Soil respiration responses of moss and lichen biocoenoses to moderate and severe rain events after summer drought in a temperate early‐successional ecosystem

It is commonly accepted that CO2 efflux increases with soil water content in aerated soils and that rewetting after periods of soil drying can result in respiration pulses. It has further been shown that soil pores may become water logged which can impede soil gas exchange. The present study aimed to quantify the carbon response of moss (Polytrichum piliferum) and lichen (Cladonia coniocraea) biocoenoses to different quantities of rain in an artificial catchment, which granted that the starting point of the development of both biocoenoses was the same. To address this aim, we conducted in situ soil moisture and soil respiration measurements, where soil respiration was hypothesised to emerge from a cryptogamic vegetation layer and from mineral soil beneath. We found that higher water‐holding capacity of the moss layer and higher accumulation of organic matter in the upper mineral soil under mosses result in higher amounts of water stored near the surface. As a consequence, evaporation of water as well as pulses of CO2 efflux after moderate rain following a period of drought were higher in the moss biocoenosis, where the upper mineral soil was of key importance. In contrast, the lichen biocoenosis facilitated penetration of rainwater into the deep soil. Superimposing rewetting pulses, near‐saturation of soil pores with water after severe rain resulted in gas exchange inhibition and diminished soil respiration until subsequent aeration in both biocoenoses.


