Geodiversity impacts plant community structure in a semi-arid region

Geodiversity refers to the variety of geological and physical elements as well as to geomorphological processes of the earth surface. Heterogeneity of the physical environment has an impact on plant diversity. In recent years, the relations between geodiversity and biodiversity has gained attention in conservation biology, especially in the context of climate change. In this study, we assessed the spatial and temporal change in plant’s community structure in a semi-arid region, Sayeret Shaked Long Term Ecosystem Research (LTER) station, Israel. Vegetation surveys were conducted on different hillslopes, either with or without rock covers in order to study the spatial trends of hillslope geodiversity. The surveys were conducted for two consecutive years (2016 and 2017), of which the second year was drier and hotter and therefore permitted to investigate the temporal change of plant’s community structure. The results of the spatial trends show that (1) geodiversity increases vegetation biodiversity and promotes perennial plants and those of the temporal change show that (2) the positive effect of geodiversity on plants’ community structure and species richness is greater in the drier year than that in a wetter year. The main insight is that in these drylands, hillslopes with higher geodiversity appear to buffer the effect of drier years, and supported a more diverse plant community than lower geodiversity hillslopes.

www.nature.com/scientificreports/ stones' size distribution and cover percentage 20 . Linked to the stoniness on the ground surface, the distribution of the stones throughout the soil profile is an important component of geodiversity [21][22][23] . The presence or absence of rock fragments affect the distribution of soil-water throughout the profile and the ramification of roots, and have an impact on soil temperature [24][25][26] . Characterization of soil profile is therefore crucial for understanding the complex impact of stoniness. Heterogeneity of the physical environment, alongside with climatic variables, have a crucial effect on vegetation living conditions and biodiversity 5,27 . It was demonstrated that geodiversity affects the distribution of vegetation 28,29 , composition of soil microbes, and the resistance of plant to drought [30][31][32] . In drylands, understanding the relation between geodiversity-governed water distributions 21 and plants viability is highly important. Recent studies from the semi-arid north-western Negev proposed that heterogeneous land units, dominated with partially-embedded stones in their ground surface, increase the spatial redistribution of water, with the resultant increase in water availability for shrubs 32 , improving their durability to prolong drought events 21 . Specifically, mass mortality of Noaea mucronata (Forssk.) shrubs was reported for hillslopes with low geodiversity level 21,32 . Recent studies proposed evidence on the buffering effect of geodiverse locations on vegetation diversity in a changing environment 33,34 as geodiversity can improve adaptability of species and acts as microclimate refugia 35,36 . More empirical data from different climate zones are needed to support the geodiversity-biodiversity relations and to be implemented in conservation projects of natural ecosystem 6,37,38 .
The N. mucronata is a perennial shrub that dominates the north-western Negev (Fig. 2), which shows an adaptive response to dryland environments at numerous levels. The adaptation to water-limited environments shaped the plant's metabolism and morphology. Aboveground adaptions comprise changes in leaf morphology, where the small narrow winter leaves that emerge after the rain (approximately in November) shed or turn into thorns during the dry season, reducing water loss through transpiration 39 . The seasonal difference in leaf morphology result in a total plant size variation from summer (smaller) to winter (larger). The green stems transform into grayish fissured bark, which enables the growth of young branches once the wet season starts 39 . The physiological traits of adaptions include the C 4 carbon fixation pathway, which correlates with high efficiency in CO 2 fixation and low transpiration loss under high temperature conditions 40 . Moreover, at the root level, N. mucronata adapt in association with ectomycorrhizal fungi, which improves the plants' water and nutrient uptake 41,42 . The N. mucronata is considered as a landscape engineer [43][44][45] , which improves habitat conditions and facilitates herbaceous community in its surroundings by modifying microclimate (reduction of temperature stress due to the shading effect) and soil properties (increasing infiltration and reducing evapotranspiration) 43,[46][47][48] . At the same time, it was reported that the N. mucronata may impose some negative effects, such as the suppression of annual plants 49 through shading and competition 50 .
