Biological soil crust (BSC) is a complex microbial association composed of microorganisms, soil particles and extracellular polymeric substances. For the arid regions of the planet, it is estimated that biocrust covers up to 45 % of the soil surface, while for the Mediterranean region the percentage of coverage is almost 70 % (Steven et al., 2014). However, while BSCs are by definition found in arid and semiarid areas, this being their key component, they are also known as inhabitants of mesic ecosystems (Bowker, 2010).

Biocrusts are highly complex systems where the abiotic environment interacts with local soil organisms, including cyanobacteria, algae, lichens and mosses, bacteria, fungi, and various types of microfauna, inducing a series of geomorphological features, such as soil stabilization (Maier et al., 2018), increased corrosion resistance (Bu et al., 2015) and formation of a rough dust-trapping texture (Williams et al., 2012). In addition, crusts have an effect on the soil hydrological characteristics by altering water infiltration (Kidron et al., 2020), runoff intensity (Rodriguez-Caballiero et al., 2012) and evaporation rate (Kidron et al., 2020). Finally, crusts alter the soil surface temperature (Couradeau et al., 2016) and soil water content (Bu et al., 2015).

Given the environmental hardness predominating in arid and semiarid environments, the interaction of crusts with the vascular plants is vital for plants' survival (Bainbridge, 2012). Vegetation has an effect on biocrust formation by altering various microclimatic and territorial properties (Zhang et al., 2016). The type of plants, their density and the size of the canopy affect parameters such as the level of radiation reaching the soil, the quality of the litter produced and the soil temperature through the layer formed by the fallen leaves. Further plants interact with the biocrusts through the activity of their root system (Zhang et al., 2016). In crusted areas enhanced porosity and increased high soil humidity are recorded beneath plants (Menon et al., 2011). For its part, biological soil crust increases soil fertility and generally has a positive effect on soil quality, a fact that is reflected in the increase in soil enzyme activity underneath the crust (Liu et al., 2014). Especially in arid areas, where nutrient concentration is low, BSC presence is associated with increased levels of soil organic matter and available nitrogen through the fixation of atmospheric nitrogen (Weber et al. 2015), thus promoting the growth of vascular plants (Havrilla et al., 2019). Apart from increasing nutrients, the organisms of BSCs secrete exopolysaccharides especially at the initial stage of their development (Cania et al., 2020) and contain lipids, which act as energy reserves, contributing further to the development of an important energy source for soil biota (Ngosong et al., 2020). However, although in most cases the effects of biocrusts on soil properties were positive, Papatheodorou et al. (2020) mentioned that the abundance of soil microbes in areas covered by Cladonia rangiformis decreased after a rainy event. Similarly, Delgado-Baquerizo et al. (2015) found contrasting soil microbial abundances under the different biocrust-forming lichens. Specifically, the lichens Fulgensia subbracteata and Diploschistes diacapsis were negatively related to the abundance of soil bacteria.

Considering the importance of the BSCs' interplay with the local vegetation, in this study we focused on the effect of the interactions between the plant Erica spp. (a drought-tolerant shrub species predominating in Mediterranean areas) and the biocrust layer dominated by the lichen Cladonia rangiformis, on soil microbial biomass; community composition; and the activity of acid phosphatase, β-glucosidase, peroxidase, polyphenol oxidase, β-N-acetylglucosamine, leucine aminopeptidase and urease soil enzymes. In addition, in order to examine the effects of the spatial heterogeneity induced by crust's presence or absence combined with the plant presence on soil microbial attributes, the samples were collected along a straight line of 100 cm from the trunk of the shrubs towards their periphery in crusted and uncrusted soils. Soil habitats in Mediterranean areas are characterized as highly heterogeneous in time and space (fine grained mosaic; Stamou, 1998), where biotic and abiotic drivers affect the assemblages of soil communities across scales ranging from hectares to square centimeters (Wu et al., 2013). In this study, we hypothesized that the spatial heterogeneity will be reflected in quantitative microbial estimates. Higher microbial biomass and enzyme activity were expected at the base of the shrubs (0 to 20 cm from the trunk) in crust-covered areas where both biotic and abiotic conditions were favorable; high amounts of organic input (dead organic material and root exudates) are inserted into the soil, while crust cover acts as a buffer system for soil life against abiotic stress. Apart from quantitative estimates, we expected differences in the soil microbial community composition and the network of interactions among the microbial members even at distances of a few centimeters.

