the origin and role of biological rock crusts in rocky

Biogeosciences 16 1133ndash1145 2019httpsdoiorg105194bg-16-1133-2019copy Author(s) 2019 This work is distributed underthe Creative Commons Attribution 40 License

The origin and role of biological rock crusts inrocky desert weatheringNimrod Wieler1 Hanan Ginat2 Osnat Gillor1 and Roey Angel31Zuckerberg Institute for Water Research Blaustein Institutes for Desert Research Ben Gurion University of theNegev Sede Boqer Campus Israel2The Dead Sea and Arava Science Center Israel3Soil and Water Research Infrastructure and Institute of Soil Biology Biology Centre CAS Czechia

Correspondence Roey Angel (roeyangelbccascz) and Osnat Gillor (gillorobguaci)

Received 9 October 2018 ndash Discussion started 6 November 2018Revised 18 February 2019 ndash Accepted 27 February 2019 ndash Published 19 March 2019

Abstract In drylands microbes that colonize rock surfaceshave been linked to erosion because water scarcity excludestraditional weathering mechanisms We studied the originand role of rock biofilms in geomorphic processes of hardlime and dolomitic rocks that feature comparable weatheringmorphologies although these two rock types originate fromarid and hyperarid environments respectively We hypothe-sized that weathering patterns are fashioned by salt erosionand mediated by the rock biofilms that originate from the ad-jacent soil and dust We used a combination of microbial andgeological techniques to characterize rock morphologies andthe origin and diversity of their biofilms Amplicon sequenc-ing of the SSU rRNA gene suggested that bacterial diversityis low and dominated by Proteobacteria and ActinobacteriaThese phyla only formed laminar biofilms on rock surfacesthat were exposed to the atmosphere and burrowed up to6 mm beneath the surface protected by sedimentary depositsUnexpectedly the microbial composition of the biofilms dif-fered between the two rock types and was also distinct fromthe communities identified in the adjacent soil and settleddust showing a habitat-specific filtering effect Moreover therock bacterial communities were shown to secrete extracel-lular polymeric substances (EPSs) that form an evaporationbarrier reducing water loss rates by 65 ndash75 The reducedwater transport rates through the rock also limit salt transportand its crystallization in surface pores which is thought tobe the main force for weathering Concomitantly the biofilmlayer stabilizes the rock surface via coating and protects theweathered front Our hypothesis contradicts common mod-els which typically consider biofilms to be agents that pro-

mote weathering In contrast we propose that the microbialcolonization of mineral surfaces acts to mitigate geomorphicprocesses in hot arid environments

1 Introduction

In arid and hyperarid stony deserts bedrock surfaces are typ-ically barren and free of vegetation or continuous soil mantleWhen these surfaces are exposed to atmospheric conditionsthey undergo weathering processes that shape the landscape(Smith 2009) Weathering is an in situ set of processes thatincludes physical chemical and mechanical forces that re-sult in the breakdown and transport of shattered fragments ofthe parent rock Weathering can appear in a range of sizesand morphologies (Smith et al 2005) including gravel shat-tering (Amit et al 1996) surface crazing (Smith 1988)ventifacts (Smith 1988) micro-rills (Smith 1988 Sweetingand Lancaster 1982) and cavernous patterns also known astafoni honeycomb or pitting (Mustoe 1983 Viles 2005)Weathering is an essential although often neglected elementin the overall denudation of hot deserts

Cavernous weathering is one of the most frequently occur-ring weathering patterns that has been observed in variousregions across the globe including humid arid cold hotcoastal and inland sites (Bruthans et al 2018) In the NegevDesert Israel cavernous weathering patterns are common incarbonate rocks in arid and hyperarid regions Upon expo-sure to the atmosphere these rocks develop a carbonate coat-ing termed calcrete or dolocrete (on limestone or dolomite

Published by Copernicus Publications on behalf of the European Geosciences Union

1134 N Wieler et al Biological rock crusts in rocky desert weathering

respectively) by displacive and replacive cementation of cal-cium or dolomite onto the rock surface (Wright and Wacey2004 Alonso-Zarza and Wright 2010) Following the ce-mentation processes typical honeycomb features are formedon the exposed parent rock typified by pits separated bythin walls that are coated by the calcrete or dolocrete Re-cent studies suggest that microbial activity also promotes theprocesses of calcrete and dolocrete formation (Alonso-Zarzaet al 2016 Alonso-Zarza and Wright 2010)

The accepted conceptual model for the formation of cav-ernous rock weathering in hot deserts involves the pres-ence of permeable rocks that are subjected to soluble saltsand repeated episodes of dryingndashrewetting cycles (Goudieet al 2002 Smith 1988 Smith et al 2005) The pro-posed mechanisms assume that cavernous weathering re-sults from physico-chemical processes including salt crys-tallization (Cooke 1979 Scherer 2004) incipient fractures(Amit et al 1996) exfoliation (Shtober-Zisu et al 2017)or stress erosion (Bruthans et al 2014 McArdle and An-derson 2001) Recently Bruthans et al (2018) conclusivelydemonstrated that moisture flux followed by salt crystalliza-tion at the boundary layer govern the case hardening modelin temperate climates In addition biological mechanismshave been proposed to promote rock weathering throughmechanisms such as flaking via colony growth (Viles 2012)acidification by bacterial extractions (Garcia-Pichel 2006Warscheid and Braams 2000) or alkalization during photo-synthesis by cyanobacteria (Buumldel et al 2004) In contrastit has been proposed that micro- and macro-organism col-onization can mitigate weathering in temperate coastal re-gions (McIlroy de la Rosa et al 2014 Mustoe 2010) viaencrustation or protection from direct rain impact Howeverit is not clear which of these mechanisms dominates or whatthe relative contribution of chemical vs biological processesto weathering in arid environments is

Microorganisms colonizing rocks form a hardy biofilmknown as the biological rock crust (BRC) which is com-mon in most arid and hyperarid regions worldwide (Gor-bushina 2007 Lebre et al 2017 Pointing and Belnap2012) Epilithic communities colonizing rock surfaces areubiquitous in arid environments whereas hyperarid rockswhich experience increased radiation and desiccation aredominated by endolithic communities that colonize inter-nal rock pores (Makhalanyane et al 2013 Pointing andBelnap 2012 Viles 1995) The BRC communities includecyanobacteria and other phototrophic and heterotrophic bac-teria but very low abundances of archaea fungi or algae(Lang-Yona et al 2018) However the BRC inoculum hasnot been resolved and has been proposed to originate fromsettled dust (Viles 2008) or the surrounding soil (Makha-lanyane et al 2015)

The goal of this study was to illuminate the origin androle of BRCs in cavernous weathering of exposed lime-stone and dolomite rocks in arid and hyperarid regions Wepredicted that the BRC communities on exposed rock sur-

faces would resemble either the ever-present dust or the sur-rounding soil supporting a subset of adapted taxa from bothsources We further hypothesized that the cavernous weath-ering morphologies of exposed rocks result from salt mobi-lization by dew causing crystallization pressure under atmo-spheric conditions The developed rock biofilms clog the sur-face rock pores through secretion of extracellular polymericsubstances (EPSs) lowering evaporation and slowing the saltcrystallization but also stabilizing the exfoliated rocks andpreventing further weathering Thus the presence of a BRCmitigates the geomorphic processes To test our hypotheseswe applied a holistic approach combining field observationsand geological geotechnical and molecular microbiologycharacterization to elucidate the morphology origin and roleof BRCs in arid cavernous weathering

2 Materials and methods

21 Study site

We focused on two sites in the Negev Desert Israel Sede-Boqer ndash an arid site and Uvda Valley ndash a hyperarid site(Fig S1 Table S1 in the Supplement) Both sites have rockyterrains underlaid predominantly by carbonate rock slopesconsisting of limestone dolomite chalk marl clay and chertfrom the Cretaceous period to the Eocene Our analyses com-pared samples from the limey Turonian age Shivta Formationlocated in the arid region with samples from the dolomiticTuronian age Gerofit Formation located in the hyperarid en-vironment The Negev Desert Israel has maintained arid tohyperarid conditions since the Holocene and has an arid-ity index (PPET) of 005ndash0005 (Amit et al 2010 Bru-ins 2012) which is similar to other arid and hyperarid ar-eas worldwide eg the Namib and Atacama deserts (Azua-Bustos et al 2012 Viles and Goudie 2007) The long-termaridity of the Negev Desert makes it a reliable site for testingthe relationships between BRCs and geological substrates