Our study investigates the spatial and temporal effect of geodiversity on plant community structure and on our model species N.mucronata. The importance of our results is its link with the climatically conditions and the novelty of the semi-arid region case. The study objectives were to assess during two compared year, a wetter and a drier one, the effect of geodiversity features such as stoniness and depth of the soil profile (1) on the physiological conditions of N. mucronata shrubs, and (2) on plant's community structure and diversity. Our overall hypotheses are that an increase of geodiversity will (1) improve the physiological conditions of N. mucronata shrubs and (2) enhance plant diversity.

Materials and methods
Regional settings. The study was conducted between February 2016 and May 2017 in the fenced area (no access for livestocks since 1990s) of the Sayeret Shaked Park (~ 20 ha) ( Fig. 1) with the permission of the Israel's LTER Network. The park is located in the semi-arid north-western Negev of Israel (31° 17′ N, 34° 37′ E; 200 m.a.s.l.), where annual average rainfall is ~ 165 and standard deviation of 58 mm/year 21 . The site's soil is aeolian loess, with sandy loamy to loamy sand texture 51 . The park is covered with dwarf shrubs (0.1-0.6 m tall) such as Atractylis serratuloides Cass., Thymelaea hirsuta (L.) Endl., and the dominant N. mucronata shrubs. The site has a large number of annuals, geophyte (e.g., Asphodelus ramosus L.) and hemicryptophyte species, as well as a rich community of biological crusts 50,[52][53][54][55] . In a previous study, a geomorphological survey followed by hillslopes mapping took place in Sayeret Shaked starting in July 2015 21 . The results showed the presence of either hillslopes without rock fragment cover and hillslopes with substantial (~ 30%) rock fragment cover (Fig. 2). Following the study, three low-geodiversity hillslopes, with a thick (> 1 m) and non-stony soil layer, without rock fragment cover or rock fragment content in the soil profile (homogeneous: HM; Fig. 2A), and three high-geodiversity hillslopes, with a thin (~ 0.1 m) stony soil layer, rock fragment cover (~ 30%) and rockfragment content in the soil profile (~ 35%) (heterogeneous: HT; Fig. 2B) were selected for the study. The mean (± SE) slope incline of all the hillslope is 5° (±) 22 . Distance between two adjacent hillslopes was at least 100 m. On each of these hillslopes, a 400 m 2 (20 × 20 m) plot was established for data collection.
To study the effect of the geodiversity features on plant community, we assessed the plant diversity at a once-a-month frequency over the growing season (Feb-May/June) of two sequential years (2016-2017). In each of these years, the last cycle of data collection took place when the annual plants were dried out, and at the point where the N. mucronata 's 'winter leaves' 56 have disappeared. Data of precipitation, solar radiation, soil temperature and soil moisture at different depths (25 and 50 cm) were obtained from the meteorological station located in the LTER site. Vegetation survey. The N. mucronata was the only perennial plant present in both HT and HM hillslope, making it the best-fit model species to assess the year-round differences in plant viability. In each plot (n = 6) we studied three individual shrubs, to a total of nine individuals per hillslopes type. In order to monitor the shrub's morphological changes during the year, we measured the maximum length of green branches. Also, we www.nature.com/scientificreports/ measured the plant's size by measuring the maximum length from a green part to another on a north-south and east-west axes, as well as the shrub height. We multiplied the three axes to calculate the maximum plant size. In addition to these nine N. mucronata plants, we sampled leaves (winter leaves only) from other nine randomly selected N. mucronata plants to estimate their physiological condition, through measuring the relative water content (RWC), membrane stability (EC), carbon-nitrogen ratio (C:N), and chlorophyll content.

Biochemical analysis.
