Current Condition of Pannonic Salt Steppes at Their Distribution Limit: What Do Indicator Species Reveal about Habitat Quality?

Little is known about the suite of ecological conditions under which characteristic species may continue to develop under the pressure of recent habitat deterioration. We aimed to determine the niche of three indicator species of the priority habitat Pannonic salt steppes and to find out how their vegetation composition, land use, and soil chemistry mirror the current condition of their typical habitat. A plot-based vegetation survey was conducted in degraded and in pristine (reference) inland salt steppes in East-Central Europe. We confirmed decreased habitat quality at their northern geographical limit. Most of the sites there showed a strong prevalence of generalists (e.g., Elytrigia repens) and lack of specialists, both resulting from lowered habitat extremity and inappropriate land use (abandonment). A small proportion of plots (19%) were in the same good condition as the reference vegetation in the central area. Soil analyses revealed that the studied halophytes are able to persist on desalinized soils if the land use is suitable. The occurrence of the annual Camphorosma annua (Amaranthaceae) was driven largely by abiotic stress; grazing alone is insufficient for its long-term persistence, while the perennial Artemisia santonicum (Asteraceae) and Tripolium pannonicum (Asteraceae) have higher survival chances as they are able to coexist with generalists. Overall habitat quality can be reliably determined from the analyzed ecological conditions of indicator species. The outcomes of the presented work are relevant for conservation practice and can serve as a quick tool for assessing the current stage of other grassland habitats.


Introduction
Natural habitats have recently been subjected to species turnover due to negative vegetation changes when the original, highly specialized species pool has been substituted by generalists with a lower naturalness value [1]. Non-forest habitats with extreme edaphic conditions suitable for specialists are disappearing [2]. Halophytic vegetation also belongs to a group of highly threatened habitats hosting endangered plant and animal species [3]. They develop on salt-affected soils in large areas with a semiarid-subhumid climate [4], in Europe for instance on the East European Plain and the Iberian Peninsula [5]. In temperate zones, inland saline habitats occur as discrete islands, for example, in the North German Plain [6], while their main center of distribution is the Pannonian Basin [7]. A. Podunajská nížina lowland-located in southwestern Slovakia, it is the northern enclave of Pannonic inland salt steppes with typical stands of the alliances Puccinellion limosae and Festucion pseudovinae [27], however, with the largest habitat loss and vegetation degradation. In the 1960s, the total area was estimated to be 8300 ha; recently, only 500 ha was reported [28]. The groundwater level is 1-5 m http://apl.geology.sk/gibges/accessed (accessed on 1 March 2021), and it has decreased throughout the area due to former water regulations resulting also in alteration of rare Solonchak soils to the Solonetz type. Land reclamations and abandonment beginning in the Communist era resulted in fragmentation of the habitat 1530 [29]. Only scattered locations remained in the prevailing agricultural landscape, and each has been declared as a protected area. Moderate grazing (cattle, grey cattle, and sheep) was re-established in most locations just recently.
B. Hortobágy-located in eastern Hungary as part of the Great Hungarian Plain. Meadow Solonetz soils are developed over loess with a salt accumulation zone in deeper soil layers [30]. Groundwater is rich in sodium bicarbonate (NaHCO 3 ), whereas the groundwater table is relatively high (usually 0.5-2.5 m). The prevailing vegetation types are alkali steppes of Festucion pseudovinae which have dominated the area since the late Pleistocene. Although river regulations in the second half of the 19th century decreased regular flooding and drainage works during the 20th century dried out the marshy depressions [7], the area is now a national park harboring diverse sodic steppes with high degree of continuality. They are known as one of the best-preserved grassland habitats in Europe thanks to traditional cattle and sheep grazing. C. Kiskunság-located in central Hungary in the Danube-Tisza Interfluve (part of the Great Hungarian Plain), the second largest saline habitat complex in the Pannonian Basin ranks first in variability, as all types of saline/alkali habitats are represented in local variations [31]. Until river regulations began at the end of the 19th century, the Danube regularly flooded its 20-30-km-wide flood plain, which enabled the formation of large sodic lakes of high natural importance [32]. Typical soils are calcareous Solonchak and Solonchak-Solonetz, covered mainly by the annual vegetation of Therosalicornietea and Crypsietea aculeatae, while perennial halophytic stands of Puccinellion limosae, Festucion pseudovinae, and Scorzonero-Juncion gerardii are also well-developed. Today, the area is a national park with dominating rangelands managed by cattle and sheep grazing.
Each of these three regions has a moderately continental climate, with hot and dry summers and mild winters. The annual precipitation is 500-550 mm and the annual average temperature is 10.5 • C [33].

