Contents lists available at ScienceDirect

Soil Biology and Biochemistry

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

Exclusion of small mammals and lagomorphs invasion interact with human-trampling to drive changes in topsoil microbial community structure andfunction in semiarid Chile

Fernando D. Alfaroa,b,c,∗, Marlene Manzanoa, Sebastian Abadesa, Nicole Trefaulta,Rodrigo de la Iglesiad, Aurora Gaxiolab,e,f, Pablo A. Marquetb,e,f,g,h, Julio R. Gutierrezb,i,Peter L. Meservej, Douglas A. Keltk, Jayne Belnapl, Juan J. Armestob,e,m

aGEMA Center for Genomics, Ecology & Environment, Universidad Mayor, Camino La Piramide, 5750, Huechuraba, Santiago, Chileb Instituto de Ecología and Biodiversidad (IEB), Casilla, 653, Santiago, Chilec Centro de Biodiversidad y Genética (CBG), Facultad de Ciencias y Tecnología, Universidad Mayor de San Simón, Sucre y Parque La Torre, Cochabamba, Boliviad Departamento de Genética Molecular y Microbiología, Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Alameda, 340, Santiago, Chilee Departamento de Ecología, Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Alameda, 340, Santiago, Chilef Laboratorio Internacional en Cambio Global (LINCGlobal, CSIC-PUC), Facultad de Ciencias Biológicas, Pontificia Universidad Católica de Chile, Santiago, Chileg The Santa Fe Institute, 1399 Hyde Park Road, Santa Fe, NM, 87501, USAh Centro de Cambio Global UC (CCGUC), Pontificia Universidad Católica de Chile, Vicuña Mackenna, 4860, Macul, Santiago, Chilei Departamento de Biología, Universidad de La Serena, and Centro de Estudios Avanzados en Zona Árida (CEAZA), La Serena, ChilejDepartment of Biological Sciences, University of Idaho, Moscow, ID, 83844, USAk Department of Wildlife, Fish, and Conservation Biology, University of California, One Shields Avenue, Davis, CA, 95616-8627, USAlU.S. Geological Survey, Southwest Biological Science Center, Moab, UT, 84532, USAm Cary Institute of Ecosystem Studies, Millbrook, NY, USA

A R T I C L E I N F O

Keywords:BiocrustsNitrogenSemi-arid ecosystemBiodiversity lossSoil stability

A B S T R A C T

Species losses and additions can disrupt the relationship between resident species and the structure and func-tioning of ecosystems. Persistent human-trampling, on the other hand, can have similar effects through thedisruption of biocrusts on surface soils of semiarid systems, affecting soil stability and fixation of carbon andnitrogen. Here, we tested the interactive and synergistic impacts of the exclusion of native mammalian herbi-vores and the effects of introduced lagomorphs in a semiarid thorn scrub ecosystem, where soils were subjectedto two different trampling intensities (i.e., trampled and non-trampled). We postulated that because of theirdifferential habitat use and fossorial activities, with respect to native small mammals, lagomorphs would havestrong negative effects on soil structure, biocrust cover, and biocrust bacterial community structure. Our ex-pectations were that changes in biocrust cover in response to trampling where native mammals were excluded,but exotic lagomorphs were present, will spread their impacts on soil chemical and physical features. To test ourhypotheses, we measured changes in soil biogeochemical properties in four experimental plots where lago-morphs (L)/small mammals (SM) were experimentally manipulated to exclude them from the plots (−), or letthem be present (+). The experimental combinations monitored were: -L/+SM, -L/-SM, +L/+SM, and +L/-SM. Results showed that human-trampling disturbance interacted with the loss of native small mammals and thepresence of non-native lagomorphs to cause large changes on biological (i.e., biocrust cover, bacterial and nifHgenes abundance), physical (i.e., soil moisture and soil stability) and chemical (i.e., TC and TN) soil features. Therelative impacts of trampling disturbance on biological and physicochemical features were strongly influencedby the presence of non-native lagomorphs. For example, larger decreases in biocrust cover and bacterialabundance were observed in treatments without lagomorphs (-L/+SM; -L/-SM). In turn, losses of biocrust cover,in addition to trampling, determined decreases in soil stability in all treatments. These results suggest that non-native lagomorphs surpass the effects of the loss of native small mammals in reducing soil quality and pro-ductivity. Therefore, human-trampling has the potential to convert low disturbed soils, as those observed in non-trampled soils in treatments -L/+SM, -L/-SM into poor soils with low biocrusts cover and concomitant lowstability, as observed in +L/+SM; +L/-SM treatments. These findings agree with previous observations thatdifferent components of global change act in synergic ways in fragile, water-limited environments. Because

https://doi.org/10.1016/j.soilbio.2018.05.019Received 31 December 2017; Received in revised form 17 May 2018; Accepted 18 May 2018

∗ Corresponding author. Centro de Genómica, Ecología y Medio Ambiente, Universidad Mayor, Santiago, Chile.E-mail addresses: fernando.alfaro@mayor.cl, fernando.alfaro@umayor.cl (F.D. Alfaro).

Soil Biology and Biochemistry 124 (2018) 1–10

Available online 28 May 20180038-0717/ © 2018 Elsevier Ltd. All rights reserved.

T

biological invasions and soil surface disturbance are becoming widespread in dryland regions globally, under-standing the long-term consequences of these interactions is essential.

1. Introduction

Above and belowground processes in ecosystems are often in-tegrated by the activities of vertebrate herbivores and this integrationcan become decoupled when dominant native herbivores are excluded(Bardgett and Wardle, 2003; Harrison and Bardgett, 2008). Accord-ingly, the loss of native herbivores is often accompanied by a decline innative functional diversity, which can substantially impact soil eco-system processes (Neely et al., 2009; IPCC, 2013; Fleming et al., 2014;Eldridge et al., 2017). This loss, in turn, can accelerate the effect ofother forms of environmental degradation, such as global warming,through altering soil carbon (C) and nitrogen (N) cycling, in addition toenhancing soil erosion (Stavi et al., 2008; Hooper et al., 2012; Fleminget al., 2014). Drylands are expected to undergo critical environmentaltransformations as a consequence of climate change drivers, as well asthrough the intense disturbances caused by exotic animal grazing andtrampling (Eldridge et al., 2016; Maestre et al., 2016). Many drylandsoils are thin and oligotrophic, with low C and N contents, and hencerecovery rates after physical disturbances can be exceedingly slow(Barger et al., 2006; Housman et al., 2006; Ferrenberg et al., 2015).Also, soil surface disruption by trampling and digging decreases soilstability and accelerates fertility loss via wind and water erosion(Eldridge and Myers, 2001; Neff et al., 2005; Eldridge and Koen, 2008).Despite increasing evidence for soil disruption due to the loss of nativeherbivores, and concomitant increases in exotic species activities andtrampling (Eldridge and Myers, 2001; Reid et al., 2005; Eldridge andKoen, 2008; Kuske et al., 2012; Ferrenberg et al., 2015), the interactiveeffects of these three critical factors upon soil structure and functionremain poorly understood.

In water-limited ecosystems, vascular plant cover is sparse and soilsbetween vegetation patches are often occupied by a biotic consortiumof cyanobacteria, lichens, heterotrophic bacteria and fungi, algae, andmosses, known collectively as biological soil crusts or biocrusts (Belnapand Lange, 2003; Belnap, 2006). This multi-trophic assembly of mi-croorganisms often represents the functional interface between the soiland the atmosphere in these semiarid ecosystems, driving the cycles ofresources (e.g., soil water) and elements (e.g., C, N, and phosphorus)(Chamizo et al., 2012; Elbert et al., 2012). As biocrusts become domi-nated by lichens or mosses, C and N fixation increase in tandem, alongwith soil aggregation and stability that prevents erosion (Eldridge andLeys, 2003; Elbert et al., 2012; Delgado-Baquerizo et al., 2013). Asdigging and trampling by exotic animals increases, cover and con-nectivity between biocrust patches decreases substantially (Eldridgeet al., 2016).