22 Field sampling

A total of 24 rock samples were collected along rockyslopes facing northward during November and December2014 these samples were comprised of 12 limestone sam-ples from the limey Turonian age Shivta Formation at the aridsite (3088 N 3478 E WGS 84 Grid the samples werenamed SB 1ndash12) and 12 dolomite samples from the limeydolomitic Turonian age Gerofit Formation at the hyperaridsite (2994 N 3497 E WGS 84 Grid the samples werenamed UV 1ndash12) Concomitantly six soil samples (ca 500 geach) were collected three samples from the arid site (re-ferred to as SBSoil 1ndash3) and three samples from the hyper-arid (referred to as UVSoil 1ndash12) site Each rock or soil sam-ple was a composite of four sub-samples that were pooledand homogenized in the lab

Biogeosciences 16 1133ndash1145 2019 wwwbiogeosciencesnet1611332019

N Wieler et al Biological rock crusts in rocky desert weathering 1135

We also collected settled dust samples using glass beadtraps (Goossens and Rajot 2008) The traps were placed dur-ing a dust storm on December 2013 and collected 3 monthslater at both the arid (the samples were named SBDust 1ndash2)and hyperarid (the samples were named UVDust 1ndash2) sitesIn total two dust samples were collected per site Each dustsample was a composite of two sub-samples that were pooledand homogenized in the lab

23 Geological analyses

The geological methods used in this study are based on di-rect field observations and a detailed characterization of thesubject lithologies (ie limestone and dolomite) which in-cluded the morphology (thin sections) mineral components(X-ray powder diffraction ndash XRD) porosity and permeabil-ity (automatic gas permeameter porosimeter) and the elas-tic properties (Schmidt hammer) Petrographic thin sections30 microm thick were prepared for each lithology to test themain components in both the BRC and the host rocks andthe sections were then examined under a light microscope(Zeiss Oberkochen Germany) The XRD analysis of min-eral components (Sandler et al 2015) was conducted on theBRC and host rocks using three replicates of each Powderedsamples were scanned using an XrsquoPert3 Powder diffractome-ter equipped with a PIXcel detector (Malvern PanalyticalAlmelo Netherlands) the scanning range was 3ndash70 2θ with a step size of 0013 and a speed of 701 s per stepThe total effective porosity (8) and the permeability (k) tests(Scherer 1999) were performed on 12 rock core cylindersamples that had a 185 mm radius and 265 mm height usingan automatic gas permeameter porosimeter (Core Laborato-ries Houston Texas USA) Six cores were prepared fromeach lithology and from the six cores three were cut paral-lel and three were cut perpendicular to the bedding planeNext 5 cm layer was mechanically removed from the top ofthe core using a diamond saw (Dremel Racine WI USA)to ensure the complete removal of the BRC Before testingporosity and permeability samples were oven dried at 110 Cfor 24 h Schmidt hammer (Lassen Aarhus Denmark) testswere applied in the field (Goudie 2016 Viles et al 2011) Atotal of 20 measurements were carried out for each lithology

24 FTIR and stable isotope analysis

Fourier transform infrared spectroscopy (FTIR) analysis wasconducted to test for the presence of extracellular polymericsubstances (EPSs) on the rock surfaces and the host rockwas used for comparison The spectra were recorded using aVertex 70 FTIR spectrometer (Bruker Billerica MA USA)with a 4 cmminus1 scan resolution A total of 1ndash2 mg of pulver-ized rock was taken from each sample (n= 2) and the spec-tra were measured twice the spectra were collected over awavenumber range from 4000 to 600 cmminus1 and a baselinecorrection was carried out The spectral absorption bands in-

dicative for EPSs were identified according to published in-formation (Ferrando et al 2018)

For δ13C and δ18O analyses 1ndash2 mg of rock surface pow-der (ie calcite or dolomite) was obtained using a micro drill(Dremel Racine WI USA) along with a cross section ofthe rock crust and its host rock Four profile measurementsof δ13C and δ18O were performed on samples UVSL 5 andUVSL 6 from the hyperarid site and samples NWSH 1 andNWSH 2 from the arid site Measurements (in duplicate) ofδ18O-H2O and δ13C-DIC were performed using a gas sourceisotope ratio mass spectrometer (GS-IRMS Thermo FisherScientific Waltham MA USA) coupled to a Gas Bench IIinterface (Thermo) after CO2 equilibration or CO2 extrac-tion by acidification for δ18O-H2O and δ13C-DIC respec-tively The samples were calibrated against internal labora-tory standards Vienna Standard Mean Ocean Water (VS-MOW) and NBS-19 carbonate standard δ13C values werealso referenced against VSMOW and valued for carbonaterelative to the Vienna PeeDee Belemnite (VPDB) standardas previously described (Uemura et al 2016) with a SD of01 permil All values are reported in per mille (permil)

25 Desiccation experiment

To test the effect of BRCs on water transport rates in therock clogging and desiccation experiments were performedon 16 rock core cylinders from both lithologies (limestoneand dolomite) Rock cylinders (with a radius of 37 mm andheight of 65 cm) were drilled using a rock core drill Each setof eight rock cores from the two different lithologies includedfour rock cores that were kept intact and four rock cores thathad their BRCs mechanically removed using a diamond saw(Dremel Racine WI USA) to a depth of 5 cm Each cylinderwas immersed in distilled water for 72 h Next the entire sur-face of the cylinder except the upper base was covered withepoxy (Devcon) and then with aluminium foil leaving onlythe top of the cylinders (with or without BRC) uncovered toallow evaporation The cylinders were then weighed (t = 0)incubated in an oven (dried) at a temperature of 44 C for48 h and weighed every 2 h during the first 12 h and thenevery 6 h to determine the residual water content Second-degree polynomial functions were fitted using the lmstatsfunction to determine evaporation rates and then comparedusing an ANOVA (both analyses were carried out in R)

26 DNA extraction PCR amplification and sequencing

For DNA extraction from all rocks the surface (ca 100 cm2)was scraped using a rasp (66ndash67 HRC hardness DieterSchmid Berlin Germany) that was cleaned with 70 technical-grade ethanol before each sampling DNA wasthen extracted from 05 g of the homogenized sample us-ing bead-beating in the presence of a CTAB buffer andphenol according to a previously published extraction pro-tocol (Angel et al 2012) A 466-bp fragment of the

wwwbiogeosciencesnet1611332019 Biogeosciences 16 1133ndash1145 2019


16S rRNA gene was amplified using the universal bacte-rial primers 341F (CCTAYGGGRBGCASCAG) and 806R(GGACTACNNGGGTATCTAAT) flanking the V3 and V4regions (Klindworth et al 2012) Library construction andsequencing were performed at the DNA Services Facility atthe University of Illinois Chicago using an Illumina MiSeqsequencer (Illumina San Diego CA USA) in the 2times 250cycle configuration (V2 regent kit)

27 Sequence processing and analysis of bacterialcommunities

Paired reads generated by the MiSeq platform were qual-ity filtered and clustered into operational taxonomic units(OTUs) using the UPARSE pipeline (Edgar 2013) withmodifications Contig assembly was carried out using thefastq_mergepairs command Contigs were then dereplicatedwith the derep_fullength command and singleton sequenceswere removed Following this OTU centroids were deter-mined with the cluster_otus command (set at 3 radius)Abundances of OTUs were determined by mapping the fil-tered contigs (before dereplication including singletons) tothe OTU centroids using the usearch_global command (setat 097 identity) Following these steps a total of ca 14 Greads remained OTU representatives were classified usingthe mothur software packagersquos implementation of a naiumlveBayesian sequence classifier (Schloss et al 2009 Wang etal 2007) against the SILVA 119 SSU NR99 database (Quastet al 2013) All downstream analyses were performed in RV344 (R Core Team 2016) Data handling and manipula-tion were undertaken using the phyloseq package (McMur-die and Holmes 2013) For α-diversity analysis all sam-ples were subsampled (rarefied) to the minimum sample sizeusing bootstrap subsampling at 1000 iterations to accountfor library size differences whereas for β-diversity analy-sis library size normalization was carried out using the geo-metric mean of pairwise ratios (GMPR Chen et al 2018)The abundance-based coverage estimator (ACE) richness es-timate (OrsquoHara 2005) and the Shannon diversity index (H)were calculated using the estimateR function in the veganpackage (Oksanen et al 2018) and tested using an ANOVAand a Tukey HSD test in the stats package Variance par-titioning and testing were undertaken using PERMANOVA(McArdle and Anderson 2001) with the vegan-functionadonis using HornndashMorisita distances Differences in phylacomposition between the sample types were tested using thenon-parametric ScheirerndashRayndashHare test (Mangiafico 2018the scheirerRayHare function in the rcompanion package)followed by the post hoc MannndashWhitney test (the wilcoxtestfunction in the stats package) and FDR corrected usingthe BenjaminindashHochberg method (Ferreira and Zwinderman2006 the padjust function in the stats package) Detection ofdifferentially abundant OTUs was carried out using ALDEx2(Fernandes et al 2014) Plots were generated using the gg-

