotic interactions between BSC
and vascular plants would modulate the influence of grazing on BSC.
Methods: Spatially explicit sampling of vegetation composition and cover was
conducted using point-sampling methods in digital images along two distur-
bance gradients in the central Monte Desert in Argentina.
Results: The grazing gradient is the major determinant of changes in the struc-
ture of plant communities at the landscape scale. Approximately 1500 to
2000 m from a watering point, there is a threshold in vegetation structure asso-
ciated with a nonlinear trend of the BSC, herb, grass and shrub cover. Bivariate
spatial patterns show attraction between BSC and shrubs in the vicinity of settle-
ments, and repulsion between BSC and both grasses and litter in less disturbed
sites.
Conclusion: Grazing affects BSC directly through trampling and indirectly by
altering vascular plant communities that interact with the BSC communities.
Both these effects vary according to the spatial scale being considered. The
results evidence that understanding of livestock impact in structuring arid eco-
systems requires an integrated analysis of BSC and vascular plant communities
at different spatial scales.
Introduction
Arid and semi-arid ecosystems are experiencing land deg-
radation at different scales. This has led to increasing
research aiming to detect early signs of degradation and
predict potential undesirable transitions (Kefi et al. 2007).
Non-irrigated arid ecosystems used for intensive or exten-
sive grazing are some of the areas most at risk of desertifi-
cation (Dawelbait & Morari 2012). Grazing disrupts
vegetation organization patterns that unbalance the flow
of materials and organisms across the landscape (Alados
et al. 2011). Much of the research has focused on the study
of vascular plant patches, and less is known about the
impact of grazing on microbiotic organisms growing on the
ground surface (Bowker et al. 2006). It seems that micro-
biotic communities are more vulnerable to degradation
than their associated vascular plant communities (Bowker
et al. 2008).
Biological soil crusts (BSC) are specialized microbiotic
communities composed of mosses, lichens, liverworts,
1
Journal of Vegetation Science
Doi: 10.1111/jvs.12204©2014 International Association for Vegetation Science
cyanobacteria and algae. They reach up to 70% of the
living ground cover in some arid lands (Belnap et al.
2004), helping to stabilize the soil, regulate infiltration and
run-off patterns, fix atmospheric N and C, retain moisture
and in some cases facilitate vascular plant establishment
(Bowker et al. 2011). Because of their importance in
the functioning of desert ecosystems, it is highly relevant
to understand the impact that grazing has on BSC
communities and how this affects the process of
desertification (Bowker et al. 2006; Eldridge et al. 2010).
Disturbance by livestock reduces lichen and moss cover
and affects the functional role of BSC (Bowker et al.
2006). However, the interaction between vascular plants
and microbiotic communities is complex, modulating the
overall effect of disturbance on their diversity and cover.
At low to intermediate levels of disturbance, bare patches
provide suitable habitat for microbiotic communities that
can compete with vascular plants for resources (Eldridge
et al. 2010). In these cases, BSC and vascular plant cover
are inversely related (Muscha & Hild 2006). In other cases,
vascular plants have positive effects on BSC by ameliorat-
ing extreme environmental microclimate conditions and
modifying locally available resources (Maestre et al. 2002;
Zhang et al. 2013). For example, grasses appear to provide
microsites that facilitate BSC establishment and develop-
ment along water-centred grazing gradients (Williams
et al. 2008; Jimenez Aguilar et al. 2009).
The composition and ecology of BSC has largely been
overlooked in the context of degradation and desertifica-
tion in South America (B€
udel 2001; Toledo & Florentino
2009). In the southern Monte of Argentina it has been
observed that grazing reduces BSC richness and cover,
and that BSC are functionally relevant at retaining vascu-
lar plant seeds (Scutari et al. 2004; Bertiller & Ares
2011). In the central Monte of Argentina cattle have a
negative impact on BSC communities that mainly
develop under the protection of vascular plants in dis-
turbed areas. Recovery of BSC communities is successful
on sites where cattle have been removed for over 40 yrs
(G
omez et al. 2012).
The process of desertification is increasing along the
South American Arid Diagonal for anthropogenic and cli-
mate change reasons (Abraham et al. 2009). In this study,
we examine the effect of grazing on the cover and spatial
pattern of BSC at landscape and local spatial scales in the
central Monte Desert in Argentina. Specifically, the follow-
ing hypotheses were evaluated. (i) At the landscape scale,
increased grazing pressure would lead to a decreasing
cover of BSC because livestock disrupt BSC cover
through trampling. (ii) At the local scale, grazing has an
indirect effect on BSC cover and distribution through
affecting vascular plant communities, and consequently
biotic interactions.