| INTRODUCTION
Mosses and lichens are widespread on earth in almost all terrestrial ecosystems and habitats.Together with heterotrophic and photoautotrophic bacteria, archaea, protists, green algae, fungi, lichens, mosses, and liverworts, they form large-scale biological soil crusts and colonise the upper millimetres of the soil (Weber et al., 2016(Weber et al., , 2022)).Those biocrusts (also known as biological soil crusts [BSCs]) alter ecosystem functions and ecological processes: They improve soil stability, reduce soil erosion, play a crucial role in changing hydrology and soil moisture (Kidron, Lichner, et al., 2022) and play an essential role in nitrogen fixation and cycling (Brankatschk et al., 2013;Kidron et al., 2015;Kidron, Fischer, & Xiao, 2022).They affect multiple physiological and ecological processes of the carbon cycle by fixing atmospheric CO 2 and increasing the amount of carbon stored in soils, as well as modulate carbon losses to the atmosphere via soil respiration (Yao et al., 2019).
As poikilohydric organisms, mosses and lichens (and the other biological components of BSCs) largely depend on precipitation in the form of rain, dew and fog to be physiologically active and to carry out photosynthesis and respiration (Fischer, Veste, Bens, & Huettl, 2012;Jung et al., 2020;Lange et al., 1992;Ma et al., 2022;Veste et al., 2008).
However, very intensive studies on this have been carried out primarily in the arid regions of the world, where BSCs are very prominently distributed (for an overview, see Belnap and Lange, 2003;and Weber et al., 2016).Depending on the seasonality of precipitation between winter and summer rainfall, there are significant differences in the physiological function of organisms and their influence on carbon cycling processes.Humid periods alternate with prolonged dry phases, so that the carbon-related physiological processes are largely controlled by rain pulses (Bowling et al., 2011).Characteristics for the precipitation in drylands are small events less than 5 mm and often lower than 1 mm.These precipitation events are still sufficient to stimulate the physiological activities of the cryptogams.The water optimum of many biocrusts organisms varied, for example, a study by Tamm et al. (2018) in the Succulent Karoo of South Africa showed 0.52-0.78mm for cyanobacteria/cyano-lichen biocrusts, 0.75-1.15mm for chloro-lichen crusts and 1.76-2.28mm for mossdominated biocrusts.However, the observed low rainfall intensities in combination with BSCs have major negative impacts on ecohydrological processes such as water percolation into deep soil or surface run-off (Li et al., 2016;Yair et al., 2011), and finally on the wetness of deeper soil layers and the related carbon processes (Kidron & Grishkan, 2023;Wu et al., 2015).
In contrast to the arid and semi-arid regions, there are few studies from temperate ecosystems, and here, too, the BSCs as well as mosses and lichens colonise open habitats, but can also be found in open-canopy forests (Corbin & Thiet, 2020;Rieser et al., 2021).Under the high precipitation in the temperate climate and the associated competition with higher plants, the cryptogams are mainly associated to edaphic dry sites, such as open grasslands, sand dunes or initial habitats.Such initial ecosystems are postmining sites.
The study site, Hühnerwasser, is a unique post-mining site in Brandenburg, Germany, which was constructed in 2005 as an experimental area to study early stages of ecosystem succession and which had been left to natural development (Gerwin et al., 2011;Schaaf et al., 2011).During the first years after the site establishment, initial stages of BSCs were dominated by cyanobacteria and green algae, which stabilised the soil surface against water erosion and deflation (Cania et al., 2020;Fischer et al., 2010;Spröte et al., 2010).Five years after construction, BSCs at different stages of development were heterogeneously spread over the study site and had formed mosaic-like patterns, which further were associated with the appearance of vascular vegetation (Spröte et al., 2010;Zaplata et al., 2013) (Cania et al., 2020;Duemig et al., 2014;Fischer, Veste, Eisele, et al., 2012;Gypser et al., 2016;Lukesova, 2001).
As the dominant life form, these cryptogams contribute significantly to soil carbon accumulation during early succession (Duemig et al., 2014).Own previous laboratory studies on carbon fluxes have shown that these are of great importance for the carbon cycle (Dietz et al., 2019;Fischer & Veste, 2018;Fischer, Veste, Eisele, et al., 2012;Gypser et al., 2016).
Even under temperate climatic conditions, pronounced prolonged dry periods are characteristic of the eastern region of Brandenburg, leading to significant drying out of the sandy soils.It can be assumed that precipitation pulses play an important role in ecohydrological processes and the associated physiological processes of photosynthesis and soil respiration.
It is commonly accepted that CO 2 efflux increases with soil water content in aerated soils and that rewetting after periods of soil drying can result in respiration pulses (Birch, 1958;Fischer, 2009;Manzoni et al., 2020).It has further been shown that soil pores may become water logged (Fischer et al., 2010), which can impede soil gas exchange (Cleveland et al., 2010;Deng et al., 2011Deng et al., , 2017)).In addition, limited supply of oxygen caused by water saturation may reduce microbial activity.The present study aimed to quantify the carbon response of soils with two different types of cryptogamic vegetation in an artificial catchment, which granted that the starting point of the development of both biocoenoses was the same.
To address this aim, we conducted in situ soil respiration measurements on moss-and lichen-dominated soils after one moderate and two severe rain events, where soil respiration is hypothesised to emerge from a cryptogamic topsoil layer and from mineral soil beneath.It is further hypothesised that-in addition to possible surface-run-off generation reported previously (Fischer et al., 2010)surface pore clogging superimposes rewetting effects on soil respiration.

| Research site
The research site is located at the artificially constructed water catchment 'Hühnerwasser/Chicken creek' (51 36 0 21 00 N, 14 15 0 54 00 E) in the Lusatian open-cast mine Welzow Süd 20 km south of Cottbus (Brandenburg, Germany).The climate is characterised as temperate and slightly continental with high summer temperatures and pronounced dry periods in the growing season.Mean annual precipitation is 566 mm, and average air temperature is 10.1 C (DWD meteorological station Cottbus, ID 880; 51 46 0 33 00 N, 14 19 0 3 00 E, 69 m.a.s.l).On an annual average of 48.4 days, covering the months of October to April with a maximum in January, the snow depth averaged 0.7 cm.
Mean annual precipitation from the weather station installed directly at the site is 567 mm and average air temperature is 10.8 C during 2005-2020(Chicken Creek data portal, 2021).There was no snow during the study period.
The dumped Quaternary soil substrates originated from the forefield of the mine and were generally characterised as sands to loamy sands containing low amounts of carbonate (Gerwin et al., 2011).The soil was classified as neutral sandy regosol.
The water catchment area consists for the most part of a mosaic build-up of ruderal vegetation.Almost the entire northern half can be attributed to this biotope type of grass and herbaceous vegetation, which is typical for post-mining areas in the region (Landeck et al., 2017).The raw substrate here has been artificially applied and provides a good substrate for pioneer species.Dominant is often Calamagrostis epigejos, while in wetter areas, Carex Carex species are typical plants (Figure 1).Here, for example, Cladonia coniocraea, Cladonia rangiferina agg., Cladonia cf.chlorphaea and other lichen species can be found on post-mining areas (Gypser et al., 2015;Landeck et al., 2017).