plot2 (Wickham 2016) and ggtern (Hamilton 2017) pack-ages

3 Results and discussion

31 Field and mineralogical observations

Weathering features were observed in about 30 of the ex-posed rocks sampled from both arid and hyperarid sites Nei-ther the prevalence of weathering nor its morphology seemedto differ between sites despite the different climates and un-derlying geology In all cases the weathering type was clas-sified as either tafoni or honeycomb weathering (Goudie etal 1997 Groom et al 2015 Fig 1a) furthermore weath-ering was coupled with the presence of sub-aerial biofilmwhich was burrowed underneath the surface and protected bysedimentary deposits (Fig 1b) The weathering and presenceof the crusts were restricted to the atmospherically exposedparts of the rock The presence of an identical weatheringmorphology and the prevalence of this morphology in dif-ferent climates and lithologies challenges the current modelwhich assumes that surface permeability moisture and thepresence of salts are primary factors that control weatheringrates (Goudie et al 2002 Smith 1988 Smith et al 2005)However the lack of a correlation between the magnitudeof tafoni weathering and climate has already been reported(Brandmeier et al 2011)

To study the possible differences between these sites weperformed geological characterization of 10 limestone anddolomite rocks collected from the arid and hyperarid sitesrespectively testing for mineral content porosity permeabil-ity and elasticity As expected our results showed differentlithological parameters between the limestone and dolomiterocks (Table S2) although they displayed similar weather-ing features Moreover petrographic thin-section analysisshowed that crusts had developed to a similar thickness of1ndash6 mm on both rock types irrespective of climatic condi-tions including mean annual precipitation (Fig 1c) How-ever microclimatic conditions such as dew or surface tem-perature may impact local morphologies Furthermore thethin sections showed that the crusts are composed of massesof micritic to microsparitic minerals that form a laminatedstructure (Fig 1c) Such laminated structures indicate thatthe crusts are stage four terrestrial calcretes and dolocretessuggesting a mature crust phase (Alonso-Zarza and Wright2010) The calcretes and dolocretes identified on the rocksrsquosurface reject the previously suggested impact of mineralizednetworks or case hardening (McBride and Picard 2004) Infact the detection of mature calcretes could serve as an indi-cation of atmospheric exposure but has also been suggestedto result from biogenic activity (Alonso-Zarza and Wright2010 Goudie 1996)



Figure 1 (a) Comparable weathering features in the exposed lime-stone and dolomite rocks at both sites as noted in field outcrops(hammer for scale 30 cm long) (b) Visual presence of a rock crustwith similar thickness (3ndash6 mm) in both rock types The crustrsquosmineralogical composition matched that of the host rock (c) Thin-section analysis of the rocks showing lamination structure in theBRCs Dashed lines indicate the interface between BRC host rocksBRCrsquos mineralogy includes micritic to microsparitic dolomitic orcalcitic crystals

32 Composition and chemical characteristics of therock crusts

To test our hypothesis that the crusts are biogenic and in-volved in rock weathering processes we characterized theirorigin and nature An XRD analysis of the crust layers andbedrock at the locations showed that the crusts are composedof a similar mineralogy to their respective host rocks indi-cating that local weathering rather than dust deposition isthe source of crust generation (Table S2)

The biogenic nature of the crusts was confirmed using across-section analysis of the stable carbon and oxygen iso-tope ratios in the crust and host rock (Fig 2a) For bothlimestone and dolomite values of δ13C increased betweenthe crust and the host rock layers and ranged from minus41 permilin the calcrete to minus09 permil in the limestone bed and from02 permil in the dolocrete to 20 permil in the dolomite bed Suchvalues are typical indicators of carbon isotope exchange ofprimary marine CaCO3 (abundant in the bedrock) with CO2released by microbial respiration (ie of carbon originat-ing from photosynthesis) with the subsequent precipitationof pedogenic calcrete (Brlek and Glumac 2014 Mora etal 1991) Analysing δ13C along with δ18O compositionsof pedogenic carbonates are a useful way of reconstructing

Figure 2 (a) Carbon and oxygen isotope ratio depth profiles of thelimestone (top) and dolomite (bottom) BRCs compared with to theirhost rocks The blue horizontal line indicates the border betweenBRC and host rock (b) Fourier transform infrared (FTIR) analysisof limestone (top) and dolomite (bottom) BRCs indicating the pres-ence of extracellular polymeric substance (EPS) molecules throughthe distinctive peak ranging between 1020 and 1040 cmminus1 whichwas absent from the host rocks

palaeo-vegetation (eg the C3C4 plant ratio (Ehleringer etal 1997 Mora et al 1991) Our δ13C results concur withδ13C values collated from speleothems (secondary mineraldeposits formed in caves) collected in the central and south-ern Negev Desert (Vaks et al 2010) which were also datedto end of the Pliocene (the past 25 million years) The lowratio detected here (Fig 2a) and by Vaks et al (2010) suggestthat the Negev region has only been able to support limitedvegetation for at least 25 Gyr if this is the case the role ofthe crust in shaping the morphology of the rock surfaces hasbeen considerable These results support the hypothesis thatcalcretes and dolocretes are of biogenic origin therefore thecrust can be referred to as BRC Moreover they indicate asimilar developmental trajectory for both the calcretes anddolocretes which is independent of aridity or lithological pa-rameters

In contrast the trend in values of δ18O differed betweenrock types In the limestone rocks the ratio ranged fromminus30 permil to minus68 permil between the BRC and the host rockwhereas in the dolomite it was higher ranging from minus54 permilto minus06 permil The decrease in δ18O in the host limestone rockcould be explained by meteoric water substitution (Sandler2006) In contrast the more negative δ18O values in thedolocrete compared with the host dolomite are attributed toisotopic differentiation of meteoric water due to condensa-tion (Rayleigh distillation) and could result from the largedistance from the Mediterranean Sea (which is the primarysource of rainfall in the area) compared with closer lime-stone rocks In speleothems similar patterns of δ18O valueshave been reported in the central and southern Negev Desert(Vaks et al 2010) The results suggest that the calcrete and



dolocrete studied here have been experiencing arid to hyper-arid climates since the Pleistocene alluding to the possiblesource of rain A similar study conducted in the Thar Desertin India also inferred sedimentary rock stable isotope patternsto paleoclimate (Andrews et al 1998)

To study the potential role of BRCs in the weathering pro-cess their composition was characterized using FTIR as waspreviously reported (Sheng et al 2010) We focused on thefunctional groups and elemental compositions in the EPSs ormicrobial aggregates and found a distinct peak in the BRClayers ranging between 1020 and 1040 cmminus1 in both lime-stone and dolomite rocks that was absent in the host rocks(Fig 2b) This peak is indicative of the presence of EPSs ofbacterial origin (Shirshova et al 2006) pointing to the sig-nificant components of asymmetric and symmetric stretch-ing of POminus2 and P(OH)2 in phosphate as well as vibrationsof CndashOH and CndashC bonds which are found in polysaccha-rides and alcohols (Jiang et al 2004) These results providestrong support for the biogenic nature of the crust as EPSis a common feature of many if not most biofilms (Drews etal 2006) The detected EPS could serve several functions inthe BRC such as a dust-particle trap to collect the dust andits nutrients a binding agent to individual members of thebiofilm (Davey and Orsquotoole 2000) or a protective agent thatdecreases evaporation retains moisture and shields from ra-diation (Or et al 2007 Roberson and Firestone 1992)

Based on these findings we hypothesized that a BRCcould in fact act as a mitigator during the weathering processby clogging the pores on the surface of the rock and therebyminimizing capillary rise Consequently crystallization ofdissolved salts considered to be the primary mechanism forrock weathering is mitigated To test this hypothesis we per-formed a drying experiment to estimate water loss from therock surfaces covered with BRCs The results suggest that (inboth limestone and dolomite rocks) water moves through therock and is lost to evaporation two or three times faster in theabsence of BRC than when it is present (Fig 3) Consideringthat salt transport due to hydraulic movement is a dominantweathering mechanism (Huinink et al 2004) reduced evap-oration due to BRC coverage will also inevitably lead to de-creased weathering rates Moreover our results differ fromthose performed on temperate sandy stones that showed nosignificant effect of BRC on decreasing water transport rates(Slaviacutek et al 2017)