Biological crusts play an important role in plants' establishment and growth and vice versa. In this study, we examined the influence of the interplay between Cladonia rangiformis and Erica spp. and whether the spatial heterogeneity created by their interactions affected the biomass, the composition and the network of interactions among the members of the soil microbial communities. Soil water content was higher in areas covered by BSC than in those where biocrust was absent. By definition, BSC is a cohesive layer that can retain higher levels of moisture (Pointing and Belnap, 2012). Although it is frequently reported that BSCs resulted in significant stocks of soil C and N in arid and semiarid areas (Chamizo et al., 2012), this was not reported in this study. Gao et al. (2010) found that BSCs had no effect on these parameters at deep soil depths, suggesting that the effect of BSCs on soil carbon and nitrogen is restricted to the upper soil layer (up to 5 cm). This is probably the explanation for the lack of effect in our study since the samples were collected from the upper 10 cm of the soil.

The saturated/unsaturated ratio is indicative of the concentration of carbon or nutrients in the soil and increases as the concentration of nutrients decreases (Fierer et al., 2003). This ratio was higher in soils covered both by Erica and BSCs, denoting possible nutrient limitation in these soils and implying competition between C. rangiformis, Erica shrubs and soil microbes for nutrient uptake. It seems that, although biological crusts are known for their ability to fix atmospheric N, a fact that was in agreement with the higher organic N recorded at the base of the shrubs (0 cm distance) in crusted soils, this was not adequate to alleviate nutrient limitations. Further, the changes in saturated/unsaturated ratio could indicate changes in the composition of the microbial community. A positive relation between the gram+/gram- ratio and the saturated/unsatured ratio was mentioned by Konstantinou et al. (2019) and is based on the fact that most gram-positive bacteria contain saturated PLFAs (Cho and Salton, 1966). The changes in community composition were supported by changes in the gram+/gram- ratio, which was increased in crusted soils. This increase could be attributed to the decreased C availability or the exhaustion of labile substrates (Fanin et al., 2019). Given that the area is covered by the drought-tolerant evergreen sclerophyllous Erica spp., which exhibit slow rates of nutrient recycling (Mediavilla and Escudero, 2003), the low labile carbon in crusted soils could be attributed to consumption of labile C by the BSC microorganisms. The latter, in combination with the fact that through fungal metabolism lichens produce phenolic compounds which are considered recalcitrant, could explain the increase in the gram+/gram- ratio. Furthermore, other possible explanations for the increased gram+/gram- ratio could be due to the resistant (against exogenous factors) cell wall of gram+ bacteria (Rohde, 2019) that protects bacteria from the excretions of the lichen.

In crusted soils significantly lower values of the iso/anteiso ratio were recorded, implying better abiotic conditions (temperature, humidity) compared to uncrusted ones since a high value of this ratio is generally used as an environmental stress indicator (Pettersson and Baath, 2003). Bacteria could decrease the content of anteiso fatty acids when growing under adverse conditions to decrease the viscosity of their cytoplasmic membranes. The high values of the ratio in uncrusted soils could be the result of the harsh temperature or moisture conditions prevailing in these soils. In fact, the water content in uncrusted soils was half of those in crusted ones.

Despite better abiotic conditions prevailing in crusted soils, microbial biomasses were almost 4 times lower compared to soils covered only with Erica shrubs. The latter contrasts literature data reporting a positive relation between the abundance of microorganisms and the presence of BSC (Guan et al., 2018) in arid and semiarid soils due to crusts' influence on the distribution of soil moisture. Therefore, the reduction in microbial biomasses in the present study may be explained as a negative by-product of Cladonia activity. It is thought that the allelochemicals' secondary metabolites of this organism containing aliphatic, aromatic and terpenoid halogenated components are involved in various chemical interactions, responsible for their antimicrobial properties (Kosanić et al., 2014). This idea is reinforced by a number of studies that show the negative effect of these lichens' substances on the biomass of soil microbes, tested both in the field and the laboratory (Pandey, 2017; Kalra et al., 2021).

Soil enzyme activity displayed an opposite pattern compared to that of microbial biomass, with higher values in crusted soils except for LAP and PPO. Besides, it is well documented that enzymes originate from a variety of sources such as living and dead microbes, plant roots, residues and soil animals (Sun et al., 2013). Thus, enzyme activity is not necessarily associated with microbial biomass. Ghiloufi et al. (2019) has reported that enzyme activity is high in soils covered by BSC because in most cases the presence of crusts is associated with high soil moisture, nutrients and organic matter content. In our study, by the inspection of the saturated/unsaturated ratio we identified nutrient limitation in soils covered by crust and Erica shrubs. Hence, soil microbial communities possibly produced enzymes for substrate degradation to increase nutrient availability for their nutritional needs according to the nutrient allocation theory (Allison and Vitousek, 2005). Also, if we consider the enzyme activity per unit of microbial biomass, this was much higher in the microbes of crusted soils, indicating more metabolically active microbial communities. The latter could be partially supported by the better abiotic conditions prevailing in these soils.