Relative water content (RWC). We added 3-5 g of N. mucronata young leaves of each individual to a 50 ml vial with a wet tissue to maintain humidity. The leaves were weighted for fresh weight using Sartorius AG Göttingen CP225D, Germany. The samples were submerged in de-ionized water for 24 h and then weighted for turgor weight. Additionally, the samples were dried at 65 °C for 24 h in the oven for dry weight.
C:N ratio. Few leaves from each shrub of N. mucronata were collected for total organic carbon (C org ) and total nitrogen content (N tot ) analysis, dried at 65 °C for 12 h and manually ground by mortar and pistil. Of these samples, 20 mg were put in a C-N analyzer (CHNS Elemental Analyzer, Thermo Scientific, USA).

Membrane stability index (MSI).
From each N. mucronata, 20-30 leaves were collected and placed in 50 ml vials filled with 20 ml of double distilled water (DDW). The electrolyte's electric conductivity (EC) was measured with a probe (Eutech Instruments, CON 510, Singapore), as initial leakage (C i ). The samples were then placed on a laboratory shaker (TOS-4838PD, MRC Lab Instruments, Holon, Israel) for 12 h at 200 rpm and the EC was measured again as C r . The samples were then autoclaved to blast cell membrane, and the maximum conductivity (C m ) was measured. Membrane stability index (MSI) was calculated according to Eq. (1): Chlorophyll content. 10 leaves of N. mucronata were collected and add into a 2 ml Eppendorf with 1 ml of Dimethyl sulfoxide (DMSO), covered with aluminum foils. In the laboratory, vials were kept in a 65 °C incubator for 72 h and then centrifuged for 14,000 rpm at 20 °C for 10 min. 200 μl of the supernatant was added to a 96 wells plate and the absorption was read with Epoch™ spectrophotometer (BioTek Instruments, Inc., Vermont, USA). Chlorophyll a (Chl a), contents was calculated using Eq. (2) 57 : In order to determine the plant diversity in the different hillslope, we calculate species richness (n) as the number of species present at the site, and species abundance as the total number of plants present at the site (N). In order to determine the diversity in the communities, we calculate the Shannon Diversity Index (Eq. 3) and Shannon Evenness Index (Eq. 4) and the Simpson Index (Eq. 5), 58,59

Results
The two years of sampling were characterized by two different rain regimes (Fig. 3B) (Fig. 3C), as well as the reduction of soil moisture at both depths (Fig. 3D). We did not find significant differences in physiological measurements of N. mucronata plants between the two hillslope types throughout the two years of sampling (SI 2). For the phenological measurements, the estimated mean N. mucronata size was greater in HM hillslopes than that in HT hillslopes (Kruskal-Wallis one-way analysis of variance, p < 0.0001).
Nevertheless, once we plotted the data for different years (2016 vs 2017), we found a significant difference for most of the parameters between the two years, and especially regarding the C:N ratio and RWC. In 2016, N. mucronata had a higher C:N ratio (HT, Dunn's Method: p < 0.05, HM Dunn's Method: p < 0.05, Fig. 4A) and higher RWC (HT Holm-Sidak method: p < 0.001, HM Dunn's Method: p < 0.05, Fig. 4B) than these in 2017. For the C:N ratio, this trend was similar for HT and HM hillslopes (a decrease of 26% and 28% from 2016 to 2017 in HT and HM, respectively). At the same time, the decrease in RWC was much smaller in HT hillslopes (5%) than that in HM hillslopes (11%). Plant diversity. We found that plant diversity can be explained by hillslope type (Analysis of similarity: R = 0.36, p = 0.0001) (Fig. 5A,B). In addition, we measured the effect of sampling year (SI 6 A, B, C and D), which was also significant, though the value explained was lower than the hillslope type (R = 0.11, p = 0.002). Moreover, we found a significant difference in life form composition, considering all species found, between the hillslopes (Pearson chi-square: Chi Square: 56.7, df = 2, p < 0.0001). HT plots had a higher mean value of perennial (39% in HT compared to 3% in HM) as well as higher mean value of herbaceous perennial (13% in HT compared to 7.5% in HM) (Fig. 5A,B). Specifically, in HM we found two shrubby species (Pituranthos tortnousus and N. mucronata), and a total of 3 individuals, while in the HT we found six shrubby species (N. mucronata, A. articulata, Dianthus monadelphus subsp. judaicus, Salvia lanigera Poir., Helianthemum stipulatum (Forssk.) C. Chr. and Helianthemum lippii (L.) Dum. Courset) and a total of 47 individuals. At the same time, the mean value of annuals was significantly higher in HM (90% in HM compared to 48% in HT). The annual plant community in HM encompassed of 25 species and 110 individuals, while in the HT it encompassed 20 species and 58 individuals.