Physical soil disturbance can modify the functional structure ofbiocrusts by transforming late successional assemblages dominated bylichens and mosses into early successional ones dominated by cyano-bacteria; these early successional communities are both less productiveand have limited ability to stabilize soils (Ferrenberg et al., 2015;Belnap et al., 2016). In semiarid environments of central Chile, under aMediterranean-type climate, with a pronounced summer drought, late-successional biocrusts are naturally absent or only occur in low abun-dance; therefore, most soils are often covered by cyanobacterial-dominated biocrusts (Castillo-Monroy and Maestre, 2011). The domi-nant native small mammals in these sites graze on annual herbs andpreferentially excavate burrows beneath shrubs, cactus, and rocks. Incontrast to native herbivores, the introduced European hare (Lepuseuropaeus) and the European rabbit (Oryctolagus cuniculus) have amarked preference for burrowing in open areas between shrubs (Jaksicet al., 1979; Fuentes et al., 1983); often resulting in soil disruption

between bushes (Jaksic et al., 1979). In addition, introduced lago-morphs facilitate invasion by exotic grasses as they forage heavily onthe native perennial species (Gutierrez et al., 2010). As in other semi-arid environments, in semiarid Chile, mammals and other native spe-cies of vertebrates (e.g., reptiles and birds) are affected by habitatfragmentation (Araya et al., 1992), species invasion (Jimenez et al.,2011; Meserve et al., 2016), livestock grazing and changes in land use(Acosta-Jamett et al., 2016).

In this research, we assessed the effects of a 10-year exclusion ofnative and introduced herbivores on soil physical and chemical fea-tures, as well as on the structure and function of the bacterial biocrustcommunity in a Chilean semiarid thorn scrub. Specifically, we hy-pothesized that (i) the presence and activities of invasive lagomorphsand exclusion of native herbivores, combined with human trampling,will reduce biocrust cover and bacterial abundance, in turn drivingchanges in community structure and composition, and (ii) changes inbacterial community organization will have a cascading effect onedaphic physicochemical properties, such as nutrient cycling processesand soil stability. Further, we hypothesize that soil fertility will declinein parallel with biocrust cover loss and decreasing bacterial abundanceacross the disturbance gradient of trampling intensity and mammalburrowing.

2. Materials and methods

2.1. Study site

The study site was located in the Fray Jorge Forest National Park (FJhenceforth), near the southern border of the Atacama Desert in centralChile (71°39′W, 30°39′S). The prevailing climate is semiarid, with astrong Mediterranean influence, with 90% of precipitation con-centrated in the winter months (May through September) and largeinter-annual variability in total rainfall. The mean annual temperature(MAT) is ∼18 °C, with maximum temperatures during the warm anddry summers (∼27 °C) and minimum temperatures during cold winters(∼4 °C) (Jimenez et al., 2011; Aguilera et al., 2016). Mean annualprecipitation over the last two decades is 127mm (Armas et al., 2016),although El Niño–Southern Oscillation (ENSO) events promote strongvariation in precipitation among years.

Soils across the study area are highly heterogeneous, with largedifferences in physical and chemical features between open patcheswithout shrub cover and sites beneath the crown of dominant shrubs(Gutiérrez et al., 1993). Whereas, soil pH remained similar betweenthese two micro-environments, electric conductivity is 5.5 fold lower inopen areas. Soil texture in open areas and below shrub cover is char-acterized as loamy sand, with low values of silt and clay. Soil nutrients(i.e., N and P) and organic matter contents in open areas are sig-nificantly lower than in the soil underneath shrubs, and these differ-ences are greater during the wet season (Gutiérrez et al., 1993; Aguileraet al., 1999).

The local plant community is characterized by a scattered popula-tion of drought-deciduous and evergreen spiny shrubs, 2–3m tall, es-tablished over an ephemeral carpet of herbaceous plants on a primarilysandy substrate (Gutierrez et al., 1993). The dominant shrub species arethe evergreen Porlieria chilensis I.M. Johnst. (Zygophyllaceae, ca. 31%cover), the drought-deciduous Proustia cuneifolia D. Don (Asteraceae,ca. 9% cover), the drought-semidecidous Adesmia bedwellii Skottsb.(Fabaceae, ca. 5% cover), and the sub-shrub Chenopodium petiolareKunth (Chenopodiaceae, ca. 16% cover) (Gutierrez et al., 1997, 2004).Mean cover of all other shrub species combined was lower than 1.5%.

F.D. Alfaro et al. Soil Biology and Biochemistry 124 (2018) 1–10

2

The native small mammal assemblage associated with this semiaridlandscape includes three caviomorph rodents, Octodon degus, O. lunatusand Abrocoma bennetti; four sigmodontine rodents, Abrothrix longipilis,A. olivaceus, Oligoryzomys longicaudatus, Phyllotis darwini; and the mouseopossum Thylamys elegans (Meserve et al., 2016). By far the mostabundant small mammal in the area is the herbivore O. degus, followedby the granivore-herbivore P. darwini and by the omnivore A. olivaceus(Gutierrez et al., 2010; Meserve et al., 2016). The introduced Europeanhare and the European rabbit, once rare or absent from the mammalianassemblage, have increased their density by ∼35% since 2002 (seeGutierrez et al., 2010).

2.2. Experimental setup and sampling design

The study was conducted in a set of long-term exclusion plots(75× 75m in area) established in 1989 to examine the response ofsmall mammals and plants to controlled biotic interactions and inter-annual climate fluctuations (Fig. 1A). This is the longest experimentalfield study running in temperate South America (Gutierrez et al., 2010,see: Supplementary material, Table S1). To test the hypotheses of thispaper, we measured soil physical and chemical features, biocrust cover,and abundance of nifH genes (indicators of the presence of nitrogenfixing soil organisms), and the relative abundance and composition ofbacterial groups in four plots randomly assigned to four treatments(hereafter: exclusion treatment). These included lagomorph (L)/smallmammal (SM) presence (+) and exclusion (−): -L/+SM, -L/-SM, +L/+SM, and +L/-SM (Table 1). To evaluate soil and plant responses, fivepairs of 0.5× 0.5m subplots were randomly placed within each mainplot. Each pair of subplots consisted of one subplot located in un-disturbed areas between plants (hereafter: non-trampled) and onesubplot located at least one meter away, in disturbed (soil compacted)plant interspaces, where the staff walks once every month to conductvegetation surveys (hereafter: trampled; see Fig. 1B and Table 1). Thisresulted in a fully factorial experimental design with two factors: typeof exclusion (with four levels; -L/+SM to + L/-SM) and tramplingdisturbance (with two levels; trampled and non-trampled soils).

2.3. Sampling and field measurements

Samples of top soil (N=5; 0–1 cm, ∼20 g) were collected usingsterilized spatulas from each subplot. Larger gravel and organic mate-rial was removed using a 2-mm mesh; samples were then bulked andstored in plastic bags that were kept on ice during transport to the la-boratory. Soil gravimetric water content was determined for oven-driedsamples (90 °C for 72 h), and bulk density was estimated by the coremethod (5 cm soil depth). The index of stability for the top 0–5mm ofsoil was measured in 10 samples per subplot using the field soil stabilitytest (Herrick et al., 2005). This test identifies and assigns samples to soilstability categories (from 1 to 6), based on their rate of loss of ag-gregation in water (see: supplementary material). Total biocrust andannual plant cover (%) from each subplot were estimated using thepoint-intercept method. Lagomorph activity was evaluated by re-cording fresh trails and feces within each subplot, and along trampledareas and non-trampled open areas between shrubs (Jaksic et al., 1979;Fuentes et al., 1983; Lopez-Cortes et al., 2007).

2.4. Physicochemical and molecular analysis of soil samples

A subsample of each soil sample collected was ground and dried fordetermination of total C, N, and δ13C and δ15N isotopic signals, using aThermo Delta V Advantage IRMS, coupled with a Flash2000 ElementalAnalyzer, at the Departamento de Ecología, Pontificia UniversidadCatólica de Chile, Santiago. Ammonium (NH4-N) and nitrate (NO3-N)concentrations in soil subsamples were determined using fractionatedsteam distillation (Pérez et al., 1998). Soil pH was measured in a 1:2soil to water suspension.

2.4.1. DNA extraction and quantitative PCR analysesThe relative abundance of bacteria in soil samples was evaluated for

DNA using qPCR (Eco Real-Time PCR System, Illumina, San Diego,California, USA). For amplification, we used primers 338f/518r (Ovreasand Torsvik, 1998). Each 20 μL reaction contained 10 μL 2x KAPASYBR® FAST qPCR Master Mix Universal (Kapa Biosystems, Boston,Massachusetts, USA) and 0.5 μM of primers (Invitrogen, Life Technol-ogies, California, USA). The qPCR reactions were performed followingEinen et al. (2008). Further information in supplementary material.