Methods
Study area
The Argentinean central Monte Desert covers a wide
northwest–southeast strip that is part of the South Ameri-
can Arid Diagonal. The study sites are located inthe Telteca
Natural Reserve (32°200S, 68°00°W, 20 700 ha), central
part of the Monte Desert. The climate is arid (total annual
precipitation about 160 mm), with cold dry winters (mean
temperature <10 °C) and warm rainy summers (mean
temperature >20 °C). The dominant landscape is a gently
undulating floodplain that presents sandy–silt soils mainly
of aeolian origin (Gonz
alez Loyarte et al. 2000). Vegeta-
tion is relatively homogeneous in physiognomy and floris-
tic composition (Abraham et al. 2009). Native woodlands
of Prosopis flexuosa are confined to inter-dune valleys
(Gonz
alez Loyarte et al. 2000).The lower stratum is a
shrub layer composed mainly by Larrea divaricata,Suaeda
divaricata,Tricomaria usillo,Atriplex lampa,Capparis atami-
squea and Lycium tenuispinosum.
In these non-irrigated areas, people live in scattered
livestock settlements supplied by hand-drilled water
wells (Abraham et al. 2009). The water table lies
between 6 and 15 m, and there are no surface water
sources (Torres 2008). The economy is subsistence, with-
out any livestock management strategy, neither in time
(continuous grazing) or space (no fencing) (Torres 2008).
The impact of livestock activity on vegetation is evident,
with changes up to 15% in cover in a 2-km area around
the settlements (Goir
an et al. 2012). Unfortunately,
there is a paucity of grazing intensity data for the area
(Meglioli et al. 2013). Other human impacts are minor
as there is no other economic activity in the area (Torres
2008).
Sampling of BSC and vascular plant cover and
distribution
Vegetation surveys were performed in early summer
along two transects extending from highly impacted set-
tlements (La Primavera and Las Delicias settlements) to
less disturbed sites. Sampling distances were ca. 10, 500,
900, 1500 and 2500 m from watering points. Landscape
homogeneity and the minimal possible influence of con-
founding factors in the area (i.e. absence of surface water
sources or other disturbance impacts), makes water-
centred grazing gradients a suitable indirect approach to
studying the effect of different grazing intensities on bio-
logical communities (Landsberg et al. 2003; Sasaki et al.
2008). We limited sampling sites to inter-dune spaces
with a slight slope in order to minimize the influence of
topography. Conditions are similar at both studied gradi-
ents since they are relatively close in the area. Our study
Journal of Vegetation Science
2Doi: 10.1111/jvs.12204 ©2014 International Association for Vegetation Science
Spatial distribution of biological soilcrust S. Tabeni et al.
focused on two individual sites that we considered as
representative of the study area since the area is rela-
tively homogeneous in its climatic, physical and biologi-
cal characteristics. This sampling strategy was chosen in
order to achieve a high spatial resolution of BSC distribu-
tion at the local scale, according to the methods
described below, although it confers a lower resolution
at the landscape scale.
At each sampling site we mapped BSC and vascular
plant cover in grids of 10 95 m divided into 200 square
quadrats of 0.25 cm
2
. We used high-resolution, image-
based methods (Booth et al. 2008). Vertical images of each
sampling quadrat were obtained using a digital camera
(6.1 MP, focal length =18–55 mm, equivalent to 27–
82.5 mm format) positioned ca. 1 m above ground level,
with the focal plane oriented parallel to the ground.
Planar scale was provided through a rigid square frame
that delimited the sampling quadrat. Digital images were
processed in order to remove linear distortion along the
measurement plane defined by the rigid frame. The
resulting dimension of each image was 1500 pixels, in
which each pixel corresponds to 0.3 mm in the field. This
method has been shown to be successful in detecting
vegetative cover changes due to grazing (Booth et al.
2008). Photographic methods have also been used for
monitoring lichens, mosses and the BSC, indicating a
close fit with other standard field monitoring methods
(e.g. Vanha-Majamaa et al. 2000; Benavides & Sastre-De
Jes
us 2009; Jespersen 2013).
The spatial distribution of BSC and vascular plant cover
was mapped through point-sampling of digital images
using SamplePoint software, based on a 225-point system-
atic grid within each photograph (Booth et al. 2006). Sam-
ple points were classified into 17 different categories,
including species of shrub (Prosopis flexuosa,Geoffroea decort-
icans,Bulnesia retama,Lycium chilense,L. tenuispinosum,Con-
dalia microphylla,Atriplex lampa,Suaeda divaricata,Ximenia
americana,Capparis atamisquea,Grahamia bracteata and an
unidentified shrub), perennial grass, herb,BSC, litter and
bare soil. This resulted in a database with 405 000 data
points for presence/absence of the different species or veg-
etation group at each point. To estimate ground cover as a
percentage of the total surface area, each sample quadrat
was divided into nine sub-quadrats, and data within each
sub-quadrat (25 data points) were averaged, resulting in a
database with 16,200 data points that were used in the
following statistical analyses.
We used hand lenses to perform floristic surveys of the
crusts present within each sampling grid. Voucher speci-
mens of the different lichen and moss species were col-
lected for later taxonomic identification in the laboratory.
The presence of cyanobacteria was checked using a com-
pound light microscope, but they were not identified to
species level. Nomenclature follows Brummitt & Powell
(1992).