33 The microbial composition and origin of the BRCs

To elucidate the identity of the bacterial communities on thelimestone and dolomite BRCs we applied multiplexed bar-coded amplicon sequencing of the small subunit RNA gene(SSU rRNA) In addition we compared the BRC communi-ties to those in the samples of the surrounding soil and settleddust in order to deduce the origin of the rock biofilm As ex-pected we found simple BRC communities (ie low-richnessand low-diversity) The communities of the BRC showed an

Figure 3 Desiccation of rock cores in the presence and absence ofBRCs as a function of time following full hydration The curvesindicate a second-degree polynomial line fitting (all fitted curveswere significantly different from one another in ANOVA tests withP values lt 001 and R2valuesgt095)

average of 182 observed phylotypes 354 predicted phylo-types and a Shannon H value of 38 (Fig 4a Table S3)for arid limestone and 129 observed phylotypes 315 pre-dicted phylotypes and a Shannon H value of 33 for hyper-arid dolomite with no significant difference between the rocktypes

The surrounding soil was significantly richer and more di-verse (Plt005) at the arid site (416 and 746 observed andpredicted OTUs respectively and the Shannon H value was56 on average) and equally rich but slightly more diverseat the hyperarid site (221 and 466 observed and predictedOTUs respectively and the ShannonH value was 38 on av-erage) The diversity of the dust samples was as poor as thatof the BRCs (169 and 107 observed and predicted OTUsrespectively and the Shannon H values were 30 and 15on average) and did not differ between sites (Fig 4a Ta-ble S3) The number of observed OTUs in the soil and theirdiversity scores were somewhat lower in this study comparedwith other reports from similar environments (Barberaacuten etal 2014 Lang-Yona et al 2018 Štoviacutecek et al 2017)which may be due to sequencing technologies and depthThe lower richness and diversity in hyperarid vs arid sam-ples and the BRC and dust vs soil samples is expected and iscomparable with trends reported in other studies (Angel andConrad 2013 Barberaacuten et al 2014 Lang-Yona et al 2018)β-diversity analysis using variance partitioning showed

statistically significant differences between samples at theOTU level based on climate sample type (ie rock soilor dust) and to a small extent also on their interac-tion Because previous field studies in dryland regions haveshowed that soil microbial communities typically exhibitrapid changes after hydration events but little to no seasonalchanges (eg Angel and Conrad 2013 Barnard et al 2013



Figure 4 Microbial community features of the BRCs the surrounding soils and settled dust at the two study sites (a) Comparison ofthe richness in the form of the observed number of OTUs (S obs) and the predicted number of OTUs (the ACE index) in addition to acomparison of α-diversity (Shannon diversity index ndash H) between the different sample types Identical lower-case letters indicate no statisticaldifference between groups in a Tukey HSD test (b) Clustering of sample types using a PCoA ordination based on the HornndashMorisita distancematrix Identical lower-case letters indicate no statistical difference between groups in a pairwise PERMANOVA test Ellipses denote 95 confidence intervals around the arid and hyperarid samples assuming multivariate normal distribution (c) Composition of bacterial phyla inthe different sample types (see Table S5 for results of statistical tests in the relative abundance of different phyla between sample types) (d)Relative contribution of each bacterial OTU to the community composition of each sample type top ndash arid site bottom ndash hyperarid site (seeFig 5 for statistical detection of preferentially abundant OTUs between each sample-type pair)

Štoviacutecek et al 2017) therefore we disregarded the sea-sonal aspect in our analysis These variables were found tosignificantly contribute to the differences in bacterial com-munities accounting for 22 40 and 38 of the to-tal variance respectively (Fig 4b Table S4) Pairwise com-parisons further showed that the two BRCs significantly dif-fered from one another (Plt001) and also from their sur-rounding soil and dust samples (Plt005 in all cases Ta-ble S4) The bacterial community in the samples was typicalfor dryland areas mostly dominated by members of the fol-lowing phyla Proteobacteria Actinobacteria Deinococcus-Thermus Chloroflexi Bacteroidetes Cyanobacteria Aci-dobacteria Firmicutes and Gemmatimonadetes (Fig 4c Ta-ble S5) Similar communities have repeatedly been reportedfor arid and hyperarid soils and rocks (Angel and Conrad2013 Barberaacuten et al 2014 Lang-Yona et al 2018) Whilecyanobacteria are typically the main primary producers in thesoil and rock communities (Weber et al 2016) recent stud-ies have shown that other autotrophs may also significantlycontribute to the energy balance of these biofilms (Ji et al2017)

The BRCs of the two rock types differed in the relativeabundance and composition of the major phyla Most no-tably Proteobacteria were significantly more dominant in thehyperarid compared with the arid samples (P = 002) com-prising 21 and 44 of the community on average in thelimestone and dolomite BRC respectively In contrast theActinobacteria showed an opposite trend (P = 003) com-prising 42 and 21 of the community on average in thelimestone and dolomite BRC respectively The two BRCsalso differed in their composition of Firmicutes Gemmati-monadetes and Chloroflexi (Plt003 Fig 4c Table S5)

The soil samples generally showed similar trends on thegross taxonomic level as their respective BRC samplesWhile none of the phyla differed significantly between thehyperarid BRC and the soil the Deinococcus-Thermus Aci-dobacteria Firmicutes and Gemmatimonadetes phyla dif-fered significantly between limestone BRC and the surround-ing arid soil (Plt004 Table S5) Lastly the arid and hyper-arid dust samples were dominated by members of the Pro-teobacteria with other phyla comprising only a minor frac-tion of the community (with the notable exception of Bac-



teroidetes which dominated one of the dust samples) How-ever these differences were not significant probably due tothe small sample size (Table S5)

Despite the general similarities in the community compo-sition between samples on the phylum level many of theOTUs found in each sample were unique to the BRC soilor dust as evident from the ternary diagrams (Fig 4d) Di-rect analysis of the differences in the OTUs detected 130(10 ) differentially abundant OTUs in the dolomite BRCand 74 (6 ) differentially abundant OTUs in the lime-stone BRC (Fig 5) Similarly several differentially abundantOTUs were also detected when comparing the BRCs withtheir respective soil and dust samples However these dif-ferentially abundant OTUs were fewer probably due to thesmall dust sample size (Fig 5)

Other BRC bacterial communities have previously beendescribed (eg rock varnish quartz pavement) (Kuhlman etal 2006 Lang-Yona et al 2018 Wong et al 2010a b) buttheir origins and roles in geomorphological processes werenot considered Our results suggest that despite the similarityin the morphology and magnitude of rock weathering fea-tures in the arid limestone and hyperarid dolomites the twoBRCs we studied harboured distinct microbial communitiesdiffering in over 16 of the OTUs in addition to their com-position at the phylum level Moreover despite the spatialproximity and continuous interaction between the limestoneand dolomite surfaces and their respective surrounding soilsand dust particles the bacterial communities of the BRCswere distinct The abilities of bacteria to disperse settle andpersist in a given location could be an important factor re-sulting in the biogeographic patterns observed here The dif-ference between arid and hyperarid soil communities couldresult from the local contribution of aeolian material whichmight affect the loess soil diversity (Crouvi et al 2008)Alternatively the hyperarid site experiences slow pedolog-ical processes whereas the arid soil formation was enhanced(Amit et al 2011) resulting in disparate bacterial communi-ties The three matrices (BRC soil and settled dust) studiedhere sparsely shared their bacterial communities In particu-lar the BRC communities had little in common with eitherthe soil or the dust communities (Fig 4) This demonstratesthe ecological filtering effect of the rock surfaces which im-poses unique abiotic challenges on the microbes living onit (Horner-Devine and Bohannan 2006) This also suggeststhat the BRCs cannot be regarded as passive deposits of mi-crobial cells originating from the surrounding soil or dustbut should rather be seen as a specific subset of adapted mi-crobes that can persist and form a biofilm under these uniqueconditions