| Experimental setup
For this comparative study, we selected two different plots (Figure 1.).
The first plot (Figure 2) was dominated by lichen C. coniocraea with more than 70% coverage in Chamber 1 (the rest were H. pilosella and A. pannonica) and around 80% in Chamber 2. The second plot was dominated by moss P. piliferum with 80% of coverage in Chamber 3 (the rest consisted of H. pilosella, A. pannonica, Vicia hesuta and Cladonia coniocrae) and almost 90% in Chamber 4 (the rest was mostly H. pilosella with little amount of C. coniocraea).The thickness of lichen-dominated cryptogamic layer was around 1 cm, and the thickness of moss-dominated cryptogamic layer was around 2 cm, not taking into account the height of single branches of both lichens and mosses.
F I G U R E 1 Experimental area with (a) lichen-dominated plot and (b) moss-dominated plot.

| Soil physico-chemical analyses
Soil samples were taken from six spots: two spots close to the pair of chambers on moss-dominated soil, two in the middle between the two sets of chambers and two close to the pair of chambers on lichens.The samples were taken at three depths at each spot (0-5 cm, 5-10 cm and 10-20 cm).Soils were passed through a 2-mm sieve to remove roots and gravel and then separated into two fractions.One soil fraction was immediately placed in the freezer for analysis of available nitrogen (AN), and the other was air-dried for further analysis.pH (soil: H 2 O 1:2.5, w/w) and electrical conductivity (EC) were measured using a conductivity meter (MultiLab 540, WTW-Wissenschaftlich-Technische Werkstätten GmbH, Weilheim i. OB, Germany).Soil texture was determined by sieving method.
The total carbon (TC) and nitrogen (TN) for all soil samples were measured by oxidative combustion with a CNS-Analyser (VARIO EL, Elemental Analyser, Hanau, Germany).The soil organic matter content was estimated using the loss-on-ignition method, after igniting the soil samples at 450 C for 12 h in a muffle furnace (Bisutti et al., 2007).The carbon was subsequently measured again, and the total organic carbon (C org ) was calculated by the difference in TC before and after muffling.Total phosphorus (TP) was measured with ICP-OES (Thermo Scientific iCAP6000 Duo, Bremen, Germany) after high temperature and pressure digestion with HNO 3 (65%) in a thermal oven (Loftfields Analytical Solutions, PDS-6 Pressure Digestion System, Neu Eichenberg, Germany).Available N was measured by CaCl 2 extraction method (Dou et al., 2000).Water-holding capacities were determined gravimetrically after water saturation using an Eijkelkamp sand bed apparatus.Samples for soil physico-chemical analysis were taken from the same moss and lichen patches, but from outside the collars.