Statistical analyses
We applied a detrended correspondence analysis (DCA) to
identify general systematic changes in the composition and
cover of the vegetation community. This method is ade-
quate since we expected species composition to be mainly
determined by the grazing gradient, and because the ordi-
nation axes can be interpreted in terms of the average SD
of species turnover among sites and can be used to predict
species distributions (Økland 1986; Jongman et al. 1995;
Ejrnaes 2000; Rydgren et al. 2003). The matrix data con-
sisted of the entire data set of percentage ground vegeta-
tion covers. Five vegetation variables were included in the
analysis, i.e. BSC, herbs, grasses, Lycium spp. and all other
species of shrub combined into a unique variable (‘other
shrubs’). Cover of Lycium spp. was considered an indepen-
dent variable since it is the dominant shrub genus in the
area. Detrending was performed by segment, and down-
weighting of rare species was applied. The length of the
first DCA axis was >3 SD units, indicating that use of a uni-
modal model of species distributions is adequate (Jongman
et al. 1995).
We also analysed the response of single vegetation
groups to the disturbance gradient and to the ecological
gradient defined by DCA axes. Ordination axes are surro-
gates of complex gradients of factors that vary more or less
in relation to each other (Økland 1986). For this purpose,
we applied a smoothing-fitting method based on locally
weighted polynomial regressions (LOESS), using cover of
vegetation groups as dependent variable and distance from
watering point or DCA axis 1 as independent variable.
These analyses allowed us to explore the major trends of
variability in the cover of the different vegetation groups
(whether linear, uni- or multimodal, or no trend). We
selected the best smoothing parameters examining plots of
the fit residuals vs the predictor variable, and we chose the
model yielding no clearly discernible information on the fit
residuals (Jacoby 2000).
Spatial pattern of vegetation groups
To test the influence of disturbance gradient on the spatial
distribution of BSC and the changes in biotic interactions
we conducted univariate and bivariate point pattern analy-
ses (Wiegand & Moloney 2004). Second-order statistics
such as the O-ring function characterize the number of
points encountered in the neighbourhood of a ring of
radius rcentred on an arbitrary point of the pattern, allow-
ing interpretation of spatial structure in terms of interac-
tions (Wiegand & Moloney 2004).
3
Journal of Vegetation Science
Doi: 10.1111/jvs.12204©2014 International Association for Vegetation Science
S. Tabeni et al. Spatial distribution of biologicalsoil crust
First, we performed univariate analyses on each of the
nine grids to examine whether spatial distributions of BSC
along grazing gradients were random, clumped or regular.
We chose a heterogeneous Poisson null model because we
observed a heterogeneous density distribution of BSC
points, corresponding to non-constant, first-order effects.
We used a circular moving window estimator and selected
a radius r=5 and 10 cells, corresponding to 17 and 35 cm,
respectively.
Bivariate O-ring statistics were performed to test the
hypothesis that the distribution of BSC is independent of
the distribution of vascular plants. We tested the relation-
ship of BSC cover against the cover of grasses, shrubs and
vascular plant litter cover. In addition, we tested the exis-
tence of the interaction between shrubs and grasses. To
detect departure from independence, we applied a toroidal
shift null model that indicates the existence of attraction or
repulsion between the two patterns. The significant depar-
ture from univariate and bivariate null models was tested
by constructing confidence envelopes with 999 Monte
Carlo simulations (Wiegand & Moloney 2004).
Results
Biological soil crust composition
Biological soil crusts are composed of free cyanobacteria,
coccoid green algae, lichens and mosses. All soil crusts
analysed contain cyanobacteria, and the proportion of
mosses in the crusts is minor. Dominant lichen species are
Collema coccophorum,C. tenax,Fuscopannaria sp.,Heppialuto-
sa,Leptogium sp., Placidium squamulosum and Placynthium
nigrum. Mosses corresponded mainly to the species Crossidi-
um sp. and Tortula inermis. A close inspection of the crusts
made in the field with hand lenses did not indicate evident
taxonomic differences along the studied gradients.
Vegetation variations along grazing gradients
The DCA ordination of sampling plots according to their
vegetation composition indicates that the first and sec-
ond axes represent the main compositional gradients
in the data set, accounting for 27.2% and 19.7% of
species variability, respectively (Fig. 1). Higher order
axes explain less than 3% of species variability each, so
they are considered unimportant. The first ordination
axis is closely associated with the disturbance gradient,
consecutively separating the centroids of sampling sites
from sites close to watering points to less disturbed sites.
The second ordination axis is mostly related to variability
within the studied sites. High local-scale variability was
expected, given the very small quadrat size used for
vegetation sampling and the typical patch spatial pattern
of vegetation in arid environments.
The nonparametric smoothing of the geographic dis-
tance from watering points against the ecological distance
estimated by DCA1 shows that the relationship approaches
a piece-wise model comprising two lines (Fig. 2), indicat-
ing a change in the rate of species turnover between 1500
to 2000 m from watering points. The relationship depicts
that vegetation changes are faster at sites closer to watering
points than at more distant and less disturbed sites.