Selection of Study Objects
The target habitat Pannonic salt steppes is a mosaic of several grassland formations reflecting different soil salinity and water regime depending on the micro-elevation of the terrain [7,34]. Three typical halophytes (flagship species for conservation) were chosen to represent the target habitat: Artemisia santonicum subsp. patens K. M. Perss (later on A. santonicum), Asteraceae, is a perennial species of the Eurasian continental steppes on sodic Solonetz/sodic soils. From an ecological viewpoint, its vegetation optimum is in the association Artemiso-Festucetum pseudovinae (alliance Festucion pseudovinae) where it is a diagnostic, constant and characteristic taxon [22,25,27]. A. santonicum is a Red List species, and on its northernmost distribution limit (Slovakia, Austria) is evaluated in the endangered category of the International Union for Conservation of Nature and Natural Resources (IUCN) [35,36]. This species indicates habitat Artemisia steppes within target habitat Pannonic salt steppes [37].
Camphorosma annua Pall. (later on C. annua), Amaranthaceae, is an annual species forming initial monodominant stands in shallow depressions of the natural Solonetz habitats. It occupies barren spots with the highest soil salinity within the salt steppes zonation [25,38]. In Slovakia, on its northernmost occurrence, it is critically endangered according to the IUCN Red List [36], and on its western distribution border (Austria and Croatia) it is endangered [35,39]. It is the indicator species of the annual salt pioneer swards (association Camphorosmetum annuae, alliance Puccinellion limosae) within the target habitat [37].
Tripolium pannonicum (Jacq.) Dobrocz. subsp. pannonicum (later on T. pannonicum), Asteraceae, is an annual, biennial, or shortly perennial species of shallow waterlogged depressions on saline and sodic soils. The main species T. pannonicum is very variable. Two subspecies were described: T. pannonicum subsp. pannonicum and T. pannonicum subsp. tripolium. In Europe, the nominate subspecies occurs in Spain, Germany, the Pannonian Basin, and south eastern Europe while subsp. tripolium is reported along most of the sea coasts from Europe and Asia [40]. It is diagnostic, characteristic, and constant species of the association Puccinellietum limosae [27]. It has the widest coeno-ecological preferences among the indicator species, occurring in Solonchak salt marshes of the order Scorzonero-Juncion geradii, in the littoral zone of sodic lakes of the class Crypsieta aculeatae [38] and in brackish marshes of the class Phragmito-Magnocaricetea [37]. In Slovakia it grows mostly in the vegetation of Puccinellion limosae. It is categorised as endangered in Slovakia [36] and Austria [35], while it is critically endangered in the Czech Republic [41]. This species indicates dense and tall Puccinellia swards [37] of the target habitat.
These obligate halophytes were selected as flagship species having high indication value of natural salty areas compared to other halophytes in the Pannonian Basin and are relatively common in the study area. C. annua and A. santonicum are considered as diagnostic species, together with T. pannonicum as constant species of Pannonic salt steppes [42]. None of them are spreading along roadsides benefiting from road salting as are Puccinellia distans agg. or Spergularia salina [43,44].

Vegetation Sampling
The field survey was conducted between 2016 and 2018 in the second half of the growing season. To determine the ecological range limits of the indicator species of Pannonic salt steppes (where they are able to grow from optimal to suboptimal conditions in their target habitat), we conducted a plot-based vegetation survey. Data on fine scale vegetation composition, land use type (grazing, mowing, or abandonment-obtained from local experts), and soil properties were collected.
Plots with a uniform size of 0.5 × 0.5 m were sampled altogether on 35 different sites throughout the Pannonian Basin; on the norther distribution border of salt steppes in Slovakia (Podunajská nížina) and in their central distribution in Hungary (Hortobágy and Kiskunság). The plots were positioned within homogenous areas bearing the defined indicator species. On the Slovak sites, we recorded the plots in the whole range of habitat naturalness from preserved to degraded stages (for A. santonicum 41 plots, for C. annua 26 plots, and for T. pannonicum 34 plots); on the Hungarian sites, we sampled plots of pristine Pannonic salt steppes serving as reference (10 plots for A. santonicum and 13 plots for C. annua in Hortobágy, 10 plots for T. pannonicum in Kiskunság). Altogether 134 plots were recorded and analyzed.
The recording was conducted on two scales. The percentage cover of vascular plants was estimated on the whole plot area (0.5 × 0.5 m). For frequency data, the plot was divided into 25 subplots (size 0.1 × 0.1 m) using a wire (see photographs in Supplementary Material S2) where the presence/absence data of vascular plants were recorded. Plot selection was semi-random due to the characteristic patchiness of inland halophytic habitats. To avoid vegetation heterogeneity, we placed the sampling wire in homogenous stands of the target vegetation types of the studied species, reflecting the characteristic species composition according to the actual habitat classification of inland saline vegetation of the Pannonian Basin [37].