34 The role of BRC in arid rock weathering ndashsynthesis

Honeycomb weathering patterns are prevalent worldwideand are found in both humid and dry ecosystems Accord-

ing to contemporary models this form of weathering is theresult of the transport of dissolved salts through the rock andtheir eventual crystallization in surface pores which leads tofractures and eventual flaking of rock material (Rodriguez-Navarro et al 1999) In this study we found that the weath-ering patterns and magnitude are similar on rocks from botharid and hyperarid sites despite the differences in precipi-tation and lithologies In arid and hyperarid regions BRCshave been shown to form once the rock is exposed to the at-mosphere (Pointing and Belnap 2012) A developed crust ofbiological origin was microscopically and isotopically appar-ent on all weathered rocks and was shown to be supported byEPSs (Fig 2) Similar to weathering magnitude the BRCsshowed no observable differences in form or depth despitethe different aridity and lithology Both BRCs comprisedbacterial taxa that are typical for xeric environments (Point-ing and Belnap 2012) and included many heterotrophs butalso dominant phototrophs or otherwise autotrophic mem-bers (Fig 4) The two BRCs did differ in their bacterial com-munities at the OTU and higher taxonomical levels demon-strating a discrepancy between community structure and eco-logical function (forming BRCs) The BRC communitiesalso differed from their surrounding soil and dust indica-tive of the specialization of the colonizing taxa to rock en-vironments In the absence of mineralized networks or casehardening (ie the addition of cementing agent to rock ma-trix material) we conclude that calcrete and dolocrete wereformed through the colonization of microorganisms and thesecretion of EPSs which served as a thin biofilm (Brantleyet al 2011 Weber et al 2016)

Our results further suggest that this biogenic layer miti-gates evaporation and reduces water transport thereby allevi-ating salt crystallization pressure in the rock pours (Scherer2004) Crystallization of calcium sulphate and sodium chlo-ride solutions which are abundant in these soils was shownto build pressure within pores and stress rocks (Scherer2004 Sperling and Cooke 1985) This process is enhancedunder low relative humidity and rapid evaporation and com-promises the durability of the rocks (Rodriguez-Navarro etal 2003 Rodriguez-Navarro and Doehne 1999) Our resultssuggest that the presence of a BRC decreases evaporationrates (Fig 3) and thus attenuates the crystallization pressureand reduces damage to the rocks Moreover the BRC mayalso stabilize the rock following exfoliation preserving theweathered structure

Arid weathering features which lead to debris formationresult from a dynamic balance between the erosive salt forcesand the mitigating effects of the BRC The role of microbialbiofilms in the protection of surfaces from mineral weather-ing was extensively studied for biomineralization and sed-imentation processes (Adams et al 1992 Dupraz et al2009) However the role of BRC in weathering processesunder atmospheric conditions in the desert has not been con-sidered before We propose that microbial colonization ofmineral surfaces protects the rocks from weathering by miti-



Figure 5 Detection of differentially abundant OTUs between each sample-type pair using ALDEx2 Each circle denotes a single OTU andits size is its average relative abundance across all samples The x axis shows the classification of each OTU whereas the y axis showsthe effect size in terms of the log2 fold difference in relative abundance between each sample-type pair Red circles are OTUs that showsignificant differential abundance at the Plt005 level Numbers next to the arrows indicate the number of significant differentially abundantOTUs that are either more abundant (up arrow) or less abundant (down arrow) in the sample-type appearing in the column compared withthe sample-type appearing in the row Numbers in brackets indicate the total number of OTUs in each pair-wise comparison

gating salt crystallization and stabilizing the weathered frontRock weathering processes are typically believed to be con-trolled at different scales ranging from the climatic scaledown to local conditions at the site and eventually the micro-scale (Smith 2009 Sperling and Cooke 1985 Viles 2001)The results presented here suggest that in arid environmentsmicro-scale conditions determine the magnitude of weather-ing that shape the landscape

Data availability The raw sequencing data were depositedinto the EMBL-ENA SRA database (httpswwwebiacukenadataviewPRJNA381483 Ben Gurion University of theNegev 2019) All data processing and plotting scripts aswell as processed data tables are available through GitHub(httpsdoiorg105281zenodo1478476 Angel 2018)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-1133-2019-supplement

Author contributions NW HG OG and RA conceptualized thestudy NW performed the field and lab work NW and RA anal-ysed the data NW HG OG and RA wrote the paper OG and RAcontributed equally to this study

Competing interests The authors declare that they have no conflictof interest

Acknowledgements RA was supported by the Czech Ministry ofEducation Youth and Sport (MEYS project nos LM2015075 andEF16_0130001782 ndash SoWa Ecosystem Research)

Review statement This paper was edited by Denise Akob and re-viewed by two anonymous referees

References

Adams J B Palmer F and Staley J T Rock weath-ering in deserts Mobilization and concentration of fer-ric iron by microorganisms Geomicrobiol J 10 99ndash114httpsdoiorg10108001490459209377910 1992

Alonso-Zarza A M and Wright V P Chapter 5 Calcretes in De-velopments in Sedimentology vol 61 225ndash267 Elsevier 2010

Alonso-Zarza A M Bustamante L Huerta P Rodriacuteguez-Berriguete Aacute and Huertas M J Chabazite and dolomite for-mation in a dolocrete profile An example of a complex alkalineparagenesis in Lanzarote Canary Islands Sediment Geol 3371ndash11 httpsdoiorg101016jsedgeo201602018 2016

Amit R Harrison J B J Enzel Y and Porat N Soils as atool for estimating ages of Quaternary fault scarps in a hyper-



arid environment ndash the southern Arava valley the Dead SeaRift Israel CATENA 28 21ndash45 httpsdoiorg101016S0341-8162(96)00028-8 1996

Amit R Enzel Y Grodek T Crouvi O Porat N and AyalonA The role of rare rainstorms in the formation of calcic soilhorizons on alluvial surfaces in extreme deserts Quat Res 74177ndash187 httpsdoiorg101016jyqres201006001 2010

Amit R Enzel Y Crouvi O Simhai O Matmon A Po-rat N McDonald E and Gillespie A R The role of theNile in initiating a massive dust influx to the Negev late inthe middle Pleistocene Geol Soc Am Bull 123 873ndash889httpsdoiorg101130B302411 2011

Andrews J E Singhvi A K Kailath A J Kuhn R DennisP F Tandon S K and Dhir R P Do Stable Isotope Datafrom Calcrete Record Late Pleistocene Monsoonal Climate Vari-ation in the Thar Desert of India Quat Res 50 240ndash251httpsdoiorg101006qres19982002 1998

Angel R Analysis scripts data files and plots used inthe manuscript roey-angelRock_weathering HTMLhttpsdoiorg105281zenodo1478476 (last access March2019) 2018

Angel R and Conrad R Elucidating the microbial re-suscitation cascade in biological soil crusts following asimulated rain event Environ Microbiol 15 2799ndash2815httpsdoiorg1011111462-292012140 2013

Angel R Claus P and Conrad R Methanogenic archaeaare globally ubiquitous in aerated soils and become ac-tive under wet anoxic conditions ISME J 6 847ndash862httpsdoiorg101038ismej2011141 2012

Azua-Bustos A Urrejola C and Vicuntildea R Life at the dry edgeMicroorganisms of the Atacama Desert FEBS Lett 586 2939ndash2945 httpsdoiorg101016jfebslet201207025 2012

Barberaacuten A Henley J Fierer N and Casamayor E O Struc-ture inter-annual recurrence and global-scale connectivity ofairborne microbial communities Sci Total Environ 487 187ndash195 httpsdoiorg101016jscitotenv201404030 2014

Barnard R L Osborne C A and Firestone M K Re-sponses of soil bacterial and fungal communities to ex-treme desiccation and rewetting ISME J 7 2229ndash2241httpsdoiorg101038ismej2013104 2013

Ben Gurion University of the Negev Arid and hyperarid rock crustsoil crust and dust microboime available at httpswwwebiacukenadataviewPRJNA381483 last access March 2019

Brandmeier M Kuhlemann J Krumrei I Kappler A and Ku-bik P W New challenges for tafoni research A new approach tounderstand processes and weathering rates Earth Surf ProcessLandf 36 839ndash852 httpsdoiorg101002esp2112 2011

Brantley S L Megonigal J P Scatena F N Balogh-BrunstadZ Barnes R T Bruns M A Cappellen P V DontsovaK Hartnett H E Hartshorn A S Heimsath A Hern-don E Jin L Keller C K Leake J R Mcdowell W HMeinzer F C Mozdzer T J Petsch S Pett-Ridge J Pregit-zer K S Raymond P A Riebe C S Shumaker K Sutton-Grier A Walter R and Yoo K Twelve testable hypothe-ses on the geobiology of weathering Geobiology 9 140ndash165httpsdoiorg101111j1472-4669201000264x 2011