Figure 3 shows the response pattern of vegetation
groups along grazing and ecological gradients. BSC, herbs
and ‘other shrubs’ (those not pertaining to the genus
Lycium) show a similar pattern along the geographic
gradient, with a peak in cover at ca. 500 m from watering
points and lower cover further away. The cover of Lycium
shrubs shows an indefinite response along the disturbance
gradient. Grasses show relatively low cover near the
watering points (approximately the closest 500 m), and a
rapid cover increase further along the gradient, resembling
a truncated unimodal curve (Fig. 3a).
Fig. 1. Ordination diagram of the DCA of vegetation composition and
cover for sampling plots. Vegetation trends are displayed as lines
connecting the centroids of the successive sites, with increasing size of
symbols corresponding to increasing distance from watering points. White
and grey symbols correspond to the two grazing gradients analysed from
the human settlements (La Primavera, LP and Las Delicias, LD) in the
Telteca Reserve (Mendoza, Argentina).
Fig. 2. Nonparametric smoothing showing the relationship between
distance from watering points and sample plot scores of DCA1. The
geographic gradient represents the changes in disturbance intensity as
distance from settlements increases, and the scores of DCA axis represent
an ecological gradient of floristic composition changes.
Journal of Vegetation Science
4Doi: 10.1111/jvs.12204 ©2014 International Association for Vegetation Science
Spatial distribution of biological soilcrust S. Tabeni et al.
The patterns of response of vegetation groups to the eco-
logical gradient depict their relative occurrence along the
community gradient (Fig. 3b). Lycium and all other shrubs
have the same location on the gradient, with a response
pattern approaching a truncated distribution model, with
their optima at the initial part of the community composi-
tion gradient. The behaviour of BSC suggests a unimodal
shape of response change along the community composi-
tion gradient, with an optimum associated with a relatively
high cover of Lycium spp. and an intermediate cover of
grasses. Grasses exhibit a bimodal structure in the response
curve, with one maximum at sites with high shrub cover
and the other at sites with relatively low shrub cover.
Herbs show increased cover at the end of the community
gradient, coinciding with a high cover of grasses.
Spatial patterns
Univariate tests indicate an aggregated spatial arrangement
of BSC at all six sampling sites where it was found with
relatively high cover (Fig. 4). The BSC occurred in clumps
ranging from 7 cm (two cells) to 21 cm (six cells) along
the gradient. The differences in aggregation scales seem
not to be related to the grazing intensity gradient.
The bivariate O-ring statistic suggests interaction
between shrubs and grasses at most of the studied sites
(Table 1). The sites located at distances <1500 m from the
watering point show, on average, more points of grasses
closer to shrubs than would be expected under indepen-
dence, thus indicating attraction. However, the distribu-
tion patterns of shrubs and grasses show repulsion with
increasing distance from the watering point.
The relationships between distribution patterns of BSC
and vascular components show a variety of responses
(Table 1). At ca. 500 m from settlements, BSC shows a sig-
nificant attraction to shrub cover, with an aggregation
scale ≤70 cm. This is the site with highest cover of BSC
(Fig. 3a). At distances >1500 m from the watering point,
the distribution of BSC was independent from shrubs. In
contrast, the interaction between BSC and grasses and
between BSC and vascular plant litter was determined
through a random pattern near the settlements and a
repulsion response on most of the sites further away from
the watering point (distance >900 m).
Discussion
In arid areas, livestock rely on water from waterholes, gen-
erating gradients of decreasing grazing activity with
increasing distances from animal concentration areas, such
as corrals and watering points. Earlier studies describe
decreasing linear responses of vascular plant diversity and
cover along the grazing gradients, but more recent studies
show thresholds in the patterns of vegetation changes and
non-linear responses of species (Landsberg et al. 2003;
Briske et al. 2005; Sasaki et al. 2008). However, few stud-
ies have included the responses of BSC communities, in
spite of their importance as an integral component of
desert environments (Belnap & Weber 2013).
As expected, our results show that the gradient from
watering points is the major determinant of community
structure at the landscape scale, determining changes
in the relative cover of both BSC and vascular plants
(Figs 1, 3a). Our survey design minimized the influence of
confounding factors other than grazing impact along the
water-centred gradients, thus the grazing gradient appears
as the prevailing disturbance regime. In addition, our
results show that the main ecological gradient is not line-
arly related to the geographic gradient, showing a break-
point between 1500 and 2000 m from settlements (Fig. 2).
According to Briske et al. (2005), an abrupt change in
floristic composition along disturbance or natural
gradients suggests the existence of an ecological threshold.
(a)
(b)
Fig. 3. Nonparametric smoothing showing the response pattern of cover
of the different vegetation groups to: (a) disturbance gradient (distance
from watering point) and (b) ecological gradient (sample plot scores of
DCA1).
5
Journal of Vegetation Science
Doi: 10.1111/jvs.12204©2014 International Association for Vegetation Science
S. Tabeni et al. Spatial distribution of biologicalsoil crust
Significant changes in vegetation cover within 2000 m of
the watering points have been found from satellite imag-
ery within the same area of our study (Goir
an et al. 2012).