The abundances of nifH gene in soil samples was evaluated using10 μL reaction mixture containing 5 μL of KAPA SYBR FAST qPCRMaster Mix (2X) (Kapa Biosystems), 0.5 μL of BSA (10mg/mL, NewEngland Biolabs), 0.4 μL of each primer 10 μM, PolF 5′- TGC GAY CCSAAR GCB GAC TC-3'and PolR 5′-ATS GCC ATC ATY TCR CCG GA-3’(Poly et al., 2001) and using the following thermal profile described inthe supplementary material. Plate readings were performed 1s at 72 °Cat the end of elongation step and every 0.5 °C during melting curve from65 °C to 95 °C (Mao et al., 2011). The specificity of PCR was evaluatedby running products on a 1% (w/v) agarose gel. Purified PCR productsfrom a common DNA from all samples were used to prepare quantifi-cation standards.

2.4.2. 16S tag sequencing and data sequence analysisTag sequencing was performed at the Centre for Comparative

Genomics and Evolutionary Bioinformatics, Dalhousie University,through the Integrated Microbiome Resource. Amplicon fragmentswere PCR-amplified from the DNA in duplicate reactions using separatetemplate dilutions (1:1 and 1:10) and Phusion High-Fidelity DNAPolymerase (NEB/Thermo Scientific). A single round of PCR was doneusing Illumina adaptors + indices + specific region primers, targetingthe 16S V6-V8 region (Comeau et al., 2011) with multiplexing. PCRproducts were visually verified running a high-throughput 96-well E-Gel (Thermo Scientific). PCR reactions from the same samples werepooled in one plate, cleaned, and normalized using the high-throughputSequalPrep 96-well Plate Kit (Thermo Scientific). Samples were pooled

Fig. 1. Study site characteristics. (A) An exclusion plot inserted in the semiaridthorn scrub in Fray Jorge National Park, Chile. (B) Schematic representation ofnon-trampled and trampled areas within exclusion plots. In different shades ofgreen, the perennial plant cover, undisturbed soils in light brown and disturbedsoils in dark brown. (Black square subplot, non-trampled area; white squaresubplot, trampled area). (C) Photo of infrequent lichenized-biocrust. (D) VR-PCA for ordination of biological, chemical and physical features of soils fromexclusion and trampling treatments (open circles, non-trampled soils; solidcircles, trampled soils; BD, bulk density; BSC cover, biocrust cover; TC, totalcarbon; TN, total nitrogen; NH4, ammonium; NO3, nitrate; SM, soil moisture;SS, soil stability). Colors of circles follow the assignment of Table 1. (For in-terpretation of the references to color in this figure legend, the reader is referredto the Web version of this article.)

F.D. Alfaro et al. Soil Biology and Biochemistry 124 (2018) 1–10

3

to make one library which was fluorometrically quantified before se-quencing. The library was run on an Illumina MiSeq using 300 bppaired-end V3 chemistry (up to 380 samples were simultaneously run).Further information is provided in supplementary material.

2.5. Statistical analyses

2.5.1. Biogeochemical characterizationA Varimax rotated Principal Component Analysis (VR-PCA,

Legendre and Legendre, 1998) was used to obtain an overview of soilphysical and chemical features across all treatment plots. Raw datawere z-score standardized, and the amount of variance explained by theprincipal axes estimated using relative Eigenvalues. Permutationalmultivariate analysis of variance (PerManova) was used to assess theeffects of exclusion and trampling on soil features. A general linearmodel (GLM) with two-way factorial design was used to assess the ef-fects of plot type and human disturbance by trampling upon soil bio-logical (i.e., biocrust cover, bacteria and nifH genes abundance), che-mical (i.e., TC, TN and pH) and physical features (i.e., soil moisture,bulk density and soil stability). All data were tested for homogeneity ofvariances and log-transformed when necessary to meet GLM assump-tions. The effects of experimental treatments (i.e., mammal exclusionand trampling) and changes in biocrust cover on soil chemical andphysical features were evaluated using a generalized linear model, withexclusion and trampling as categorical factor and biocrust cover ascontinuous variable. A similar procedure was used to assess the effectsof trampling and changes in biocrust cover upon chemical and physicalsoil properties within exclusion treatments.

2.5.2. Bacterial community structure and composition analysisWe used rarefaction analysis to compare observed OTUs at 97%

similarity (Fig. S3). The ordination patterns of bacterial communitiesand their possible causal relationships to physical and chemical vari-ables were assessed using redundancy analysis (RDA) after a Hellingertransformation (Legendre and Gallagher, 2001). These analyses werecarried out using the VEGAN package of R version 3.2.0 (RDevelopment Core Team, 2011).

Table 1Experimental design and plot characteristics. Mean (±1SE) are given for an-nual plant cover and trail density.

Treatments Code Annul plant cover(%)

Lagomorph trail density (n/m2)

-Lagomorphs/+small mammals (-L/+SM)Non-trampled N1 10.62 (1.49) 0

Trampled T1 3.89 (0.30) 0

-Lagomorphs/-small mammals (-L/-SM)Non-trampled N2 12.99 (0.96) 0

Trampled T2 3.77 (0.80) 0

+Lagomorphs/+small mammals (+L/+SM)Non-trampled N3 5.65 (0.18) 0.64 (0.29)

Trampled T3 1.68 (0.42) 1.60 (0.87)

+Lagomorphs/-small mammals (+L/-SM)Non-trampled N4 4.30 (0.36) 0.60 (0.38)

Trampled T4 1.17 (0.54) 3.20 (1.31)

Fig. 2. Effects of treatments and tramplingupon biological (A–C), chemical (D–F) andphysical (G–I) characteristics of soils. Mean(±1SE) are given for each variable. Colorsof circles and lines follow the assignment ofTable 1 (P-values show the effect of treat-ments and trampling and the interaction in aGeneral Linear Model with factorial design).(For interpretation of the references to colorin this figure legend, the reader is referred tothe Web version of this article.)

F.D. Alfaro et al. Soil Biology and Biochemistry 124 (2018) 1–10

4

3. Results

3.1. Changes in biological properties among treatments

Cover (%) of herbaceous annual plants was significantly higher innon-trampled soils from all plots (Table 1). In trampled soils, herbac-eous plant cover in -L/+SM and -L/-SM was significantly greater thanthat in +L/-SM (F = 6.52; p < 0.004 Table 1). Not surprisingly, signsof lagomorph activity (feces, trails) were higher in trampled soils from+L/+SM and +L/-SM (Table 1). Biocrust cover and numbers of nifHgene copies differed across both exclusion and trampling treatments,whereas bacterial abundance was only affected by trampling (Fig. 2A, Band C). Biocrust cover was higher in non-trampled soils from all plots,except + L/-SM (Fig. 2A). Bacterial abundance was significantly higherin non-trampled soils from -L/+SM (F = 5.74, p = 0.04), -L/-SM(F = 6.14, p = 0.03) and +L/+SM (F = 7.02, p = 0.02). On the otherhand, the abundance of nifH genes declined significantly from non-trampled to trampled soils in plots -L/+SM (F = 5.29, p = 0.04) and+L/+SM (F = 11.79, p = 0.008) and did not change in -L/-SM and+L/-SM (Fig. 2C).

3.2. Changes in physical and chemical soil features

Overall, we found high variability of soil biogeochemical propertiesamong and within plots; differences were particularly large betweentrampled and non-trampled soils. Principal component analysis in-dicated that trampled soils had similar features, regardless of the ex-clusion treatment, while non-trampled soils showed a more dispersedpattern. We note that non-trampled soils from +L/-SM showed soilfeatures that resembled trampled soils from the other treatments(Fig. 1D). Two way PerManova confirmed significant effects of bothexclusion (F=4.61; p= 0.013) and trampling (F= 2.37; p < 0.022),although the interaction was not significant (F=−1.07; p= 0.416).

Soil moisture was higher in non-trampled soils within -L/+SM(F = 9.22, p = 0.01) and -L/-SM (F = 12.13, p = 0.008). Bulk densitywas higher in trampled soil within -L/+SM (F = 10.12, p = 0.01), -L/-SM (F = 5.85, p = 0.04) and +L/-SM (F = 10.96, p = 0.01). Soilstability was higher in non-trampled soils in all plot treatmentsexcept + L/+SM.