| Experimental design and measurements of soil CO 2 fluxes
Soil chamber collars (20 cm diameter Â 10 cm height, PVC, Figure 2) were inserted into the soil 3 weeks before the measurements and remained in place for the study duration.Soil CO 2 fluxes were measured using an automatic closed dynamic soil CO 2 flux system (LI-

| Microclimate
Local microclimatic measurements were recorded in 10-min intervals directly at the study site with a microclimate station with a ZL-6 data-logger (MeterGroup, Munich, Germany) equipped with air temperature, relative humidity, barometric pressure sensor (Atmos-14), sonic anemometer for wind speed and direction (Atmos-22), global radiation sensor (Apogee Instruments, Logan, Utah, USA) at 2-m height and a high resolution rain gauge rainfall (ECRN-100) at 1-m height.Soil temperature and moisture echo probes (ECH2O EC-TM, Decagon Devices, Inc., Pullman, Washington, USA) were installed in adjacent soil profiles at each plot between the two chambers at three soil depths (0-5 cm, 5 cm and 10 cm) below the cryptogamic layer.

| Data and statistical analysis
All microclimatic data collected at 10-min frequencies were aggregated to hourly sampling intervals to maintain consistency across data F I G U R E 2 Selected measurements plots.From left to right: Chamber 1 with lichen-dominated BSC, Chamber 2 with lichen-dominated BSC, Chamber 3 with moss-dominated BSC and Chamber 4 with moss-dominated BSC.
sets.The CO 2 raw data were automatically fitted to exponential function by the LI-COR software SoilFluxPro.Soil moistures and soil temperatures between the moss-and lichen-dominated variants were compared using paired Wilcoxon signed-rank tests after verifying that no lag between the time series existed using the ccf function of R's tseries package.Multiple linear regression was used to identify statistical significance of the driving factors for soil CO 2 efflux using the lm function of R. Relations between soil temperature, soil moisture and soil CO 2 efflux were visualised using the scatter3d function of R's car package with quadratic surface fit.

| Soil physico-chemical and hydrological properties
The soil was classified as neutral sandy regosol.Water saturation was estimated at a volumetric water content of 39%-40% by volume, where below approximately 3% by volume, the remaining water can be considered as not bioavailable.The mosses accumulated significantly higher amounts of carbon and nitrogen in the vegetation layer and throughout the soil profile than the lichens (Table 1).

| Environmental factors and soil respiration dynamics
During the study period, 68% of all rain events had amount of less than 5 mm per event, and 42% had even less than 1 mm per event.Soil moisture had a pulse-driven pattern following rainfall events.The volumetric soil water content at lichen-and moss-dominated soils reached on average 0.08 and 0.09 m 3 m À3 at a depth of 0-5 cm, 0.07 and 0.07 m 3 m À3 at 5 cm, and 0.09 and 0.08 m 3 m À3 at 10 cm, respectively, with highest hourly mean values of 0.29 and 0.34 m 3 m À3 at 0-5 cm soil depth, respectively.The soil temperature at lichen-dominated and moss-dominated soils reached 17.6 C and 17.1 C at a depths of 0-5 cm, 17.6 C and 17.0 C at 5 cm, 17.0 C and 17.2 C at 10 cm on average, respectively, with the highest hourly mean values of 48.7 C and 44.3 C at 0-5 cm, respectively.The differences in soil volumetric water content and soil temperature between lichen-dominated and mossdominated soils were significant at all depths (p < 0.001, paired Wilcoxon signed-rank test).
The hourly soil respiration rate at the soils with lichendominated communities in Chamber 1 reached a maximum of 15.20 μmol CO 2 m À2 s À1 on 26 August 2020 and a minimum of 1.56 μmol CO 2 m À2 s À1 on 17 August 2020, and the average value was 6.50 ± 2.58 μmol CO 2 m À2 s À1 ; in Chamber 2, they reached a maximum of 9.96 μmol CO 2 m À2 s À1 on 09 September 2020 and a minimum of 1.04 μmol CO 2 m À2 s À1 on 20 October 2020, and the average value was 3.39 ± 1.19 μmol CO 2 m À2 s À1 ; the soils with moss-dominated communities in Chamber 3 reached a maximum of 5.89 μmol CO 2 m À2 s À1 on 09 September 2020 and a minimum of 0.83 -μmol CO 2 m À2 s À1 on 14 October 2020, and the average value was 2.44 ± 0.75 μmol CO 2 m À2 s À1 ; and in Chamber 4, they reached a maximum of 7.85 μmol CO 2 m À2 s À1 on 14 September 2020 and a minimum of 0.94-μmol CO 2 m À2 s À1 on 20 October 2020, and the average value was 2.97 ± 0.93 μmol CO 2 m À2 s À1 (Figure 3).
The daily C release by soil respiration of the lichen-dominated patch in Chamber 1 varied from 2.12 to 11.20 g C m À2 with a mean of 6.73 ± 2.47, in Chamber 2 from 1.56 to 5.43 g C m À2 with a mean 3.51 ± 0.97, in Chamber 3 from 1.29 to 4.12 g C m À2 with a mean 2.53 ± 0.66, in Chamber 4 from 1.42 to 4.48 g C m À2 with a mean 3.08 ± 0.79.During the period of the experiment by lichendominated soils, the total C amount of 417.54 g C m À2 was respired in Chamber 1 and the C amount of 217.68 g C m À2 in Chamber 2; by moss-dominated soils, the amount of 156.70 g C m À2 was respired in Chamber 3 and 190.69 g C m À2 in Chamber 4 (n = 63 full days).