Our study reinforces these results, indicating that changes
in vegetation cover are associated with major community
structural changes. Probably the threshold indicates differ-
ences between the vegetation communities developing
within and outside the area of major influence of the graz-
ers. A concentrated effect of grazing on the vegetation in
small areas in the close proximity to watering pointsis typi-
cal of rangelands not subjected to animal management
strategies (Andrew & Lange 1986; Meglioli et al. 2013).
The structural threshold we found between 1500 and
2000 m from watering points is mainly associated with an
increasing dominance of grasses in the community
(Fig. 3a,b). It is well known that selective grazing of live-
stock on grasses determines positive response curves for
reduced grazing intensity (Williams et al. 2008; Wesuls
et al. 2013). However, we found a bimodal pattern of grass
response to the ecological gradient (Fig. 3b), suggesting
that other factors might also be important for explaining
the observed vegetation changes. Bimodal curves along
0.24
0.18
0.12
500 m
900 m
1500 m
Minimum distance from settlement
2500 m
0.06
0.12
0.09
0.06
0.03
0
0.25
0.2
0.15
0.1
0
0.05
Oll(r)
O11(r)
O11(r)
O11(r)
0.2
0.16
0.12
0.08
0.04
O11(r) O11(r)
0.2
0.16
0.12
0.08
0.04
0.2
0.16
0.12
0.08
0.04
0
0123456789
Spatial scale r [cells]
‘La Primavera’ settlement ‘Las Delicias’ settlement
Univariate O-ring statistic (W-M)
Univariate O-ring statistic (W-M) Univariate O-ring statistic (W-M)
Univariate O-ring statistic (W-M)
Univariate O-ring statistic (W-M)
Insufficient data
Univariate O-ring statistic (W-M)
10 11 12 13 14 15 16 17 18 19 20
0123456789
Spatial scale r [cells]
10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 8 9
Spatial scale r [cells]
10 11 12 13 14 15 16 17 18 19 20
0123456789
Spatial scale r [cells]
10 11 12 13 14 15 16 17 18 19 20 01234567 89
Spatial scale r [cells]
10 11 12 13 14 15 16 17 18 19 20
0 123456789
Spatial scale r [cells]
10 11 12 13 14 15 16 17 18 19 20
Fig. 4. Univariate point pattern analysis using the O-ring statistic (bold lines) in the sampling grids. Dashed lines indicate 95% confidence envelopes for the
null model. Insufficient data indicate that the analysis could not be performed due to low cover of BSC in the study site.
Journal of Vegetation Science
6Doi: 10.1111/jvs.12204 ©2014 International Association for Vegetation Science
Spatial distribution of biological soilcrust S. Tabeni et al.
ecological gradients are usually related to mechanisms of
species competitive interaction (Austin & Smith 1989).
Our results of the bivariate analysis of the distribution of
grass and shrub covers within grids (local scale) concur
with this interpretation, showing a pattern of attraction on
heavily grazed sites and of repulsion on moderate to less
grazed sites (Table 1). Attraction is interpreted as a positive
relationship between shrubs and grasses, probably related
to the protection of grasses from being grazed beneath
shrub canopies, and repulsion might be associated with the
colonization of shrub interspaces by grasses as disturbance
through grazing is relaxed.
The BSC are sensitive to trampling by livestock, thus
reducing their cover as intensity of disturbance increases
(Berkeley et al. 2005; Thomas & Dougill 2006, 2007). Wil-
liams et al. (2008) found exponential increases in the
cover of cyanobacteria-dominated BSC as density of cattle
decreased along a 500-m long, water-centred grazing gra-
dient. To the best of our knowledge, our results are the first
to provide insight into the responses of BSC cover along
larger grazing gradients (2500 m), showing a non-linear
response pattern with a peak at an intermediate level of
grazing (Fig. 3a). This pattern has been commonly
reported for several groups of vascular plants, and is inter-
preted to result from the ability of the species to tolerate
intermediate grazing pressures, while plant decline at less
disturbed sites is due to competitive interactions (Sasaki
et al. 2008; Wesuls et al. 2013).
The distribution of increased BSC cover found at inter-
mediate levels of disturbance (500 m from watering
points) is closely related to the distribution of shrub cover,
while at less disturbed sites further away from watering
points, the distributions of BSC and shrubs are indepen-
dent (Table 1). This suggests facilitation of BSC develop-
ment beneath shrub canopies at relatively disturbed
sites, while this interaction is not apparent at less disturbed
sites. These results agree with those found in Kalahari
rangelands in showing that BSC development is restricted
to sites beneath shrub canopies at high levels of distur-
bance, while interspaces and shrub subcanopies provide
equally suitable habitat for BSC development under lim-
ited disturbance (Berkeley et al. 2005). They are also in
agreement with the results of a previous study in the
Argentinean central Monte Desert, which found that BSC
communities are restricted to Larrea subcanopies at dis-
turbed sites, but successfully recover in open spaces at sites
where cattle have been removed for more than 40 yrs
(G
omez et al. 2012).