Soil TC, TN and NH4-N were significantly affected by mammal ex-clusion and trampling treatments (Fig. 2D–E and Fig. S1A), whilechanges in NO3-N were unrelated to experimental treatments (Fig.

S1B). Soil pH values showed a significant interaction of exclusiontreatment and trampling effects (Fig. 2F). Physical soil properties suchas moisture, bulk density and stability were significantly affected byboth exclusion and trampling disturbances (Fig. 2G–I), Pairwise com-parisons within plots showed that TC and TN were significantly higherin non-trampled soils within -L/+SM (F = 8.69, p = 0.018 andF = 9.43, p = 0.015 respectively), and NH4-N showed larger values intrampled soil of +L/+SM (F = 9.36, p = 0.015). No changes weredetected in soil NO3-N between trampling treatments. Soil pH washigher in trampled soils from +L/+SM (F = 7.25, p = 0.02) and innon-trampled soils within +L/-SM (F = 7.42, p = 0.02). However, for-L/+SM and -L/-SM treatments no significant differences were ob-served (F = 4.92, p = 0.06; F = 0.51, p = 0.49; respectively).

3.3. δ13C and δ15N dynamics

In soils overlain with biocrusts, values of δ13C and δ15N ranged from−28.2 to −25.13‰ and 4.21 to 7.78‰, respectively. The δ13C signalof soils was altered by trampling (Fig. S1C); whereas δ15N changedprimarily across mammal exclusion treatments (Fig. S1D). Pairwisecomparisons between non-trampled and trampled soils within eachexclusion treatment showed that δ13C differed only within -L/-SM.Values of δ13C increased with total soil C (R2=0.51; p < 0.001). Thelowest δ15N and the highest δ13C signals were found in non-trampledsoils from -L/+SM (Fig. 3B), and these treatments also had the highestsoil C and N content (Fig. 2D–E).

3.4. Cascading effects of biocrusts productivity changes

The effects of changes in biocrust cover in interaction with experi-mental treatments (exclusion and trampling) were only detected in thecontents of soil moisture; whereas, changes in soil stability were relatedto each individual predictor variable (Table 2). Decreases in biocrustscover were accompanied with reductions in soil stability, in particularin non-trampled soils within +L/+SM and –L/+SM and in trampledsoil from all treatments (Fig. S2). In this model, variation in soil che-mical features were explained exclusively by exclusion and trampling.Analyses within each exclusion plot revealed that biocrust cover wasrelated to changes in soil stability, and also to changes in soil moisturein -L/+SM (Table 3). On the other hand, as observed for chemicalfeatures in the global analysis, biocrust cover was related to changes ofNO3-N in +L/+SM (Table 3).

Fig. 3. Relationship between isotopic signatures and soil nutrient content. (A) Soil N and (δ15N); and (B) soil C and (δ13C). Continuous line, non-trampled soils;dashed line, trampled soils (in panel 3A); gray line, general trend; open circles, non-trampled soils; solid circles, trampled soils. (-L/+SM, green centroids; -L/-SM,yellow centroids; +L/+SM, blue centroids; +L/-SM, red centroids). (For interpretation of the references to color in this figure legend, the reader is referred to theWeb version of this article.)

F.D. Alfaro et al. Soil Biology and Biochemistry 124 (2018) 1–10

5

3.5. Changes in soil bacterial community structure and composition

Three abundant phyla (Actinobacteria, Bacteroidetes, andProteobacteria) explained 87.8% of the relative abundance of bacteriaacross all soils (86.5% in non-trampled and 89.1% in trampled soils;Fig. 4A). In both, trampled and non-trampled soils, Bacteroidetes werethe most abundant phylum, encompassing 38.1 and 37.4% of theoverall relative abundance respectively, followed by Proteobacteria(24.6 and 27.2%, respectively), and Actinobacteria (23.5 and 24.5%,respectively). Other groups that showed significant differences in re-lative abundance between trampled and non-trampled soils (Fig. S4)were Acidobacteria (5.3% and 3.5%, respectively), Cyanobacteria(0.9% and 1.4%, respectively), Gemmatimonadetes (2.2% and 1.6%,respectively), and Verrucomicrobia (2.1% and 1.8%, respectively).

Actinobacteria and Bacteroidetes abundances were not influencedby exclusion or trampling effects, but their interaction was significant(Fig. 4B–C). For the third abundant group (Proteobacteria), exclusionbut not trampling treatments influenced abundance (Fig. 4D). Pairwisecomparisons between non-trampled and trampled soils within experi-mental plots showed that declines in Actinobacteria abundance weremarginally significant within -L/+SM (F = 4.09, p = 0.077) and sig-nificant within -L/-SM (F = 9.72, p = 0.014), whereas there was nodifference in both + L/+SM and +L/-SM. Bacteroidetes abundancetended to be lower in non-trampled soils within treatments -L/+SMand -L/-SM, and in trampled soils in +L/+SM and +L/-SM, althoughthese patterns were only close to significance (F = 4.66, p = 0.062;F = 3.07, p = 0.117; F = 3.48, p = 0.099; F = 2.33, p = 0.170

Table 2Generalized linear model showing the effects of experimental treatments (exclusion and trampling) and changes in biocrust cover on soil chemical and physicalfeatures. In bold values of log –likelihood (Type I) with statistically significant effect at P < 0.05.

Treatments gl TC TN NH4 NO3 pH SM BD SS

Exclusion 3 −10.38 81.44 −107.09 −131.27 11.42 145.67 69.67 −23.85Trampling 1 −7.42 84.21 −103.22 −131.27 13.06 151.17 81.59 −3.86Biocrust cover 1 −7.15 84.52 −101.41 −130.17 13.12 153.09 81.53 10.26Exclusion* Trampling 3 0.48 91.51 −96.64 −129.70 18.86 159.60 83.60 35.23Exclusion* Biocrust cover 3 0.84 91.87 −95.90 −127.14 20.55 162.24 85.24 36.35Trampling*Biocrust cover 1 0.91 92.30 −94.88 −126.67 20.64 163.23 86.03 36.56Exclusion* Trampling* Biocrust cover 3 1.28 93.04 −92.16 −125.60 21.28 170.80 87.42 37.52

Table 3Generalized linear model showing the effects of trampling and biocrust coveron soil chemical and physical features within each exclusion treatment. In boldvalues of log –likelihood (Type I) with statistically significant effect atP < 0.05.

Treatments gl TC TN NH4 NO3 pH SM BD SS

-L/+SMTrampling (T) 1 −5.3 19.1 −24.4 −24.7 1.3 35.9 20.2 4.1Biocrust cover

(Bc)1 −5.3 19.3 −23.4 −23.6 1.5 37.1 20.8 11.2

T*Bc 1 −5.2 19.3 −23.4 −23.5 1.5 41.0 20.9 11.3

-L/-SMTrampling (T) 1 −3.2 19.8 −22.7 −36.2 7.0 45.4 19.6 3.4Biocrust cover

(Bc)1 −3.2 20.0 −22.5 −35.7 7.8 45.4 19.9 6.4

T*Bc 1 −3.2 20.3 −22.4 −35.2 7.9 45.5 20.8 6.4

+L/+SMTrampling (T) 1 7.4 29.4 −26.9 −40.5 9.9 45.1 26.4 7.5Biocrust cover

(Bc)1 8.2 29.8 −26.4 −38.1 10.0 45.4 26.8 9.5

T*Bc 1 8.8 30.5 −25.3 −37.1 11.0 45.3 28.5 10.3

+LM/-SMTrampling (T) 1 3.3 26.0 −24.8 −29.1 5.9 43.6 22.3 3.6Biocrust cover

(Bc)1 4.4 26.9 −24.1 −28.8 6.1 44.2 22.6 11.3

T*Bc 1 4.6 27.0 −19.7 −28.8 6.5 44.3 22.9 11.3

Fig. 4. Effect of exclusion and trampling treatments on the relative abundance of bacterial groups. (A) Relative abundance of the three major taxonomic phyla (i.e.,Actinobacteria [purple colors], Bacteroidetes [green colors] and Proteobacteria [yellow colors]) at the class level. (T, trampled; N, non-trampled). (B) Changes inabundance at phylum level for Actinobacteria, (C) Bacteroidetes, and (D) Proteobacteria. (E) Changes in abundance at class level for Actinobacteria, (F)Sphingobacteria, and (G) Alphaproteobacteria. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of thisarticle.)