The accumulated cover of four species contributed to 52.8% of the differences between the two hillslope types. These species included Stipa capensis Thunb. (19.6%), N. mucronata (18.8%), Anabasis articulata (Forssk.) Moq. (8.6%), and Onobrychista crista-gali (L.) Lam. (5.8%) (Similarity percentage analysis, SI 5). The main differences were caused by two perennial plants that were absent from the HM transects: N. mucronata and A. articulata. Additionally, once we plotted the annuals' diversity data per year (SI 6 A-D), we found that the annual community structure in the HT hillslopes during the shift from 2016 to 2017 became more even, with a more uniform abundance of species (SI 6 A and B). An opposite trend was found for the HM hillslopes, where an increase in abundance of dominant species was observed (SI 6 C and D).
Species richness and abundance are reported in Table 1   To gain more insights about the community structure, we calculated the Shannon Diversity Index and Simpson Diversity Index (Table 1). Overall, HT hillslope had a higher diversity (Shannon Diversity Index in HT was 3.5% higher than that in HM). Further, once we calculated the yearly data, we found that in 2017 the drop of diversity in HM hillslopes was significant, whilst a significant increase was observed for HT hillslopes. Moreover, in 2017, the Shannon Evenness Index showed an increase in HT hillslopes and a decrease in HM hillslopes. This suggests that in HM hillslopes, the functional group encompass few dominant species with high abundance and few sparser species with low abundance. In HT hillslopes, the community structure was more even.

Discussion
The aims of our study were to assess the spatial effect of geodiversity, express as stoniness and depth of the soil profile in different hillslopes, on the physiological conditions of N. mucronata shrubs and plant's community structure and diversity. We showed that the geodiversity features promote a more diverse plant community (Fig. 5A,B). During the two years monitoring (2016-2017) the second year happened to be significant drier with repercussion on soil temperature and moisture at different depth (Fig. 3). This fact allowed us to have insights about the temporal effect of geodiversity features to mitigate drier years. In fact, annual communities on hillslopes with lower geodiversity tend to suffer a decrease in diversity during the drier year (SI 6 and Table 1). Previous studies from the semi-arid north-western Negev suggested that hillslope-scale geodiversity improves the source-sink relations 29 and positively affect soil quality and geo-ecosystem functions. Specifically, it was reported that the HT hillslopes had 22% greater mean hygroscopic moisture 22 then that in HM hillslopes. This result can indirectly indicates a similar availability of soil water for plants. The comparatively favorable conditions in the HT hillslopes were proposed to be the key of survival of N. mucronata plants during the mass mortality that occurred during different drought events started in 1999 32 . Geodiversity components such as geomorphology, topography, geology, and hydrology are associated with energy and nutrients, which regulate biodiversity 61,62 . Nevertheless, only recently, the impact of geodiversity on biodiversity has gained attention [63][64][65][66] . Considering global climatic change, it was proposed that high-geodiversity land units are potentially more capable to support biodiversity because of their intrinsic resilience 65,[67][68][69] . In our study, the favorable soil conditions in HT hillslopes were not straightforward translated into a better physiological state of plants (SI 2), as we expected in our first hypothesis. Our biochemical analysis did not show a prominent effect of geodiversity features over the physiological conditions of N.mucronata, when we compared the plots (HT vs HM) (SI 2). Despite that, during 2017, the shrubs in HM hillslopes faced stronger water stress compared to these in HT hillslopes (Fig. 4B). The positive effect of increased geodiversity on hygroscopic moisture 22 , and the overall higher soil-water content in HT hillslopes 70 , might explain the reduced water stress for shrubs in HT hillslopes 32,71 .