| Wetting-drying cycle inducing respiration pulses
Initially, the cryptogamic layer and the mineral soil were dry.During the first precipitation event (P 1 : 17 August 2021, 7:30 PM-7:40 PM totalling to 1 mm) the precipitation was entirely intercepted by the cryptogamic layers and caused a respiration flush (denoted as 'Birch effect cryptogams' in Figure 4).Until a third precipitation event (P 3 ) at 18 August 2021 08:30 AM-10:40 AM (totalling to 4.8 mm), it is very unlikely that the cryptogamic covers dried out overnight (r.h. was 78 ± 5%), so the respiration increased to base level for wet cryptogams (bR wet cryptogams) after the respiration flush.These findings are substantiated by a second precipitation event P 2 at 18 August 2021 5:50 AM-6:00 AM (totalling to 0.4 mm), which did not percolate into the mineral soil nor did it induce a respiration flush or altered cryptogamic basal respiration.It further can be concluded that the Birch effect was fully pronounced and not limited by water deficiency for cryptogamic covers of all variants.
After the third event P 3 , a second respiration flush was observed, which can be attributed to the underlying mineral soil after moistening by infiltrating rainwater (denoted as 'Birch effect mineral soil' in T A B L E 1 Carbon storage at various depths and water-holding capacities of the cryptogamic layer, mean ± standard deviation (n = 2).

Mosses Lichens
Cryptogams    where CO 2 efflux is shown for Chamber 1 (Lichens).During hydration States 1 and 2, the rain events inducing the respiration pulse were followed by a period with occasional light rain.Similar to the respiration pulse (Figure 4a), the water contents of the top 5 cm of the mineral soil were significantly higher under the moss layer (Figure 5a), pointing to