Brlek M and Glumac B Stable isotopic (δ13C and δ18O)signatures of biogenic calcretes marking discontinuity sur-faces a case study from Upper Cretaceous carbonates of cen-

tral Dalmatia and eastern Istria Croatia Facies 60 773ndash788httpsdoiorg101007s10347-014-0403-7 2014

Bruins H J Ancient desert agriculture in the Negevand climate-zone boundary changes during averagewet and drought years J Arid Environ 86 28ndash42httpsdoiorg101016jjaridenv201201015 2012

Bruthans J Soukup J Vaculikova J Filippi MSchweigstillova J Mayo A L Masin D KletetschkaG and Rihosek J Sandstone landforms shaped by negativefeedback between stress and erosion Nat Geosci 7 597ndash601httpsdoiorg101038ngeo2209 2014

Bruthans J Filippi M Slaviacutek M and Svobodovaacute EOrigin of honeycombs Testing the hydraulic and casehardening hypotheses Geomorphology 303 68ndash83httpsdoiorg101016jgeomorph201711013 2018

Buumldel B Weber B Kuumlhl M Pfanz H Suumlltemeyer Dand Wessels D Reshaping of sandstone surfaces by cryp-toendolithic cyanobacteria bioalkalization causes chemicalweathering in arid landscapes Geobiology 2 261ndash268httpsdoiorg101111j1472-4677200400040x 2004

Chen L Reeve J Zhang L Huang S Wang X and Chen JGMPR A robust normalization method for zero-inflated countdata with application to microbiome sequencing data PeerJ 6e4600 httpsdoiorg107717peerj4600 2018

Cooke R U Laboratory simulation of salt weathering pro-cesses in arid environments Earth Surf Process 4 347ndash359httpsdoiorg101002esp3290040405 1979

Crouvi O Amit R Enzel Y Porat N and Sandler A Sanddunes as a major proximal dust source for late Pleistoceneloess in the Negev Desert Israel Quat Res 70 275ndash282httpsdoiorg101016jyqres200804011 2008

Davey M E and Orsquotoole G A Microbial biofilms from ecologyto molecular genetics Microbiol Mol Biol Rev 64 847ndash867httpsdoiorg101128MMBR644847-8672000 2000

Drews A Lee C-H and Kraume M Membrane fouling ndasha review on the role of EPS Desalination 200 186ndash188httpsdoiorg101016jdesal200603290 2006

Dupraz C Reid R P Braissant O Decho A W NormanR S and Visscher P T Processes of carbonate precipita-tion in modern microbial mats Earth-Sci Rev 96 141ndash162httpsdoiorg101016jearscirev200810005 2009

Edgar R C UPARSE highly accurate OTU sequencesfrom microbial amplicon reads Nat Methods 10 996ndash998httpsdoiorg101038nmeth2604 2013

Ehleringer J R Cerling T E and Helliker B R C4 photosyn-thesis atmospheric CO2 and climate Oecologia 112 285ndash299httpsdoiorg101007s004420050311 1997

Fernandes A D Reid J N Macklaim J M McMurroughT A Edgell D R and Gloor G B Unifying the analy-sis of high-throughput sequencing datasets characterizing RNA-seq 16S rRNA gene sequencing and selective growth exper-iments by compositional data analysis Microbiome 2 1ndash15httpsdoiorg1011862049-2618-2-15 2014

Ferrando D Toubiana D Kandiyote N S Nguyen T H Ne-jidat A and Herzberg M Ambivalent role of calcium in theviscoelastic properties of extracellular polymeric substances andthe consequent fouling of reverse osmosis membranes Desali-nation 429 12ndash19 httpsdoiorg101016jdesal2017120062018



Ferreira J A and Zwinderman A H On the BenjaminindashHochberg method Ann Stat 34 1827ndash1849httpsdoiorg101214009053606000000425 2006

Garcia-Pichel F Plausible mechanisms for the boring on carbon-ates by microbial phototrophs Sediment Geol 185 205ndash213httpsdoiorg101016jsedgeo200512013 2006

Goossens D and Rajot J L Techniques to measure the dry aeo-lian deposition of dust in arid and semi-arid landscapes a com-parative study in West Niger Earth Surf Process Landf 33178ndash195 httpsdoiorg101002esp1533 2008

Gorbushina A A Life on the rocks Environ Microbiol 9 1613ndash1631 httpsdoiorg101111j1462-2920200701301x 2007

Goudie A S Organic agency in calcrete development J AridEnviron 32 103ndash110 httpsdoiorg101006jare199600101996

Goudie A S Quantification of rock control ingeomorphology Earth-Sci Rev 159 374ndash387httpsdoiorg101016jearscirev201606012 2016

Goudie A S Viles H A and Parker A G Monitor-ing of rapid salt weathering in the central Namib Desertusing limestone blocks J Arid Environ 37 581ndash598httpsdoiorg101006jare19970297 1997

Goudie A S Wright E and Viles H A The roles of salt (sodiumnitrate) and fog in weathering a laboratory simulation of condi-tions in the northern Atacama Desert Chile CATENA 48 255ndash266 httpsdoiorg101016S0341-8162(02)00028-0 2002

Groom K M Allen C D Mol L Paradise T R and HallK Defining tafoni Re-examining terminological ambiguity forcavernous rock decay phenomena Prog Phys Geogr 39 775ndash793 httpsdoiorg1011770309133315605037 2015

Hamilton N ggtern An Extension to ldquoggplot2rdquo for the Creationof Ternary Diagrams available at httpsCRANR-projectorgpackage=ggtern (last access March 2019) 2017

Horner-Devine M C and Bohannan B J M Phyloge-netic clustering and overdispersion in bacterial communi-ties Ecology 87 S100ndashS108 httpsdoiorg1018900012-9658(2006)87[100PCAOIB]20CO2 2006

Huinink H P Pel L and Kopinga K Simulating thegrowth of tafoni Earth Surf Process Landf 29 1225ndash1233httpsdoiorg101002esp1087 2004

Ji M Greening C Vanwonterghem I Carere C R BayS K Steen J A Montgomery K Lines T Beardall JDorst J van Snape I Stott M B Hugenholtz P and Fer-rari B C Atmospheric trace gases support primary produc-tion in Antarctic desert surface soil Nature 552 400ndash403httpsdoiorg101038nature25014 2017

Jiang W Saxena A Song B Ward B B Beveridge T J andMyneni S C B Elucidation of Functional Groups on Gram-Positive and Gram-Negative Bacterial Surfaces Using InfraredSpectroscopy Langmuir 20 11433ndash11442 2004

Klindworth A Pruesse E Schweer T Peplies J Quast CHorn M and Glockner F O Evaluation of general 16S ribo-somal RNA gene PCR primers for classical and next-generationsequencing-based diversity studies Nucleic Acids Res 41 e1httpsdoiorg101093nargks808 2012

Kuhlman K R Fusco W G La Duc M T Allenbach L BBall C L Kuhlman G M Anderson R C Erickson I KStuecker T Benardini J Strap J L and Crawford R LDiversity of Microorganisms within Rock Varnish in the Whip-

ple Mountains California Appl Environ Microbiol 72 1708ndash1715 httpsdoiorg101128AEM7221708-17152006 2006

Lang-Yona N Maier S Macholdt D S Muumlller-Germann IYordanova P Rodriguez-Caballero E Jochum K P Al-AmriA Andreae M O Froumlhlich-Nowoisky J and Weber B In-sights into microbial involvement in desert varnish formation re-trieved from metagenomic analysis Environ Microbiol Rep10 264ndash271 httpsdoiorg1011111758-222912634 2018

Lebre P H Maayer P D and Cowan D A Xerotolerant bacte-ria surviving through a dry spell Nat Rev Microbiol 15 285ndash296 httpsdoiorg101038nrmicro201716 2017

Makhalanyane T P Valverde A Birkeland N-K Cary S CMarla Tuffin I and Cowan D A Evidence for successional de-velopment in Antarctic hypolithic bacterial communities ISMEJ 7 2080ndash2090 httpsdoiorg101038ismej201394 2013

Makhalanyane T P Valverde A Gunnigle E Frossard ARamond J-B and Cowan D A Microbial ecology of hotdesert edaphic systems FEMS Microbiol Rev 39 203ndash221httpsdoiorg101093femsrefuu011 2015

Mangiafico S rcompanion Functions to Support ExtensionEducation Program Evaluation available at httpsCRANR-projectorgpackage=rcompanion (last access March 2019)2018