F.D. Alfaro et al. Soil Biology and Biochemistry 124 (2018) 1–10

6

respectively). On the other hand, increases in Proteobacteria abundancewere only significant within treatment -L/+SM (F = 5.63, p = 0.045).Mammalian exclusions affected Actinobacteria class abundance(Fig. 4E), with significant differences between -L/-SM and +L/-SM,whereas no relationship was observed for Sphingobacteria (Fig. 4F) orAlphaproteobacteria (Fig. 4G). Pairwise comparisons showed that onlyAlphaproteobacteria abundance within -L/+SM was higher in non-trampled than in trampled soils (F = 7.34, p = 0.024).

Bacterial community structure and composition was also affected byexperimental exclusions and trampling effects (PerManova,Fexclusion= 2.74; p < 0.001; Ftrampling = 2.10; p = 0.008). Pairwisecomparisons between non-trampled and trampled soils within eachexclusion treatment revealed that overall bacterial composition differedsignificantly within +L/+SM and +L/-SM (Table S2). In turn, re-dundancy analysis showed that physical soil changes associated withtrampling disturbance were the major drivers of changes in bacterialcommunity composition (Fig. 5A). At the phylum level, increases inBacteroidetes abundance, in both non-trampled and trampled soils,were negatively correlated with reductions in the relative abundancesof Actinobacteria (Fig. 5B) and Proteobacteria (Fig. 5C).

4. Discussion

We provide here substantial evidence that the experimental exclu-sion of native small mammals from semiarid sites and the invasion bylagomorphs (hare and rabbits) enhanced the negative effects of tram-pling on soil biocrust cover and the disruption of bacterial communitystructure and function. These effects were associated with large de-clines in soil physical properties that sustain primary productivity suchas soil moisture and stability. These results are similar to previouslyreported patterns, where trampling and grazing by invasive mammalspecies had large and disruptive effects on physical and chemical soilproperties, including changes in soil biogeochemistry and edaphic ag-gregation (Eldridge and Myers, 2001; Eldridge and Koen, 2008). Ourfindings emphasize the individual and additive detrimental impacts ofbiodiversity loss and species invasions on biocrust structure and soilbiogeochemistry, in particular when these processes interact with per-sistent trampling disturbance, another disruptive factor.

As expected, impacts of trampling on physical and chemical soilfeatures were higher in undisturbed plots (control treatment, –L/+SM)and decreased in mammal-disturbed soils in the follow sequence: -L/+SM > -L/-SM>+L/+SM>+L/-SM. Changes in biocrust cover

and bacterial and nifH gene abundance were accompanied by largechanges in soil physical properties (Fig. 2). In contrast, declines inbiocrust cover across mammal exclusions and trampled soils, was notassociated with changes in soil chemical properties. Exclusion andtrampling treatments had different effects on soil chemical featuresmeasured in this study, whereas soil physical conditions showed similarpatters across treatments. The specific and interactive effects ofmammal exclusion and human trampling are discussed below.

4.1. Reductions in biocrust cover and bacterial abundance and keyecosystem functions

4.1.1. Biocrust coverThe large decline in biocrust cover documented in all trampled soils

(except within +L/-SM) suggests that trampling may not only impactcover, but also drive important changes in ecosystem components orprocesses associated with biocrusts, including their structure and or-ganization, as shown by other studies (Warren, 2003; Kuske et al.,2012; Faist et al., 2017). The decline of biocrusts in non-trampled soilswithin -L/+SM and -L/-SM relative to + L/-SM partially confirmed ourthird hypothesis, which postulated that biocrust cover from trampledand non-trampled soils will decline within treatments with lagomorphs,where native mammals were excluded. We note that no differences inbiocrust cover were observed among trampled soils from all exclusiontreatments. Changes in biocrust cover were accompanied by largesuccessional changes in bacterial community structure, with dark cya-nobacteria-dominated biocrusts occurring in non-trampled soils withintreatments -L/+SM and -L/-SM, while trampled soils from all plotsshowed early successional stages of light cyanobacteria and greenalgae. We also observed large openings of bare soil within +L/+SMand +L/-SM. The lower biocrust cover and the lack of differences be-tween non-trampled and trampled soil within +L/-SM suggest an ad-ditive effect of the loss of small mammal biodiversity and concomitantlagomorph invasion (Fig. 2A). Significantly lower values of biocrustcover in non-trampled soils within +L/+SM and +L/-SM compared to-L/+SM and -L/-SM underscore the detrimental impacts of lagomorphson biocrust establishment and succession, particularly when nativesmall mammals are absent (i.e., within +L/-SM). This is presumablybecause lagomorphs dig their warrens in both disturbed and un-disturbed open areas in central Chile (Jaksic et al., 1979; Root-Bernstein and Jaksic, 2015). Their habitat use patterns are supported bythe similar trail densities and other signals of lagomorph activities

Fig. 5. Changes in bacterial composition and abundancein exclusion and trampling treatments. (A) RedundancyAnalysis (RDA) of the effects of soil physical and chemicalproperties upon bacterial community composition (allOTUs) across exclusion and trampling treatments. (B)Correlation analysis for Bacteroidetes and Actinobacteria.(C) Correlation analysis for Bacteroidetes andProteobacteria. (Open circles: non-trampled soils, Solidcircles: trampled soils).

F.D. Alfaro et al. Soil Biology and Biochemistry 124 (2018) 1–10

7

when comparing + L/+SM and +L/-SM (Table 1).

4.1.2. Bacterial abundance and functionIn addition to conspicuous effects of mammal exclusion and tram-

pling on bacterial abundances, we also found strong interactive effectson bacterial functions. Abundance of the nifH gene was, on average,1.2-fold higher in non-trampled than in trampled soils within -L/+SM,-L/-SM and +L/+SM. Moreover, the difference observed betweentrampled and non-trampled soils within +L/-SM indicates that nifHgene abundance is also responsive to the loss of small mammals and theinvasion of lagomorphs. This last result supports the additive nature ofthe different drivers.

4.2. Cascading effects on biogeochemical cycles and soil stability

Exclusion and human-trampling effects on soil physical features hadindirect effects mediated by changes in biocrust cover (Table 2). Lowerbiocrust cover in trampled soils was significantly related to reduced soilstability in all plots (Fig. S2). These patterns suggest that reductions inbiocrust cover in non-trampled soils had greater impacts on soil stabi-lity than changes in biocrust in trampled soils. In other words, lowerbiocrust cover in non-trampled soils from both + L/+SM and +L/-SMplots (potentially determined by lagomorph activities) could have si-milar impacts as persistent human trampling. The cascading effects ofdecreases in biocrust productivity on soil stability reported in this studyhave been preciously observed in semiarid environments, where in-duced disturbance drives large changes in soil physical properties viachanges in biocrust cover (Eldridge and Leys, 2003; Warren, 2003; Faistet al., 2017).

Despite weak effects of small mammal exclusion treatments andtrampling on soil chemical characteristics in this semiarid ecosystem,we observed declines in soil pH within -L/+SM and +L/-SM, plots thatlacked small mammals (Fig. 2F). This was surprising, in plots with la-gomorphs, as their activity tends to bring buried calcareous layers tothe surface, thereby increasing soil pH (Eldridge and Myers, 2001;Eldridge and Koen, 2008). The differences observed in NH4-N betweentreatments were as expected determined by high heterogeneity andhigh levels of NH4-N in trampled soils, given the lagomorph depositionsof urine and feces on soils in +L/+SM plots (Daryanto et al., 2013).