Soil Sampling
At the time of vegetation recording, one soil sample was taken per plot at a depth of 0-12 cm. Altogether 77 samples were collected, 25 in the vegetation with A. santonicum (15 from Slovakia, 10 from Hortobágy), 21 in C. annua (11 from Slovakia, 10 from Hortobágy), and 31 in T. pannonicum (21 from Slovakia, 10 from Kiskunság).
They were subjected to laboratory analysis of the following properties: EC (electric conductivity), SAR (sodium adsorption ratio), Na_exch (exchangeable sodium), pH, Cox (organic carbon), and Ntot (total nitrogen). The pH and EC were determined by the method of Sotáková et al. [45]. A saturated soil extract was obtained by centrifugation, and pH and conductivity were determined in the supernatant using a pH electrode and conductometer, respectively. Exchangeable cations (Na_exch) were determined according to ISO 13536. The exchangeable cations were displaced by barium chloride buffered at pH 8.1. The concentration of displaced cations was determined by F-AAS. For calculating SAR, concentrations of Ca 2+ , Mg 2+ , and Na + (determined by F-AAS in the saturated soil extract) were used. Simultaneous determination of carbon and nitrogen (Cox and Ntot) was carried out by dry combustion. First, soil samples were dispensed into the combustion tube. The result of combustion is a mixture of gases, nitrogen oxides, CO 2 , H 2 O, SO 2 , and the excess oxygen. The mixture of gases was passed through the reduction catalyst, which eliminates redundant oxygen and reduces the nitrogen oxides to nitrogen alone. This mixture was split on a thin-layer chromatography column and analyzed by a thermal conductivity detector. This method also determines levels of inorganic carbon originating from CaCO 3 , NaHCO 3 , etc. In order to exclusively measure Cox concentration, inorganic carbon must be removed by decalcification using a dilute HCl solution. The analytical database (Supplementary Material S3) was checked for consistency according to the suggestions of Richards [46].

Data Processing
The plot dataset of 134 relevés (cover scale) and the subplot dataset of 3350 relevés (presence/absence) were stored in the TURBOVEG database [47] and exported for analysis in the JUICE program [48]. The recorded species were divided into characteristic species (including indicator species), accompanying halophytic species, and generalists according to the list of halophytes compiled by V. Krist [49] from the northern part of the Pannonian Basin. Details of classification criteria are contained in Tables S1-S3 in Supplementary Material S1. Characteristic species have a percentage frequency in reference plots of at least 5%. For this purpose, the scale of 0.01 m 2 was used. Due to the higher number of records, this scale captures the typical species composition more exactly and suppresses random occurrences of species with low cover (frequency below 5%), which would otherwise be recorded at the 0.25 m 2 scale (for details, see Figures S1-S3 in Supplementary Material S1).
The recorded plots were divided into five groups of naturalness, based on the species composition of the three surveyed plant communities of Pannonic salt steppes (Artemiso-Festucetum pseudovinae, Camphorosmetum annuae, and Puccinellietum limosae) represented by the three indicator species. Classification criteria are shown in Table 1. Table 1. Classification criteria into five groups of naturalness based on the proportion of characteristic, accompanying halophytic and generalist species (for supporting data see S1). cording to the list of halophytes compiled by V. Krist [49] from the northern part of the Pannonian Basin. Details of classification criteria are contained in Tables S1-S3 in Supplementary Material S1. Characteristic species have a percentage frequency in reference plots of at least 5%. For this purpose, the scale of 0.01 m 2 was used. Due to the higher number of records, this scale captures the typical species composition more exactly and suppresses random occurrences of species with low cover (frequency below 5%), which would otherwise be recorded at the 0.25 m 2 scale (for details, see Figures S1-S3 in Supplementary Material S1). The recorded plots were divided into five groups of naturalness, based on the species composition of the three surveyed plant communities of Pannonic salt steppes (Artemiso-Festucetum pseudovinae, Camphorosmetum annuae, and Puccinellietum limosae) represented by the three indicator species. Classification criteria are shown in Table 1.
Group 1: vegetation of the highest naturalness, typical association of the indicator species, high representation of specialists, no occurrence of species indicating degradation (generalists).
Group 2: vegetation of high naturalness, typical association of the indicator species, only random occurrence of generalists.
Group 3: vegetation with decreased naturalness and minor signs of degradation, not typical association of the indicator species, low proportion of generalists.
Group 4: degraded vegetation of the association of the indicator species with high representation of generalists.
Group 5: strongly altered vegetation not assignable to any target association, generalists predominate. Table 1. Classification criteria into five groups of naturalness based on the proportion of characteristic, accompanying halophytic and generalist species (for supporting data see S1).