McArdle B H and Anderson M J Fitting multivariate modelsto community data a comment on distance-based redundancyanalysis Ecology 82 290ndash297 httpsdoiorg1018900012-9658(2001)082[0290FMMTCD]20CO2 2001

McBride E F and Picard M D Origin of honeycombs and re-lated weathering forms in Oligocene Macigno Sandstone Tus-can coast near Livorno Italy Earth Surf Process Landf 29713ndash735 httpsdoiorg101002esp1065 2004

McIlroy de la Rosa J P Warke P A and Smith BJ The effects of lichen cover upon the rate of solu-tional weathering of limestone Geomorphology 220 81ndash92httpsdoiorg101016jgeomorph201405030 2014

McMurdie P J and Holmes S phyloseq An R Pack-age for Reproducible Interactive Analysis and Graphicsof Microbiome Census Data PLOS ONE 8 e61217httpsdoiorg101371journalpone0061217 2013

Mora C I Driese S G and Seager P G Carbondioxide in the Paleozoic atmosphere Evidence fromcarbon-isotope compositions of pedogenic carbon-ate Geology 19 1017 httpsdoiorg1011300091-7613(1991)019lt1017CDITPAgt23CO2 1991

Mustoe G E Cavernous weathering in the Capitol ReefDesert Utah Earth Surf Process Landf 8 517ndash526httpsdoiorg101002esp3290080603 1983

Mustoe G E Biogenic origin of coastal honeycombweathering Earth Surf Process Landf 35 424ndash434httpsdoiorg101002esp1931 2010

OrsquoHara R B Species richness estimators how many speciescan dance on the head of a pin J Anim Ecol 74 375ndash386httpsdoiorg101111j1365-2656200500940x 2005

Oksanen J Blanchet F G Friendly M Kindt R Legendre PMcGlinn D Minchin P R OrsquoHara R B Simpson G LSolymos P Stevens M H H Szoecs E and Wagner H ve-gan Community Ecology Package available at httpsCRANR-projectorgpackage=vegan (last access March 2019) 2018



Or D Smets B F Wraith J M Dechesne A and Friedman SP Physical constraints affecting bacterial habitats and activityin unsaturated porous media ndash a review Adv Water Resour 301505ndash1527 httpsdoiorg101016jadvwatres2006050252007

Pointing S B and Belnap J Microbial colonization and con-trols in dryland systems Nat Rev Microbiol 10 551ndash562httpsdoiorg101038nrmicro2831 2012

Quast C Pruesse E Yilmaz P Gerken J SchweerT Yarza P and Glockner F O The SILVA riboso-mal RNA gene database project improved data process-ing and web-based tools Nucl Acid Res 41 D590ndashD596httpsdoiorg101093nargks1219 2013

R Core Team R A Language and Environment for Statistical Com-puting R Foundation for Statistical Computing Vienna Aus-tria available at httpswwwR-projectorg (last access March2019) 2016

Roberson E B and Firestone M K Relationship between desic-cation and exopolysaccharide production in a soil Pseudomonassp Appl Environ Microbiol 58 1284ndash1291 1992

Rodriguez-Navarro C and Doehne E Salt weath-ering influence of evaporation rate supersatura-tion and crystallization pattern Earth Surf ProcessLandf 24 191ndash209 httpsdoiorg101002(SICI)1096-9837(199903)243lt191AID-ESP942gt30CO2-G 1999

Rodriguez-Navarro C Doehne E and Sebastian E Originsof honeycomb weathering The role of salts and wind GeolSoc Am Bull 111 1250ndash1255 httpsdoiorg1011300016-7606(1999)111lt1250OOHWTRgt23CO2 1999

Rodriguez-Navarro C Rodriguez-Gallego M Chekroun KB and Gonzalez-Muntildeoz M T Conservation of Orna-mental Stone by Myxococcus xanthus-Induced CarbonateBiomineralization Appl Environ Microbiol 69 2182ndash2193httpsdoiorg101128AEM6942182-21932003 2003

Sandler A Meunier A and Velde B Mineralogi-cal and chemical variability of mountain redbrownMediterranean soils Geoderma 239ndash240 156ndash167httpsdoiorg101016jgeoderma201410008 2015

Scherer G W Crystallization in pores Cement and Con-crete Research 29 1347ndash1358 httpsdoiorg101016S0008-8846(99)00002-2 1999

Scherer G W Stress from crystallizationof salt Cem Concr Res 34 1613ndash1624httpsdoiorg101016jcemconres200312034 2004

Schloss P D Westcott S L Ryabin T Hall J R Hart-mann M Hollister E B Lesniewski R A Oakley BB Parks D H Robinson C J Sahl J W Stres BThallinger G G Van Horn D J and Weber C F Introduc-ing mothur Open-Source Platform-Independent Community-Supported Software for Describing and Comparing Micro-bial Communities Appl Environ Microbiol 75 7537ndash7541httpsdoiorg101128AEM01541-09 2009

Sheng G-P Yu H-Q and Li X-Y Extracellular polymeric sub-stances (EPS) of microbial aggregates in biological wastewa-ter treatment systems A review Biotechnol Adv 28 882ndash894httpsdoiorg101016jbiotechadv201008001 2010

Shirshova L T Ghabbour E A and Davies G Spectroscopiccharacterization of humic acid fractions isolated from soil us-

ing different extraction procedures Geoderma 133 204ndash216httpsdoiorg101016jgeoderma200507007 2006

Shtober-Zisu N Amasha H and Frumkin A Inland notcheslithological characteristics and climatic implications of subaerialcavernous landforms in Israel Inland Notches Lithological char-acteristics and climatic implications Earth Surf Process Landf42 1820ndash1832 httpsdoiorg101002esp4135 2017

Slaviacutek M Bruthans J Filippi M Schweigstillovaacute JFalteisek L and Rihošek J Biologically-initiated rockcrust on sandstone Mechanical and hydraulic propertiesand resistance to erosion Geomorphology 278 298ndash313httpsdoiorg101016jgeomorph201609040 2017

Smith B J Weathering of superficial limestone debris ina hot desert environment Geomorphology 1 355ndash367httpsdoiorg1010160169-555X(88)90007-4 1988

Smith B J Weathering Processes and Forms in Geomorphologyof Desert Environments edited by Parsons A J and AbrahamsA D 69ndash100 Springer Netherlands Dordrecht 2009

Smith B J Warke P A McGreevy J P and KaneH L Salt-weathering simulations under hot desertconditions agents of enlightenment or perpetuatorsof preconceptions Geomorphology 67 211ndash227httpsdoiorg101016jgeomorph200403015 2005

Sperling C H B and Cooke R U Laboratory simulation ofrock weathering by salt crystallization and hydration processesin hot arid environments Earth Surf Process Landf 10 541ndash555 httpsdoiorg101002esp3290100603 1985

Štoviacutecek A Kim M Or D and Gillor O Microbial communityresponse to hydration-desiccation cycles in desert soil Sci Rep7 45735 httpsdoiorg101038srep45735 2017

Sweeting M M and Lancaster N Solutional and wind erosionforms on limestone in the central Namib Desert Z Geomorphol26 197ndash207 1982

Uemura R Nakamoto M Asami R Mishima S Gibo MMasaka K Jin-Ping C Wu C-C Chang Y-W and ShenC-C Precise oxygen and hydrogen isotope determination innanoliter quantities of speleothem inclusion water by cavity ring-down spectroscopic techniques Geochim Cosmochim Acta172 159ndash176 httpsdoiorg101016jgca201509017 2016

Vaks A Bar-Matthews M Matthews A Ayalon A andFrumkin A Middle-Late Quaternary paleoclimate of north-ern margins of the Saharan-Arabian Desert reconstruction fromspeleothems of Negev Desert Israel Quat Sci Rev 29 2647ndash2662 httpsdoiorg101016jquascirev201006014 2010

Viles H Ecological perspectives on rock surface weatheringTowards a conceptual model Geomorphology 13 21ndash35httpsdoiorg1010160169-555X(95)00024-Y 1995

Viles H Self-organized or disorganized Towards a general expla-nation of cavernous weathering Earth Surf Process Landf 301471ndash1473 httpsdoiorg101002esp1287 2005

Viles H Goudie A Grab S and Lalley J The use ofthe Schmidt Hammer and Equotip for rock hardness as-sessment in geomorphology and heritage science a com-parative analysis Earth Surf Process Landf 36 320ndash333httpsdoiorg101002esp2040 2011

Viles H A Scale issues in weathering studies Geomorphol-ogy 41 63ndash72 httpsdoiorg101016S0169-555X(01)00104-0 2001