4.3. Proportional changes in soil features across disturbance treatments

The above results provide partial support to our first hypothesisstating that edaphic conditions (i.e., chemical and physical) will changein response to native mammal exclusion and trampling effects. Thesefindings are summarized in Fig. 6. When comparing between trampledand non-trampled soils, the largest chemical and physical differences insoil properties were observed within -L/+SM plots, where other typesof disturbance (e.g., lagomorphs) were absent. This suggests thathuman-trampling played a proportionally larger role in modifyingbiogeochemical and physical aspects of less disturbed soils under nat-ural conditions (i.e., -L/+SM). In contrast, the smaller differences insoil properties between trampled and non-trampled soils within +L/+SM and +L/-SM plots, suggest that the presence of exotic lago-morphs alters soil bulk density and reduces soil moisture, even in non-trampled sites. For example, stability of non-trampled soils within +L/+SM and +L/-SM was significantly lower than non-trampled soilsfrom -L/+SM and -L/-SM (Fig. 2F). The exclusion of small mammalsalone, in the absence of lagomorphs (i.e., -L/-SM), resulted in greatlyreduced differences in chemical conditions between trampled and non-trampled soils. This difference could be associated with the absence ofthe native herbivore rodent (i.e., O. degus), which often feeds and de-fecates in non-trampled areas. Accordingly, the individual and com-bined effects of species loss and invasions promote substantial changein soil physical and chemical properties that amplify the negative ef-fects of trampling.

4.4. Changes in soil isotopic composition

Variation in stable N isotopes in soils can indicate alterations in therelative contribution of cyanobacterial N-fixation (Aranibar et al.,2003). As biocrust species vary in their ability to fix N, this may alsoreflect changes in biocrust community composition (Aranibar et al.,2004; Barger et al., 2016). The high 15N/14N ratios in all sites suggestthat, in general, biocrusts are making small contributions to fixed N.Similar patterns occur in the Mojave Desert (Billings et al., 2003).However, we failed to find consistent differences in δ15N betweentrampled and non-trampled soils across exclusion treatments, althoughthe δ15N in soils from +L/+SM were significantly higher than soilsfrom -L/+SM (Fig. S2).

Cyanobacterial communities often show δ13C signatures close to−12‰, similar to C4 plants (Palmqvist, 1993; Maguas et al., 1995),whereas mosses and lichens generally have δ13C signatures rangingfrom −23 to −26‰, more similar to C3 plants (Evans and Johansen,1999). The δ13C signatures of lichens often vary among species, andchanges are dependent on photobiont type (Beck and Mayr, 2012) andhabitat (Batts et al., 2004). Because lichen cover was low in these plots,the more negative δ13C signatures observed in non-trampled than intrampled soils within -L/+SM and -L/-SM (Fig. S1C) are more likelyrelated to the changes in biocrust biomass (Fig. 2A).

4.5. Changes in bacterial community composition and dominance undermammals exclusion and trampling treatments

Contrary to our prediction, bacterial community composition dif-fered more between trampled and non-trampled soils where lago-morphs were present (+L/+SM and +L/-SM) (Fig. 5A and Table S2).Surprisingly, bacterial composition in plots without lagomorphs (-L/+SM and -L/-SM) remained similar between trampled and non-tram-pled soils (Figs. 1 and 2). Although soil pH is considered a good pre-dictor of changes in soil bacterial community composition and diversity(Fierer and Jackson, 2006), in our study area this factor had no effectson bacterial patterns. Large differences in soil bacterial compositionbetween trampled and non-trampled soils within exclusion treatments(+L/+SM and +L/-SM), suggest that after prolonged soil disturbance,different soil bacterial communities may develop, as seen with somebiocrust communities (Concostrina-Zubiri et al., 2014). Alternatively,our results could reflect different bacterial strategies for acquiringlimited resources in trampled versus non-trampled soils as in otherwater-limited environments (Kuske et al., 2012; Steven et al., 2015).

In the semiarid system studied, all bacterial communities were

Fig. 6. Overview of proportional changes in chemical and physical features,and bacterial abundance and composition, between non-trampled and trampledsoils across all four treatments of exclusion. (-L/+SM, green blocks; -L/-SM,yellow blocks; +L/+SM, blue blocks; +L/-SM red blocks). Differences be-tween samples were extracted from PCA for each latent (i.e., chemical, physicaland biotic) variable. (For interpretation of the references to color in this figurelegend, the reader is referred to the Web version of this article.)

F.D. Alfaro et al. Soil Biology and Biochemistry 124 (2018) 1–10

8

dominated (> 85%) by three major groups (Actinobacteria,Bacteroidetes and Proteobacteria). These results are similar to thosereported by Makhalanyane et al. (2015) in a worldwide survey of hotdeserts, where these three phyla were dominant across micro-habitats,particularly in the Atacama Desert just north of our study site. Ex-perimental and observational evidence suggest that these three phylaoften exclude other clades, partly due to higher growth rates (Fiereret al., 2007; Goberna et al., 2014). Our findings are partially consistentwith this statement because Bacteroidetes outnumbered Actinobacteriaand Proteobacteria, particularly in more natural settings (non-trampledsoils with native herbivores present, but lacking lagomorphs [-L/+SMand -L/-SM], Fig. 4). Interestingly, changes in abundance of Bacter-oidetes among exclusion treatments were negatively correlated withabundances of Actinobacteria and Proteobacteria (Fig. 5B and C). No-tably, the lack of relationship between changes in soil variables and theabundances of these three phyla, suggests that Bacteroidetes may in-teract negatively with Actinobacteria and Proteobacteria; however,such conclusion should be treated cautiously when considering patternsat the phylum level.

5. Conclusions

In Chilean arid and semiarid environments, native small mammals,such as the abundant O. degus, tend to restrict their activities to areasbeneath shrub canopies to escape from predation. In contrast, twoexotic lagomorph species recently introduced in this ecosystem, due tolower predator pressure, tend to use open areas between bushes morefrequently (Jaksic et al., 1979). As a consequence, soil recovery afterprolonged human-trampling in open areas could be greatly delayed bythe activities of these common exotic species. Our results point stronglyto the fact that, in contrast to exotic lagomorphs, soil heterogeneitycaused by the activity of native small mammals favors the developmentof biocrust cover, enhancing bacterial abundance, which in turn, in-creases nutrient availability and soil stability. The experimental long-term exclusion of native small mammal species and the novel activitiesof invasive exotic mammals exacerbate the negative effects that per-sistent trampling has on functionally important soil bacterial commu-nities in this semiarid thorn scrub.

Acknowledgements

We thank Renzo Vargas, Juan Monardez and Martina Le-bert forfield assistance and/or laboratory work. F.D.A. received support fromFONDECYT 3150179. S.A. and N.T. were supported by FONDECYT1170995. F.D.A., P.A.M., J.R.G., and J.J.A. were supported by GrantsICM-MINECON, P05-002 and CONICYT PFB-0023 (Chile) to theInstitute of Ecology and Biodiversity. J.R.G. thanks support ofFONDECYT 1160026. D.A.K. and P.L.M. were supported by NSF grant1456729. J.B. is supported by the USGS Land Change and Ecosystemsprograms. Research reported here is a contribution to the program ofthe Long-Term Socio-Ecological Research Network- Chile (LTSER-Chile). Any use of trade, firm, or product names is for descriptivepurposes only and does not imply endorsement by the U.S.Government.

Conflicts of interest

The authors declare no conflict of interest.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.soilbio.2018.05.019.

References

Acosta-Jamett, G., Gutiérrez, J.R., Kelt, D.A., Meserve, P.L., Previtali, M.A., 2016. El NiñoSouthern Oscillation drives conflict between wild carnivores and livestock farmers ina semiarid area in Chile. Journal of Arid Environments 126, 76–80.

Aguilera, L.E., Gutierrez, J.R., Meserve, P.L., 1999. Variation in soil micro-organisms andnutrients underneath and outside the canopy of Adesimia bedwellii (Papilionaceae)shrubs in arid coastal Chile following drought and above average rainfall. Journal ofArid Environments 42, 61–70.

Aguilera, L.E., Armas, C., Cea, A.P., Gutierrez, J.R., Meserve, P.L., Kelt, D.A., 2016.Rainfall, microhabitat, and small mammals influence the abundance and distributionof soil microorganisms in a Chilean semi-arid shrubland. Journal of AridEnvironments 126, 37–46.

Aranibar, J.N., Anderson, I.C., Ringrose, S., Macko, S.A., 2003. Importance of nitrogenfixation in soil crusts of southern African arid ecosystems: acetylene reduction andstable isotope studies. Journal of Arid Environments 54, 345–358.

Aranibar, J.N., Otter, S.A., Macko, C.J., Feral, H.E., Epstein, P.R., Dowty, F., Eckardt,H.H., Swap, R.J., 2004. Nitrogen cycling in the soil-plant system along a precipitationgradient in the Kalahari sands. Global Change Biology 10, 359–373.