The shrub patch is the driver of the cyclic succession of plants community 48,[72][73][74] . According to the biodiversity cyclic hypothesis 48 , once the patch is consolidated, the dissimilarity in community structure between the shrubby patches and interpatch spaces increases. Therefore, shrublands patchiness is critical for sustaining spatial heterogeneity 75 . In our study region, the HT hillslopes have higher patchiness, mainly due to higher geodiversity that supports shrub durability to droughts. Regarding our second objective on the effect of geodiversity on the plant community structure, we found that the perennials in HM were 2% of the community structure, whilst being 38% in HT (Fig. 5A,B). In HM, the annuals contributed 90% of the total community structure, while in HT they contributed 48% (Fig. 5A,B). The high cover of annuals in the HM hillslopes compared to the HT hillslope displays one possible response of the ecosystem to prolonged droughts that is a transition from shrubland to grassland. This transition did not occur in the HT hillslopes due to the effect of geodiversity that increase their resistance to drought. The ecosystem in the HM does not collapse to the bare soil as predicted by many mathematical models of vegetation patterns in drylands that include only wood vegetation but it is actually shifted to another stable state 76,77 .
The annual community structure faced a shift from the wetter 2016 to the dryer 2017 in both hillslopes (SI 6 A-D). In HT hillslopes, the annuals community become more even, having fewer dominant species with high abundance (SI 6 A and B). At the same time, in HM hillslopes, the annuals' community structure became less even (SI 6 C and D). In our diversity analysis, we showed that in the drier year, plant diversity increased in HT hillslopes (Table 1) and decreased in HM hillslopes. Different simulation studies, where plant communities underwent abiotic stresses related to climate changes-such as higher temperature or lower precipitations-showed similar results 78,79 . It was shown that community response to changes is rapid (two seasons), and that stresses can change patterns of plant dominance and evenness 78,79 . Changes in dominance in community composition can have consequences for species coexistence and ecosystem functions 80 . Recent studies show that locations with high geodiversity have the potential role to act as hotspot for biodiversity and to maintain community's structure due to their function to buffer climatic changes 66,81 . In our study, we show that the effect of geodiversity features on community structure and species richness is greater in the drier year than that in a wetter year. Our results suggested that hillslopes with higher geodiversity might be in situ refugia that facilitate the existing community and in shifting climate periods might serve as biodiversity spots [82][83][84][85][86] . www.nature.com/scientificreports/ Our results aligned with the findings of recent studies that suggest that integrating geodiversity in conservation nature projects becomes a necessity to protect ecosystems and their services 10,65,87 and offer the opportunity to broad the data availability. In fact the connection between biodiversity and geodiversity allow to use geodiversity measurements, more spatial consistent and therefore available at broader scale, and relating them to biodiversity 64,88 .

Conclusion
The motivation of this study was to understand how different geodiversity features, expressed by the degree of stoniness and soil thickness, affect the physiological state of N. mucronata and the plant life form variability. In conclusion, our data show that in a semi-arid regions, hillslopes with higher geodiversity better buffer the effect of drier years, and supported a more diverse plant community compared to lower geodiversity hillslopes. The potential role of sites with higher geodiversity to mitigate climate change effects in terms of persistence of biodiversity should be take into consideration when planning for conservation actions and ecosystem management 84,89,90 In this context, additional studies should be conducted in other drylands of the world in order to verify the mechanisms through which geodiversity regulates the structure and composition of vegetation community.