It is interesting that spatial heterogeneity had a limited effect on microbial biomasses and enzyme activity. We observed a common pattern concerning the effect of distances on the microbial biomass, regardless of the presence of BSC. The microbial groups that were affected negatively by shrub presence were bacteria and actinomycetes, which displayed the lowest biomass at 0 cm distance from the shrubs and the highest at 40 cm. The fact that bacteria exhibited some kind of spatial heterogeneity is probably related to their limited ability to move and absorb nutrients. The active modes of their dispersal are limited to the scale of (connected) soil pores or aggregates (Yang and van Elsas, 2018). In contrast, the hyphal growth form of fungi likely conveys several advantages to these organisms, chief of which is the translocation of nutrients and resources from microsites where they are abundant to sites where they are limiting (Frey et al., 2003). As concerns the decrease in total microbial biomass at the shrub base, this could be attributed to the roots' increased density underneath the trunk, resulting in stronger competition for nutrients. Furthermore, phenolic compounds, which seem to have antimicrobial effects, are abundant in Ericaceae species (Malik, 1995). In some Ericaceae species such as E. andevalensis, a variety of phenols and flavonoids have been isolated and have proven to have allelopathic antimicrobial activity (Marquez-Garcia et al., 2009). Also, it has been found that this plant requires an acidic soil environment for its growth; however, this characteristic may provoke a quite inhospitable environment for the majority of soil microorganisms (Vazquez et al., 2020).

Beside quantitative measurements such as biomass and activity that displayed limited spatial heterogeneity, the composition of the local microbial communities was totally different from one point to another. At each soil point along a line of 100 cm from the trunk to the periphery of the shrubs, unique microbial communities were identified. This can be viewed as an effect of the fine-grained mosaic that characterizes the Mediterranean ecosystems (Stamou, 1998). The effect of spatial heterogeneity was mirrored at the microbial network topological features as well. The network in the area without cover (bare soils; Fig. 3f) was very well organized; the crust itself resulted in network fragmentation (Fig. 3c) probably due to released antimicrobial substances that acted selectively to the members of the microbial communities. Through negative effects on microbes, because of the acid soil environment and allelochemical substances contained in litter, the network in areas with only Erica seems to have collapsed (Fig. 3d). Remarkably, when both Erica and crust acted together (Fig. 3a), the microbial network was organized quite well as if the selective action of the crust on microbes were compensated by the selective action of Erica spp. or the opposite. Also, by comparing the networks in Fig. 3a and f we observed that the substances leached by Erica spp. and crust acted much more selectively for the microbial interactions than the adverse abiotic conditions prevailing in bare soil. The well-organized network in soils uncovered by Erica and crust soils with a high number of direct interconnections of relatively high intensity reflects a microbial community adapted to adverse abiotic conditions of the Mediterranean areas (Stamou et al., 1998). However, this was the only network that did not show characteristics of small-worldness. According to Peng et al. (2016) there is a trade-off between small-worldness and robustness. Thus, the organization of interactions between the members of the microbial communities in Mediterranean bare soils revealed them as the more robust. For these communities the risk of transition from stability to instability in the case of future disturbances such as trampling, fire or extreme watering is low. A similar response was reported by Papatheodorou et al. (2020) for the same Mediterranean area; the composition and the biomass of soil microbial communities in bare soils showed higher resistance to frequent-watering events than the communities in crust-covered soils.

We studied the effect of the spatial heterogeneity created by the crust C. rangiformis and the shrub Erica spp. on microbial biomass, community composition, network of interactions and enzyme activity in the soil. The quantitative measurements such as biomass and activity showed limited and no response to spatial heterogeneity, respectively; they changed mainly in relation to coverage regime. In contrast, the qualitative characteristics of the microbial communities such as their composition and networks' topology confirmed the idea of a fine-grained mosaic existing in the Mediterranean areas within distances of a few centimeters. Although the most well-organized and robust network was that in bare soils, it seems that Erica spp. and crust coverage acted in concert and conditioned the development of more connected networks compared to that produced by each agent alone. The interactions between soil microbial community, soil functionality, BSC and local vegetation are particularly complex. Hence, a multifaceted approach is recommended where qualitative and quantitative aspects are examined jointly.