Viles H A Understanding Dryland Landscape Dynamics Do Bi-ological Crusts Hold the Key Geogr Compass 2 899ndash919httpsdoiorg101111j1749-8198200800099x 2008

Viles H A Microbial geomorphology A neglected link be-tween life and landscape Geomorphology 157ndash158 6ndash16httpsdoiorg101016jgeomorph201103021 2012

Viles H A and Goudie A S Rapid salt weather-ing in the coastal Namib desert Implications forlandscape development Geomorphology 85 49ndash62httpsdoiorg101016jgeomorph200603025 2007

Wang Q Garrity G M Tiedje J M and Cole J R NaiveBayesian classifier for rapid assignment of rRNA sequencesinto the new bacterial taxonomy Appl Environ Microbiol 735261ndash5267 httpsdoiorg101128AEM00062-07 2007

Warscheid T and Braams J Biodeterioration of stonea review Int Biodeterior Biodegrad 46 343ndash368httpsdoiorg101016S0964-8305(00)00109-8 2000

Weber B Buumldel B and Belnap J Eds Biological Soil CrustsAn Organizing Principle in Drylands 1st ed 2016 editionSpringer New York NY 2016

Wickham H ggplot2 Elegant Graphics for Data AnalysisSpringer-Verlag New York USA available at httpggplot2org(last access March 2019) 2016

Wong F K Y Lau M C Y Lacap D C Aitchison J CCowan D A and Pointing S B Endolithic Microbial Col-onization of Limestone in a High-altitude Arid EnvironmentMicrob Ecol 59 689ndash699 httpsdoiorg101007s00248-009-9607-8 2010a

Wong F K Y Lacap D C Lau M C Y Aitchison J C CowanD A and Pointing S B Hypolithic Microbial Community ofQuartz Pavement in the High-Altitude Tundra of Central TibetMicrob Ecol 60 730ndash739 httpsdoiorg101007s00248-010-9653-2 2010b

Wright D T and Wacey D Sedimentary dolomite a re-ality check Geol Soc Lond Spec Publ 235 65ndash74httpsdoiorg101144GSLSP20042350103 2004


Abstract

Introduction

Materials and methods

Study site

Field sampling

Geological analyses

FTIR and stable isotope analysis

Desiccation experiment

DNA extraction PCR amplification and sequencing

Sequence processing and analysis of bacterial communities

Results and discussion

Field and mineralogical observations

Composition and chemical characteristics of the rock crusts

The microbial composition and origin of the BRCs

The role of BRC in arid rock weathering -- synthesis

Data availability

Supplement

Author contributions

Competing interests

Acknowledgements

Review statement

References


respectively) by displacive and replacive cementation of cal-cium or dolomite onto the rock surface (Wright and Wacey2004 Alonso-Zarza and Wright 2010) Following the ce-mentation processes typical honeycomb features are formedon the exposed parent rock typified by pits separated bythin walls that are coated by the calcrete or dolocrete Re-cent studies suggest that microbial activity also promotes theprocesses of calcrete and dolocrete formation (Alonso-Zarzaet al 2016 Alonso-Zarza and Wright 2010)

The accepted conceptual model for the formation of cav-ernous rock weathering in hot deserts involves the pres-ence of permeable rocks that are subjected to soluble saltsand repeated episodes of dryingndashrewetting cycles (Goudieet al 2002 Smith 1988 Smith et al 2005) The pro-posed mechanisms assume that cavernous weathering re-sults from physico-chemical processes including salt crys-tallization (Cooke 1979 Scherer 2004) incipient fractures(Amit et al 1996) exfoliation (Shtober-Zisu et al 2017)or stress erosion (Bruthans et al 2014 McArdle and An-derson 2001) Recently Bruthans et al (2018) conclusivelydemonstrated that moisture flux followed by salt crystalliza-tion at the boundary layer govern the case hardening modelin temperate climates In addition biological mechanismshave been proposed to promote rock weathering throughmechanisms such as flaking via colony growth (Viles 2012)acidification by bacterial extractions (Garcia-Pichel 2006Warscheid and Braams 2000) or alkalization during photo-synthesis by cyanobacteria (Buumldel et al 2004) In contrastit has been proposed that micro- and macro-organism col-onization can mitigate weathering in temperate coastal re-gions (McIlroy de la Rosa et al 2014 Mustoe 2010) viaencrustation or protection from direct rain impact Howeverit is not clear which of these mechanisms dominates or whatthe relative contribution of chemical vs biological processesto weathering in arid environments is

Microorganisms colonizing rocks form a hardy biofilmknown as the biological rock crust (BRC) which is com-mon in most arid and hyperarid regions worldwide (Gor-bushina 2007 Lebre et al 2017 Pointing and Belnap2012) Epilithic communities colonizing rock surfaces areubiquitous in arid environments whereas hyperarid rockswhich experience increased radiation and desiccation aredominated by endolithic communities that colonize inter-nal rock pores (Makhalanyane et al 2013 Pointing andBelnap 2012 Viles 1995) The BRC communities includecyanobacteria and other phototrophic and heterotrophic bac-teria but very low abundances of archaea fungi or algae(Lang-Yona et al 2018) However the BRC inoculum hasnot been resolved and has been proposed to originate fromsettled dust (Viles 2008) or the surrounding soil (Makha-lanyane et al 2015)

The goal of this study was to illuminate the origin androle of BRCs in cavernous weathering of exposed lime-stone and dolomite rocks in arid and hyperarid regions Wepredicted that the BRC communities on exposed rock sur-

faces would resemble either the ever-present dust or the sur-rounding soil supporting a subset of adapted taxa from bothsources We further hypothesized that the cavernous weath-ering morphologies of exposed rocks result from salt mobi-lization by dew causing crystallization pressure under atmo-spheric conditions The developed rock biofilms clog the sur-face rock pores through secretion of extracellular polymericsubstances (EPSs) lowering evaporation and slowing the saltcrystallization but also stabilizing the exfoliated rocks andpreventing further weathering Thus the presence of a BRCmitigates the geomorphic processes To test our hypotheseswe applied a holistic approach combining field observationsand geological geotechnical and molecular microbiologycharacterization to elucidate the morphology origin and roleof BRCs in arid cavernous weathering

2 Materials and methods

21 Study site

We focused on two sites in the Negev Desert Israel Sede-Boqer ndash an arid site and Uvda Valley ndash a hyperarid site(Fig S1 Table S1 in the Supplement) Both sites have rockyterrains underlaid predominantly by carbonate rock slopesconsisting of limestone dolomite chalk marl clay and chertfrom the Cretaceous period to the Eocene Our analyses com-pared samples from the limey Turonian age Shivta Formationlocated in the arid region with samples from the dolomiticTuronian age Gerofit Formation located in the hyperarid en-vironment The Negev Desert Israel has maintained arid tohyperarid conditions since the Holocene and has an arid-ity index (PPET) of 005ndash0005 (Amit et al 2010 Bru-ins 2012) which is similar to other arid and hyperarid ar-eas worldwide eg the Namib and Atacama deserts (Azua-Bustos et al 2012 Viles and Goudie 2007) The long-termaridity of the Negev Desert makes it a reliable site for testingthe relationships between BRCs and geological substrates

22 Field sampling

A total of 24 rock samples were collected along rockyslopes facing northward during November and December2014 these samples were comprised of 12 limestone sam-ples from the limey Turonian age Shivta Formation at the aridsite (3088 N 3478 E WGS 84 Grid the samples werenamed SB 1ndash12) and 12 dolomite samples from the limeydolomitic Turonian age Gerofit Formation at the hyperaridsite (2994 N 3497 E WGS 84 Grid the samples werenamed UV 1ndash12) Concomitantly six soil samples (ca 500 geach) were collected three samples from the arid site (re-ferred to as SBSoil 1ndash3) and three samples from the hyper-arid (referred to as UVSoil 1ndash12) site Each rock or soil sam-ple was a composite of four sub-samples that were pooledand homogenized in the lab






































































References

















































































































Abstract

Introduction


Study site

Field sampling

Geological analyses










Data availability

Supplement


Competing interests

Acknowledgements

Review statement

References





































































References

















































































































Abstract

Introduction


Study site

Field sampling

Geological analyses










Data availability

Supplement


Competing interests

Acknowledgements

Review statement

References












































































































Abstract

Introduction


Study site

Field sampling

Geological analyses










Data availability

Supplement


Competing interests

Acknowledgements

Review statement

References

the origin and role of biological rock crusts in rocky

Documents