Araya, P.I., Benoit, J., Cerda, B., Contreras, R., Cuevas, R., Hernández, R., Layana, A.,Noton, E., Peña, E., Soto, A., 1992. Plan de manejo Parque Nacional Fray Jorge.(Programa patrimonios silvestres, región Coquimbo). Corporación Nacional ForestalIV, Santiago, Chile.

Armas, C., Gutiérrez, J.R., Kelt, D.A., Meserve, P.L., 2016. Twenty-five years of researchin the north-central chilean semiarid zone: the fray Jorge long-term socio-ecologicalresearch (LTSER) site and Norte Chico. Journal of Arid Environments 126, 1–6.

Bardgett, R.D., Wardle, D.A., 2003. Herbivore-mediated linkages between abovegroundand belowground communities. Ecology 84, 2258–2268.

Barger, N.N., Herrick, J.E., Van Zee, J.W., Belnap, J., 2006. Impacts of biological soil crustdisturbance and composition on C and N losses from water erosion. Biogeochemistry77, 247–263.

Barger, N.N., Weber, B., Garcia-Pichel, F., Zaady, E., Belnap, J., 2016. Patterns andcontrols on nitrogen cycling of biological soil crusts. In: In: Weber, B., Büdel, B.,Belnap, J. (Eds.), Biological Soil Crusts: an Organizing Principle in Drylands.Ecological Studies (Analysis and Synthesis), vol. 226 Springer, Cham.

Batts, J.E., Calder, L.J., Batts, B.D., 2004. Utilizing stable isotope abundances of lichens tomonitor environmental change. Chemical Geology 204, 345–368.

Beck, A., Mayr, C., 2012. Nitrogen and carbon isotope variability in the green-algal lichenXanthoria parietina and their implications on mycobiont–photobiont interactions.Ecology and Evolution 2, 3132–3144.

Belnap, J., Lange, O.L., 2003. Biological Soil Crusts: Structure, Function andManagement. Springer-Verlag, Berlin, pp. 503.

Belnap, J., 2006. The potential roles of biological soil crusts in dryland hydrologic cycles.Hydrological Processes 20, 3159–3178.

Belnap, J., Weber, B., Büdel, B., 2016. Biological soil crusts as an organizing principle indrylands. In: In: Weber, B., Büdel, B., Belnap, J. (Eds.), Biological Soil Crusts: anOrganizing Principle in Drylands. Ecological Studies (Analysis and Synthesis), vol.226 Springer, Cham.

Billings, S.A., Schaeffer, S.M., Evans, R.D., 2003. Nitrogen fixation by biological soilcrusts and heterotrophic bacteria in an intact Mojave Desert ecosystem with elevatedCO2 and added soil carbon. Soil Biology and Biochemistry 35, 643–649.

Castillo-Monroy, A.P., Maestre, F.T., 2011. La costra biologica del suelo: avances recientesen el conocimiento de su estructura y funcion ecologica. Revista Chilena de HistoriaNatural 84, 1–21.

Chamizo, S., Canton, Y., Rodrıguez-Caballero, E., Domingo, F., Escudero, A., 2012. Runoffat contrasting scales in a semiarid ecosystem: a complex balance between biologicalsoil crust features and rainfall characteristics. Journal of Hydrology 452, 130–138.

Comeau, A.M., Li, W.K.W., Tremblay, J.E., Carmack, E.C., Lovejoy, C., 2011. Arctic Oceanmicrobial community structure before and after the 2007 record sea ice minimum.PLoS One 6, e27492.

Concostrina-Zubiri, L., Huber-Sannwald, E., Martinez, E., Flores, J.L., Reyes-Agüero, J.A.,Escudero, A., Belnap, J., 2014. Biological soil crusts across disturbance – recoveryscenarios: effect of grazing regime on community dynamics. Ecological Applications24, 1863–1877.

Daryanto, S., Eldridge, D.J., Wang, L., 2013. Ploughing and grazing alter the spatialpatterning of surface soils in a shrub-encroached woodland. Geoderma 200, 67–76.

Delgado-Baquerizo, M., Maestre, F.T., Gallardo, A., 2013. Biological soil crusts increasethe resistance of soil nitrogen dynamics to changes in temperatures in a semi-aridecosystem. Plant and Soil 366, 35–47.

Einen, J., Thorseth, I.H., Ovreas, L., 2008. Enumeration of Archaea and Bacteria in sea-floor basalt using real-time quantitative PCR and fluorescence microscopy. FEMSMicrobiology Letters 282, 182–187.

Elbert, W., Weber, B., Burrows, S., Steinkamp, J., Büdel, B., Andreae, M.O., Pöschl, U.,2012. Contribution of cryptogamic covers to the global cycles of carbon and nitrogen.Nature Geoscience 5, 459–462.

Eldridge, D.J., Myers, C.A., 2001. The impact of warrens of the European rabbit(Oryctolagus cuniculus L.) on soil and ecological processes in a semi-arid Australianwoodland. Journal of Arid Environments 47, 325–337.

Eldridge, D.J., Leys, J.F., 2003. Exploring some relationships between biological soilcrusts, soil aggregation and wind erosion. Journal of Arid Environments 53, 457–466.

Eldridge, D.J., Koen, T., 2008. Formation of nutrient-poor patches in a semi-arid wood-land by the European rabbit (Oryctolagus cuniculus L.). Austral Ecology 33, 88–98.

Eldridge, D.J., Delgado-Baquerizo, M., Woodhouse, J.N., Neilan, B.A., 2016. Mammalianengineers drive soil microbial communities and ecosystem functions across a dis-turbance gradient. Journal of Animal Ecology 85, 1636–1646.

F.D. Alfaro et al. Soil Biology and Biochemistry 124 (2018) 1–10

9

Eldridge, D.J., Delgado-Baquerizo, M., Travers, S.K., Val, J., Oliver, I., 2017. Do grazingintensity and herbivore type affect soil health? Insights from a semi-arid productivitygradient. Journal of Applied Ecology 54, 976–985.

Evans, R.D., Johansen, J.R., 1999. Microbiotic crusts and ecosystem processes. Solicitedreview. Critical Reviews in Plant Sciences 18, 183–225.

Faist, A.M., Herrick, J.E., Belnap, J., Van Zee, J.W., Barger, N.N., 2017. Biological soilcrust and disturbance controls on surface hydrology in a semi-arid ecosystem.Ecosphere 8, e01691.

Ferrenberg, S., Reed, S.C., Belnap, J., 2015. Climate change and physical disturbancecause similar community shifts in biological soil crusts. Proceedings of the NationalAcademy of Sciences of the United States of America 112, 12116–12121.

Fierer, N., Jackson, R.B., 2006. The diversity and biogeography of soil bacterial com-munities. Proceedings of the National Academy of Sciences of the United States ofAmerica 103, 626–631.

Fierer, N., Bradford, M.A., Jackson, R.B., 2007. Toward and ecological classification ofsoil bacteria. Ecology 88, 1354–1364.

Fleming, P.A., Anderson, H., Prendergast, A.S., Bretz, M.R., Valentine, L.E., Hardy, G.E.,2014. Is the loss of Australian digging mammals contributing to a deterioration inecosystem function? Mammal Review 44, 94–108.

Fuentes, E.R., Jaksic, F.M., Simonetti, J.A., 1983. European rabbits versus native rodentsin central Chile: effects on shrub seedlings. Oecologia 58, 411–414.

Goberna, M., Garcia, C., Verdu, M., 2014. A role for biotic filtering in driving phyloge-netic clustering in soil bacterial communities. Global Ecology and Biogeography 23,1346–1355.

Gutiérrez, J.R., Meserve, P.L., Jaksic, J.M., Contreras, L.C., Herrera, S., Vásquez, H., 1993.Structure and dynamics of vegetation in a Chilean semiarid thorn-scrub community.Acta Oecologica 14, 271–285.

Gutierrez, J.R., Meserve, P.L., Contreras, L.C., Vásquez, H., Jaksic, F.M., 1993. Spatialdistribution of soil nutrients and ephemeral plants underneath and outside the ca-nopy of Porlieria chilensis shrubs (Zygophyllaceae) in arid coastal Chile. Oecologia95, 347–352.

Gutierrez, J.R., Meserve, P.L., Herrena, S., Contreras, L.C., Jaksic, F.M., 1997. Effects ofsmall mammals and vertebrate predators on vegetation in the Chilean semiarid zone.Oecologia 109, 398–406.