| Wetting-drying cycle and wetting event inducing gas diffusion inhibition
(a) Precipitation (blue columns, left axis) and volumetric water content of the mineral soil (solid red line: lichens 0-5 cm, dashed red line: lichens 5-10 cm, solid green line: mosses 0-5 cm, dashed green line: mosses 5-10 cm).(b) Soil respiration (in μmol C m À2 s À1 ) for Chambers 1-4 (blue, red: lichens; green, yellow: mosses); bR: basal respiration.The hydration status of cryptogams and upper mineral soil is shown on top, where hydration of the soil profile is symbolised in bluish tones.
T A B L E 2 Basal respiration and total amount of C released during respiration pulses (background corrected Birch effect) before and after events P 1 and P 3 , explanation of abbreviations: see Figure 4.   4; wetted phase in Figure 5).A subsequent drop of the soil moisture content to about 0.15 m 3 m À3 was accompanied with an increase of soil respiration to the maximum recovery level before the rain event.
Hydration states 5 and 6 were characterised by soil dry-out, which coincided with a decrease of soil respiration (dry-out phase in Figure 5).Tables 5 and 6.Figures 6 and 7 exemplify the relation between soil CO 2 efflux, moisture and temperature of the mineral soil in Chamber 1 (lichens) during the first wetted and dry-out phases (Figure 6) and the second wetted phase (Figure 7).
We observed a significantly positive response of the CO 2 efflux from soil moisture during the dry-out phase in all variants and a significantly negative response of the CO 2 efflux from soil moisture in Chambers 1-3 during the first wetted phase.Lower regression coefficients indicate that the negative effect was more pronounced in the lichen cover during the wetted phases.There was a negative correlation between soil temperature and soil respiration during the recovery phase (except Chamber 3) and during the dry-out phase (except Chamber 1).During the second wetted phase, the correlation between soil moisture and soil respiration was negative and between soil temperature and soil respiration was positive.Due to moist conditions, the second wetted phase was not followed by an immediate dry-out phase (Figure 3).Soil moisture contents decreased faster in the moss variant compared to the lichen variant during the dry-out phase, pointing to enhanced evaporation through the moss layer (Figure 5a).

| DISCUSSION
Our data agree well with literature findings.For lichen-dominated biocrusts, Zaady et al. (2000) reported maximum respiration rates of T A B L E 3 Recovery phase (20-30 August 2020): regression coefficients of soil respiration (in μmol C m À2 s À1 ) with time (in μmol C h m À2 s À1 ), volumetric water content (Θ in m 3 m À3 ) and temperature (T in C) of the mineral soil (0-5 cm) underlying the cryptogamic layer.4.1 | Rewetting induced respiration pulses Jarvis et al. (2007) and Barnard et al. (2020) reported that rewetting pulses in soils emerge as a result of (1) disruption of soil aggregates and exposition of previously protected organic matter for decomposition (Denef et al., 2001); (2) microbial necromass by soil drying followed by decomposition of postmortal microbial biomass on rewetting (Bottner, 1985); (3) spontaneous rapid increase in microbial biomass in response to the availability of water (Griffiths & Birch, 1961;Jager & Bruins, 1975;Orchard & Cook, 1983;Scheu & Parkinson, 1994); and (4) release of compatible solutes as a microbial hypo-osmotic stress response (Fierer & Schimel, 2003;Kieft et al., 1987).Along with abiotic CO 2 release caused by solubilisation of carbonates, CO 2 displacement, changes in water film connectivity and degassing of CO 2 dissolved in rain and due to the decrease in barometric pressure over time, Barnard et al. (2020)  and an autotrophic alga and that mosses, in contrast, are formed exclusively from phytomass, we expected a different response to rewetting.Surprisingly, there is no significant difference in the rewetting effect between the lichen and moss layers (Table 2).The assumption that the amount of precipitation at P 1 was insufficient to fully express the rewetting pulse due to lower water bioavailability must be rejected, because in this case a pulse would have occurred after P 2 (Figure 4).Morillas et al. (2017) reported that small water pulses can activate the metabolism of carbon in soils through lichens, mosses and cyanobacteria associated with biocrusts while deeper soil layers remain dormant.Although both variants of our experiment received the same amount of precipitation (P 1 , P 2 and P 3 ), the increase in mineral soil water content at 0-5 cm depth was higher under the moss cover after P 3 (solid green line in Figure 4a).Both lower water-holding capacity (Table 1) and percolation into deeper soil (dashed red line in Figure 4a) point to lower storage of water near the surface in the lichen variant.
The lower amounts of organic matter in the upper mineral soil (Table 1) consequently led to a significantly lower second respiration pulse under the lichen cover.Manzoni et al. (2020) found rewetting respiration pulses peaked at around 1 mg CO 2 -C g À1 soil C for a wide range of mineral and organic soils.In our study, rewetting respiration pulses averaged to 0.54 and 1.90 mg CO 2 -C g À1 soil C for mosses and lichens, respectively.In our study, the relative amount of biomass and organic matter vulnerable to rewetting respiration pulses is with this range for the mosses, but higher for the lichens, which may indicate ongoing progression of organic matter accumulation and development of heterotrophic decomposers (Odum, 1969).