The distribution of BSC cover shows patterns of repul-
sion towards grass and litter cover at sites with relatively
low disturbance further than 900 m from watering points
(Table 1). This result is in contrast to those studies that
found a significant association between grasses and BSC
due to modification of soil properties and buffering of abi-
otic stressors beneath tussocks (Bowker 2007; Read et al.
2008), but agrees with others that report the spatial segre-
gation of BSC and grasses due to competition for natural
resources or living space (Bowker et al. 2010; Peterson
2013). According to Bowker et al. (2010), occupation of
more living space is the most important competitive factor
structuring BSC communities, given that more space
secures access to more resources. In addition, the observed
negative spatial relationship between BSC and plant leaf
litter can be related to BSC burial reducing the availability
of light for photosynthesis (Berkeley et al. 2005; Thomas &
Dougill 2007). Therefore, the diminished availability of
bare soil due to both the increased cover of grasses and lit-
ter fall has a detrimental effect on the development of BSC
at relatively undisturbed sites in our study area.
Conclusions
Our study analysed the changes in vascular plant and BSC
in an arid ecosystem of central Argentina as a function of
Table 1. Results of the bivariate point pattern analysis along a grazing disturbance gradient showing the scales of attraction and repulsion between the
cover of different vegetation groups and the biological soil crust (BSC). Not available (na), indicates that the analysis could not be performed due tolow
cover or absence of BSC in the sites. Names of the sites as in Fig. 1.
Distance from watering point (m)
10 500 900 1500 2500
LP1 LD1 LP2 LP3 LD2 LP4 LD3 LP5 LD4
Shrubs 9Grasses Attraction
(≤70 cm)
Attraction
(≤70 cm)
Attraction
(14–31 cm)
Random Attraction
(≤63 cm)
Repulsion
(≤17 cm)
Random Repulsion
(≥56–70 cm)
Repulsion
(≤63 cm)
Shrubs 9BSC na na Attraction
(10–70 cm)
Random Random Random na na Random
Grasses 9BSC na na Random Repulsion
(63–70 cm)
Random Repulsion
(17–56 cm)
na na Random
Litter 9BSC na na Random Repulsion
(≤10 cm)
Random Repulsion
(42–70 cm)
na na Repulsion
(≤17 cm)
7
Journal of Vegetation Science
Doi: 10.1111/jvs.12204©2014 International Association for Vegetation Science
S. Tabeni et al. Spatial distribution of biologicalsoil crust
major disturbance by livestock. The results suggest
nonlinear responses of the different vegetation groups
and the BSC to the disturbance intensity gradient.
Development of grasses is possible in relatively highly
impacted sites due to their being protected under shrub
canopies. At these sites, development of BSC is also
facilitated beneath shrubs. However, as livestock distur-
bance relaxes, it seems that increase in grass cover and
plant leaf litter in interspaces between shrubs limits
development of the BSC. Further studies should analyse
the generality of our results for multiple ecological sites,
and how the changes in vegetation patterns relate to
processes that regulate the functioning of this desert
ecosystem.
Acknowledgements
This study was partially funded by ANPCyT (grant PICT
1417). We also are most grateful to Nelly Horak forrevising
the English version of the manuscript. We thank Benjamin
Bender, Agustina Barros and Mauro Britos Navarro for
assistance with analysis of the images.
References
Abraham, E., del Valle, H.F., Roig, F., Torres, L., Ares, J.O., Coro-
nato, F. & Godagnone, R. 2009. Overview of the geography
of the Monte Desert biome (Argentina). Journal of Arid Envi-
ronments 73: 144–153.
Alados, C.L., Puigdef
abregas, J. & Mart
ınez-Fern
andez, J. 2011.
Ecological and socio-economical thresholds of land and
plant-community degradation in semi-arid Mediterranean
areas of southeastern Spain. Journal of Arid Environments 75:
1368–1376.
Andrew, M.H. & Lange, R.T. 1986. Development of a new pio-
sphere in arid chenopod shrubland grazed by sheep. 1.
Changes to the soil surface. Australian Journal of Ecology 11:
395–409.
Austin, M.P. & Smith, T.M. 1989. A new model for the contin-
uum concept. Vegetatio 83: 35–47.
Belnap, J. & Weber, B. 2013. Biological soil crusts as an integral
component of desert environments. Ecological Processes 2: 11.
Belnap, J., Phillips, S.L. & Miller, M.E. 2004. Response of desert
biological soil crusts to alterations in precipitation frequency.
Oecologia 141: 306–316.
Benavide s, J.C. & Sastre-De Jes
us, I. 2009. Digitized images pro-
vide more accuracy and efficiency to estimate bryophyte
cover. The Bryologist 112: 12–18.
Berkeley, A., Thomas, A.D. & Dougill, A.J. 2005. Cyanobacterial
soil crusts and woody shrub canopies in Kalahari rangelands.
African Journal of Ecology 43: 137–145.