Gutierrez, J.R., Meserve, P.L., Kelt, D.A., 2004. Estructura y dinámica de la vegetación delecosistema semiárido del Parque Nacional Bosque Fray Jorge entre 1989 y 2002. In:Squeo, F.A., Gutiérrez, J.R., Hernández, I.R. (Eds.), Historia Natural del ParqueNacional Bosque Fray Jorge. Ediciones Universidad de La Serena, La Serena, Chile,pp. 115–134.

Gutierrez, J.R., Meserve, P.L., Kelt, D.A., Engilis, A., Previtali, M.A., Milstead, W.B.,Jaksic, F.M., 2010. Long-term research in Bosque fray Jorge National Park: twentyyears studying the role of biotic and abiotic factors in a Chilean semiarid scrubland.Revista Chilena de Historia Natural 83, 69–98.

Harrison, K.A., Bardgett, R.D., 2008. Impacts of grazing and browsing by large herbivoreson soils and soil biological properties. In: Gordon, I.J., Prins, H.H.T. (Eds.), TheEcology of Browsing and Grazing. Springer, Berlin, pp. 201–216.

Herrick, J.E., Van Zee, J., Havstad, K.M., Burkett, L.M., Whitford, W.G., 2005. Monitoringmanual for grassland, shrubland and savanna ecosystems. Volume I: Quick Start.Volume II: design, supplementary methods and interpretation. Monitoring manual forgrassland, shrubland and savanna ecosystems. Volume I: Quick Start. Volume II:design, supplementary methods and interpretation.

Hooper, D.U., Adair, E.C., Cardinale, B.J., Jarrett, E., Byrnes, K., Hungate, B.A., Matulich,K.L., Gonzalez, A., Duffy, J.E., Gamfeldt, L., O'Connor, M.I., 2012. A global synthesisreveals biodiversity loss as a major driver of ecosystem change. Nature 486, 105–108.

Housman, D.C., Powers, H.H., Collins, A.D., Belnap, J., 2006. Carbon and nitrogen fixa-tion differ between successional stages of biological soil crusts in the ColoradoPlateau and Chihuahuan Desert. Journal of Arid Environments 66, 620–634.

IPCC, 2013. Climate Change 2013: the Physical Science Basis. Contribution of WorkingGroup I to the Fifth Assessment Report of the Intergovernmental Panel on ClimateChange. Cambridge University Press, Cambridge and New York.

Jaksic, F.M., Fuentes, E.R., Yanez, J.L., 1979. Spatial distribution of the old world rabbit(Oryctolagus cuniculus) in central Chile. Journal of Mammalogy 60, 207–209.

Jimenez, M.A., Jaksic, F.M., Armesto, J.J., Gaxiola, A., Meserve, P.L., Kelt, D.A.,Gutiérrez, J.R., 2011. Extreme climatic events change the dynamics and invasibilityof semi-arid annual plant communities. Ecology Letters 12, 1227–1235.

Kuske, C.R., Yeager, C.M., Johnson, S., Ticknor, L.O., Belnap, J., 2012. Response andresilience of soil biocrust bacterial communities to chronic physical disturbance inarid shrublands. The ISME Journal 6, 886–897.

Legendre, P., Gallagher, E.D., 2001. Ecologically meaningful transformations for ordi-nation of species data. Oecologia 129, 271–280.

Legendre, P., Legendre, L., 1998. Numerical Ecology: Second English Edition. Elsevier,Amsterdam, The Netherlands.

Lopez-Cortes, F., Cortes, A., Miranda, E., Rau, J.R., 2007. Dietas de Abrothrix andinus,Phyllotis xanthopygus (Rodentia) y Lepus europaeus (Lagomorpha) en un ambientealtoandino de Chile. Revista Chilena de Historia Natural 80, 3–12.

Maestre, F.T., Eldridge, D.J., Soliveres, S., Kéfi, S., Delgado-Baquerizo, M., Bowker, M.A.,García-Palacios, P., Gaitán, J., Gallardo, A., Lázaro, R., Berdugo, M., 2016. Structureand functioning of dryland ecosystems in a changing world. Annual Review ofEcology, Evolution and Systematics 47, 215–237.

Maguas, C., Griffiths, H., Broadmeadow, M.S.J., 1995. Gas exchange and carbon isotopediscrimination in lichens: evidence for interactions between CO2-concentrating me-chanisms and diffusion limitation. Planta 196, 95–102.

Makhalanyane, T.P., Valverde, A., Gunnigle, E., Frossard, A., Ramond, J.B., Cowan, D.A.,2015. Microbial ecology of hot desert edaphic systems. FEMS Microbiology Reviews39, 203–221.

Mao, Y., Yannarell, A.C., Mackie, R.I., 2011. Changes in N-transforming Archaea andBacteria in soil during the establishment of bioenergy crops. PLoS One 6, e24750.

Meserve, P.L., Kelt, D.A., Gutierrez, J.R., Previtali, M.A., Milstead, W.B., 2016. Bioticinteractions and community dynamics in the semiarid thorn scrub of Bosque FrayJorge National Park, north-central Chile: a paradigm revisited. Journal of AridEnvironments 126, 81–88.

Neely, C., Bunning, S., Wilkes, A., 2009. Review of Evidence on Dryland Pastoral Systemsand Climate Change. Implications and Opportunities for Mitigation and Adaptation.Food and Agriculture Organization of the United Nations Land and Water DiscussionPaper 8.

Neff, J.C., Reynolds, R.L., Belnap, J., Lamothe, P., 2005. Multi-decadal impacts of grazingon soil physical and biogeochemical properties in southeast Utah. EcologicalApplications 15, 87–95.

Ovreas, L., Torsvik, V., 1998. Microbial diversity and community structure in two dif-ferent agricultural soil communities. Microbial Ecology 36, 303–315.

Palmqvist, K., 1993. Photosynthetic CO2-use efficiency in lichens and their isolatedphotobionts: the possible role of a CO2-concentrating mechanism. Planta 191, 48–56.

Pérez, C.A., Hedin, L.O., Armesto, J.J., 1998. Nitrogen mineralization in two unpollutedold-growth forests of contrasting biodiversity and dynamics. Ecosystems 1, 361–373.

Poly, F., Ranjard, L., Nazaret, S., Gourbiere, F., Monrozier, L.J., 2001. Comparison of nifHgene pools in soils and soil microenvironments with contrasting properties. Appliedand Environmental Microbiology 67, 2255–2262.

R Development Core Team, 2011. R: a Language and Environment for StatisticalComputing. R Foundation for Statistical Computing, Vienna, Austria Available at:http://www.R-project.org.

Reid, W.V., Mooney, H.A., Cropper, A., Capistrano, D., Carpenter, S.R., Chopra, K.,Dasgupta, P., Dietz, T., Duraiappah, A.K., Hassan, R., Kasperson, R., Leemans, R.,May, R.M., McMichael, T., Pingali, P., Samper, C., Scholes, R., Watson, R.T., Zakri,A.H., Shidong, Z., Ash, N.J., Bennett, E., Kumar, P., Lee, M.J., Raudsepp-Hearne, C.,Simons, H., Thonell, J., Zurek, M.B., 2005. Millennium Ecosystem AssessmentSynthesis Report. Millennium Assessment and World Resources Institute,Washington, DC.

Root-Bernstein, M., Jaksic, F., 2015. Ecosystem process interactions between centralChilean habitats. Global Ecology and Conservation 3, 776–788.

Stavi, I., Ungar, E.D., Lavee, H., Sarah, P., 2008. Grazing-induced spatial variability of soilbulk density and content of moisture, organic carbon and calcium carbonate in asemi-arid rangeland. Catena 75, 288–296.

Steven, B., Kuske, C.R., Gallegos-Graves, L.V., Reed, S.C., Belnap, J., 2015. Climatechange and physical disturbance manipulations result in distinct biological soil crustcommunities. Applied and Environmental Microbiology 81, 7448–7459.

Warren, S.D., 2003. Synopsis: influence of biological soil crusts on arid land hydrologyand soil stability. In: Belnap, J., Eldridge, D. (Eds.), Biological Soil Crusts: Structure,Function, and Management. Springer International Publishing, Berlin Heidelberg,Germany, pp. 349–360.

F.D. Alfaro et al. Soil Biology and Biochemistry 124 (2018) 1–10

10