| Wetting-drying cycle inducing gas diffusion inhibition
It is well accepted that soil respiration increases with soil moisture, or soil water bioavailability.At wet sites, however, soil CO 2 efflux has been reported to be insensitive to moisture change and may be inhibited after rain events because excessive water would decrease soil oxygen diffusion (Cleveland et al., 2010;Deng et al., 2011Deng et al., , 2017)).It has further been reported that limited aeration in a moss-dominated biocrust also explained the high concentration of total ammonia and nitrite (Kidron et al., 2015).In our study, we observed a negative correlation between moisture and CO 2 efflux immediately following severe rain (Figures 6 and 7, Tables 5 and 6), which points to replacement of air by water in the soil pore system (pore clogging) and which coincides with risk of Hortonian surface run-off reported previously (Fischer et al., 2010;Kidron, Lichner, et al., 2022).
After drainage of water into deep soil (Figure 3), we observed a positive correlation between moisture and CO 2 efflux during the dryout phase, which we attribute to the opening of pores which coincides with the onset of aeration.To explain the positive relation between temperature and respiration in both wetted phases but the negative relation in dry-out phase (Table 5), it needs to be considered that respiration originates from both cryptogamic layer and mineral soil.During the dry-out phase, the cryptogamic layer dries out first, resulting in faster drying with increasing temperature and causing respiration earlier to cease at higher temperatures.Voortman et al. (2014) reported that the modulating effect of mosses on evaporation possibly differs between wet and dry climates and that under dry conditions, mosses and lichens are able to maintain a moisture supply from the soil, leading to a higher evaporation rate than mineral soils.Moreover, higher water-holding capacities of the vegetation layers may retain higher amounts of water near the surface, where increased evaporation takes place shortly afterwards due to the decrease of the relative humidity in the atmospheric layer and the lower albedo of the crust which facilitates an increase in surface and subsurface temperatures (Kidron, Fischer, & Xiao, 2022).Our results best fit a pattern of evaporation for dry climates, and there is no support of the notion that cryptogamic biocoenoses impeded evaporation in our study.
All phenomena associated with water saturation of the cryptogamic layer and the mineral soil generally require hydration first, which implies overcoming possible water repellency.A biocrust dominated by Polytrichum has been reported to be water repellent (Lichner et al., 2018), which may result in Hortonian surface run-off when dry (Kidron, Lichner, et al., 2022), and in the formation of biocrust surface patches (Fischer et al., 2014) that facilitate preferential water flow into the deep soil (Lichner et al., 2018).By installing closed rings down to the mineral soil, we could not take into account the influence of lateral water flows, which imply a loss of water for the biocoenosis.On the other hand, we are not aware of any system that can simultaneously record surface run-off and gas exchange in the field.We see this limitation as a major challenge for small-scale research into the relationship between hydrological surface processes and the C balance of cryptogamic biocoenoses.