Bertiller, M.B. & Ares, J.O. 2011. Does sheep selectivity along
grazing paths negatively affect biological crusts and soil seed
banks in arid shrublands? A case study in the Patagonian
Monte, Argentina. Journal of Environmental Management 92:
2091–2096.
Booth, D.T., Cox, S.E. & Berryman, R.D. 2006. Point sampling
digital imagery with “Samplepoint”. Environmental, Monitor-
ing & Assessment 123: 97–108.
Booth, D.T., Cox, S.E., Meikle, T. & Zuuring, H.R. 2008. Ground-
Cover measurements: assessing correlation among aerial and
ground-based methods. Environmental Management 42: 1091–
1100.
Bowker, M.A. 2007. Biological soil crust rehabilitation in theory
and practice: an underexploited opportunity. Restoration Ecol-
ogy 15: 13–23.
Bowker, M.A., Belnap, J., Davidson, D.W. & Goldstein, H. 2006.
Correlates of biological soil crust abundance across a contin-
uum of spatial scales: support for a hierarchical conceptual
model. Journal of Applied Ecology 43: 152–163.
Bowker, M.A., Miller, M.E., Belnap, J., Sisk, T.D. & Johnson,
N.C. 2008. Prioritizing conservation effort through the use of
biological soil crusts as ecosystem function indicators in an
arid region. Conservation Biology 22: 1533–1543.
Bowker, M.A., Soliveres, S. & Maestre, F.T. 2010. Competition
increases with abiotic stress and regulates the diversity of
biological soil crusts. JournalofEcology98: 551–
560.
Bowker, M.A., Mau, R.L., Maestre, F.T., Escolar, C. & Castillo-
Monroy, A.P. 2011. Functional profiles reveal unique
ecological roles of various biological soil crust organisms.
Functional Ecology 25: 787–795.
Briske, D.D., Fuhlendorf, S.D. & Smeins, F.E. 2005. State-and-
transiti on models, thresholds, and rangeland heal th: a
synthesis of ecological concepts and perspectives. Rangeland
Ecology & Management 58: 1–10.
Brummitt, R.K. & Powell, C.E. 1992. Authors of Plants Names,
p. 732. Royal Botanical Garden, Kew, UK.
B€
udel, B. 2001. Biological soil crusts of South America. In:
Belnap, J. & Lange, O.L. (eds.) Biological Soil Crusts: Structure,
Function, and Management,pp.51–55. Springer, Berlin, DE.
Dawelbait, M. & Morari, F. 2012. Monitoring desertification in a
Savannah region in Sudan usingLandsat images and spectral
mixture analysis. Journal of Arid Environment 80: 45–55.
Ejrnaes, R. 2000. Can we trust gradients extracted by Detrended
Correspondence Analysis? Journal of Vegetation Science 11:
565–572.
Eldridge, D.J., Val, J., James, A.I. 2010. Abiotic effects predomi-
nate under prolonged livestock-induced disturbance.Austral
Ecology 36: 367–377.
Goir
an, S., Aranibar, J.N. & Gomez, M.L. 2012. Heterogeneous
spatial distribution of traditional livestock settlements and
their effects on vegetation cover in arid groundwater coupled
ecosystems in the Monte desert (Argentina). Journal of Arid
Environment 87: 188–197.
G
omez, D.A., Aranibar, J.N., Tabeni, S., Villagra, P.E., Garibotti,
I.A. & Atencio, A. 2012. Biological soil crust recovery
after long-term grazing exclusion in the Monte Desert
(Argentina). Changes in coverage, spatial distribution, and
soil nitrogen. Acta Oecologica 38: 33–40.
Journal of Vegetation Science
8Doi: 10.1111/jvs.12204 ©2014 International Association for Vegetation Science
Spatial distribution of biological soilcrust S. Tabeni et al.
Gonz
alez Loyarte, M., Rodeghiero, A.G., Buk, E. & Trione, S.
2000. An
alisis comparativo de dos comunidades en el bosque
de Prosopis flexuosa dc. del NE de Mendoza, Argentina. Mult-
equina 9: 75–89.
Jacoby, W.G. 2000. Loess: a nonparametric, graphical tool for
depicting relationships between variables. Electoral Studies
19: 577–613.
Jespersen, R.G. 2013. Experimental Investigations into the Ecological
Functions of Cryptogams in Alpine Plant Communities.School
of Environmental and Forest Sciences, University of
Washington, Washington, DC, US.
Jimenez Aguilar, A., Huber-Sannwald, E., Belnap, J., Smart,
D.R. & Arredondo Moreno, J.T. 2009. Biological soil crusts
exhibit a dynamic response to seasonal rain and release from
grazing with implications for soil stability. Journal of Arid
Environment 73: 1158–1169.
Jongman, R.H.G., ter Braak, C.J.F. & van Tongeren, O.F.R.
1995. Data Analysis in Community and Landscape Ecology.
Cambridge University Press, Cambridge, UK.
Kefi, S., Rietkerk, M., Alados, C.L., Pueyo, Y., Papanastasis, V.P.,
Elaich, A. & de Ruiter, P.C. 2007. Spatial vegetation patterns
and imminent desertification in Mediterranean arid ecosys-
tems. Nature 449: 213–218.