| CONCLUSIONS
We conclude that the allocation of infiltrating rainwater depends on vertical distribution patterns of organic matter in cryptogamic biocoenoses.Higher water-holding capacity of the moss layer and higher accumulation of organic matter in the upper mineral soil under mosses result in higher amounts of water stored near the surface.As a consequence, evaporation of water as well as pulses of CO 2 efflux after moderate rain following a dry period were higher in the moss biocoenosis, where the upper mineral soil is of key importance.In contrast, the lichen biocoenosis facilitated penetration of rainwater into the deep soil.Superimposing rewetting pulses, saturation of soil pores with water after severe rain result in gas exchange inhibition and diminished soil respiration until subsequent aeration in both biocoenoses.
species can be found.Other common species are Trifolium arvense, Daucus carota, Leontodon taraxacoides, Oenothera parviflora agg., Echium vulgare, Achillea pannonica, Agrostis capillaris, and with progressive succession, shrubs and pioneer trees such as Elaeagnus rhamnoides (syn.Hippophae rhamnoides), Genista pilosa, Pinus sylvestris and Robinia pseudoacacia become established.Due to their high light demand, well-developed cryptogamic covers dominated by mosses and soil lichens are associated with the more open grassland patches were Hieracium pilosella, A. pannonica, Vicia hesuta and 8100, LI-COR Inc., Lincoln, Nebraska, USA).It consisted of an infrared gas analyser equipped with an 8-Port Multiplexer unit (LI-8150, LI-COR Inc., Lincoln, Nebraska, USA) connected to four long-term dark respiration chambers (20 cm in diameter; system volume 4093 cm 3 ; LI-8100-104, LI-COR Inc.).Chambers were oriented to face south when closed to minimise shading effects inside the collars.The dark respiration was measured in each chamber in turn once per hour for 3 min and 15 s, including a 30-s pre-purge, 45-s post-purge and 2-min measurement period.The number of replicates was estimated based on the size of the cryptogamic vegetation patches, and the consideration that a spatial expansion of the installation would lead to a mixing of the within-patch and the between-patch variability.It was on the other hand not possible to increase the density of collars within the patches to minimise the impact of the hardware installation on the biocoenosis.The experimental design was based on the study of Tucker et al. (2017).Measurements were carried out from 11 August to 21 October 2020.

Figure 4 )
Figure 4), again levelling off to basal respiration (bR wet cryptogams + mineral soil).While rainwater confined in the upper 5 cm of the mineral soil below the mosses, it percolated into deeper layers below the lichens.Finally, a slow decrease due to soil desiccation was observed.All these respiration components likely are superimposed

Figure 5
Figure 5 exemplifies the three ecohydrological phases identified following the rewetting-induced respiration pulse (17-21 August 2020), lower water-holding capacity or enhanced drainage by mosses.The soil respiration of Chamber 1 (lichens) increased until Aug 31 from about 3 μmol C m À2 s À1 to maximum values reaching 14 μmol C m À2 s À1 , which we attribute to recovery of the biocoenosis after droughtinduced dormancy (recovery phase in Figure5).This increase was less pronounced but yet significant in Chambers 2-4.Table3summarises the regression coefficients during the recovery phase, where the trend over time represents the recovery of the biocoenosis.During hydration States 3 and 4, a severe rain event from August 30-31 summing to 70.8 mm coincided with a rapid decrease of CO 2 efflux to a minimum of 3.6 μmol C m À2 s À1 , where the maximal moisture content of the mineral soil (Θ max = 0.35 m 3 m À3 in both variants) approached water saturation (Θ r = 0.39 m 3 m À3 in Table The regression coefficients of soil respiration, volumetric water content and temperature of the mineral soil are given in F I G U R E 5 Hydration states and ecohydrological phases from 20 August 2020 to 11 September 2020.(a) Precipitation (blue columns, left axis) and volumetric water content in the top 5 cm of the mineral soil (red line: lichens, green line: mosses).(b) Three phases of CO 2 efflux (in μmol CO 2 m À2 s À1 ) from chamber 1 (lichens).The hydration states of cryptogams and upper mineral soil is shown on top, where hydration of the soil profile is symbolised in bluish tones, water saturation in plain blue colour.

T A B L E 5
Regression coefficients of soil respiration (in μmol C m À2 s À1 ) volumetric water content (Θ in m 3 m À3 ) and temperature (T in C) of the mineral soil (0-5 cm) underlying the cryptogamic layer during the first wetted (31 August-04 September 2020) and dry-out (05-11 September 2020) phases.
photodegradation of plant litter.These findings refer to the heterotrophic microbial community in soils, and it must be assumed that autotrophic cryptogams contribute to rewetting pulses in addition to these heterotrophs.Based on the fact that lichens are a symbiosis of a heterotrophic fungus