Landsberg, J.D., James, C.D., Morton, S.R., Muller, W.J. & Stol,
J. 2003. Abundance and composition of plant species along
grazing gradients in Australian rangelands. Journal of Applied
Ecology 40: 1008–1024.
Maestre, F.T., Huesca, M.T., Zaady, E., Bautista, S. & Cortina, J.
2002. Infiltration, penetration resistance and microphytic
crust composition in contrasted microsites within a Mediter-
ranean semi-arid steppe. Soil Biology & Biochemistry 34:
895–898.
Meglioli, P.A., Aranibar, J.N., Villagra, P.E., Alvarez, J.A. &
Jobb
agy, E.G. 2014. Livestock stations as foci of groundwater
recharge and nitrate leaching in a sandy desert of the Central
Monte, Argentina. Ecohydrology 7: 600–611.
Muscha, J.M. & Hild, A.L. 2006. Biological soil crusts in grazed
and ungrazed Wyoming sagebrush steppe. Journal of Arid
Environment 67: 195–207.
Økland, R. 1986. Rescaling of ecological gradients. I. Calculation
of ecological distance between vegetation stands by means
of their floristic composition. Nordic Journal of Botany 6:
651–660.
Peterson, E.B. 2013. Regional-scale relationship among biologi-
cal soil crusts, invasive annual grasses, and disturbance.
Ecological Processes 2: 2. doi:10.1186/2192-1709-2-2.
Read, C.F., Duncan, D.H., Vesk, P.A. & Elith, J. 2008. Biological
soil crust distribution is related to patterns of fragmentation
and land use in a dryland agricultural landscape of southern
Australia. Landscape Ecology 23: 1093–1105.
Rydgren, K., Økland, R.H. & Økland, T. 2003. Species response
curves along environmental gradients. A case study from SE
Norwegian swamp forests. Journal of Vegetation Science 14:
869–880.
Sasaki, T., Okayasu, T., Jameran, U. & Takeuchi, K. 2008.
Threshold changes in vegetation along a grazing gradient in
Mongolian rangela nds. Journal of Ecology 96: 145–
154.
Scutari, N., Bertiller, M.B. & Carrera, A. 2004. Soil-associated
lichens in rangelands of north-eastern Patagonia. Their
potential as bioindicators of conservation and grazing distur-
bance. The Lichenologist 36: 405–412.
Thomas, A.D. & Dougill, A.J. 2006. Distribution and characteris-
tics of cyanobacterial soil crusts in the Molopo Basin, South
Africa. Journal of Arid Environment 64: 270–283.
Thomas, A.D. & Dougill, A.J. 2007. Spatial and temporal distri-
bution of cyanobacterial soil crusts in the Kalahari: implica-
tions for soil surface properties. Geomorphology 85: 17–29.
Toledo, V. & Florentino, A. 2009. Las costras microbi
oticas del
suelo. Revista de Investigaci
on 68: 199–216.
Torres, L.M. 2008. Hilos de agua, lazos de sangre:enfrentando la
escasez en el desierto de Lavalle (Mendoza, Argentina).
Ecosistemas 17: 46–59.
Vanha-Majamaa, I., Salemaa, M., Tuominen, S. & Mikkola, K.
2000. Digitized photographs in vegetation analysis –a
comparison of cover estimates. Applied Vegetation S cience 3:
89–94.
Wesuls, D., Pellowski, M., Suchrow, S., Oldeland, J., Jansen, F.
& Dengler, J. 2013. The grazing fingerprint: modelling
species responses and trait patterns along grazing gradients
in semi-arid Namibian rangelands. Ecological Indicators 27:
61–70.
Wiegand, T. & Moloney, K.A. 2004. Rings, circles, and null-
models for point pattern analysis in ecology. Oikos 104:
209–229.
Williams, W.J., Eldridge, D.J. & Alchin, B.M. 2008. Grazing and
droughtreducecyanobacterialsoilcrustsinanAustralian
Acacia woodland. Journal of Arid Environment 72: 1064–1075.
Zhang, J.H., Wu, B., Li, Y.H., Yang, W.B., Lei, Y.K., Han, H.Y. &
He, J. 2013. Biological soil crust distribution in Artemisia
ordosica communities along a grazing pressure gradient in
Mu Us Sandy Land, Northern China. Journal of Arid Land 5:
172–179.
9
Journal of Vegetation Science
Doi: 10.1111/jvs.12204©2014 International Association for Vegetation Science
S. Tabeni et al. Spatial distribution of biologicalsoil crust
Graphical Abstract
The contents of this page will be used as part of the graphical abstract of html only. It will not be published as part
of main.
Evidence suggests that the main biotic factors structuring biological soil crust communities in areas disturbed by grazing
are the presence of shrubs and grasses, with opposite patterns across the gradient. There is a relationship of attraction
between BSC and shrubs in the vicinity of the settlements, and of repulsion between BSC and both grasses and litter in less
disturbed sites.