Data Analysis
For each plot with a percentage cover scale, indicator values for light, temperature, moisture, soil reaction, and nutrients were calculated and weighted by species cover. Taking into account the strong representation of Pannonian species, indicator values Generalists indicating degradation 0% >0% to <5% >0% to <25% (>0%*) ≥25% to <50%
Group 1: vegetation of the highest naturalness, typical association of the indicator species, high representation of specialists, no occurrence of species indicating degradation (generalists).
Group 2: vegetation of high naturalness, typical association of the indicator species, only random occurrence of generalists.
Group 3: vegetation with decreased naturalness and minor signs of degradation, not typical association of the indicator species, low proportion of generalists.
Group 4: degraded vegetation of the association of the indicator species with high representation of generalists.
Group 5: strongly altered vegetation not assignable to any target association, generalists predominate.

Data Analysis
For each plot with a percentage cover scale, indicator values for light, temperature, moisture, soil reaction, and nutrients were calculated and weighted by species cover. Taking into account the strong representation of Pannonian species, indicator values appropriate for East-Central Europe were used by Borhidi [50]. Ecological indication values express optimum ecological niche of plant species on the gradient of nutrients, moisture, soil reaction, continentality, temperature, light, and salinity on the ordinal scale (mostly 1-9, possibly 1-12 for humidity). They are often used to rapidly estimate site conditions from species composition when measured values of environmental variables are not available. Values are based on field observations, and partly on evidence from ecological experiments and measurements of environmental variables [51]. Ecological indication values of species  tc. In order to exclusively measure Cox concentraved by decalcification using a dilute HCl solution. tary Material S3) was checked for consistency acs [46]. cover scale) and the subplot dataset of 3350 relevés TURBOVEG database [47] and exported for analycorded species were divided into characteristic specompanying halophytic species, and generalists acpiled by V. Krist [49] from the northern part of the ation criteria are contained in Tables S1-S3 in Supic species have a percentage frequency in reference e, the scale of 0.01 m 2 was used. Due to the higher s the typical species composition more exactly and ecies with low cover (frequency below 5%), which 0.25 m 2 scale (for details, see Figures S1-S3 in Supd into five groups of naturalness, based on the speed plant communities of Pannonic salt steppes (Arorosmetum annuae, and Puccinellietum limosae) reps. Classification criteria are shown in Table 1. est naturalness, typical association of the indicator alists, no occurrence of species indicating degradaralness, typical association of the indicator species, s. sed naturalness and minor signs of degradation, not ecies, low proportion of generalists. f the association of the indicator species with high tion not assignable to any target association, gener-   originating from CaCO3, NaHCO3, etc. In order to exclusively measure Cox concentration, inorganic carbon must be removed by decalcification using a dilute HCl solution.
The analytical database (Supplementary Material S3) was checked for consistency according to the suggestions of Richards [46].

Data Processing
The plot dataset of 134 relevés (cover scale) and the subplot dataset of 3350 relevés (presence/absence) were stored in the TURBOVEG database [47] and exported for analysis in the JUICE program [48]. The recorded species were divided into characteristic species (including indicator species), accompanying halophytic species, and generalists according to the list of halophytes compiled by V. Krist [49] from the northern part of the Pannonian Basin. Details of classification criteria are contained in Tables S1-S3 in Supplementary Material S1. Characteristic species have a percentage frequency in reference plots of at least 5%. For this purpose, the scale of 0.01 m 2 was used. Due to the higher number of records, this scale captures the typical species composition more exactly and suppresses random occurrences of species with low cover (frequency below 5%), which would otherwise be recorded at the 0.25 m 2 scale (for details, see Figures S1-S3 in Supplementary Material S1).
The recorded plots were divided into five groups of naturalness, based on the species composition of the three surveyed plant communities of Pannonic salt steppes (Artemiso-Festucetum pseudovinae, Camphorosmetum annuae, and Puccinellietum limosae) represented by the three indicator species. Classification criteria are shown in Table 1.
Group 1: vegetation of the highest naturalness, typical association of the indicator species, high representation of specialists, no occurrence of species indicating degradation (generalists).
Group 2: vegetation of high naturalness, typical association of the indicator species, only random occurrence of generalists.
Group 3: vegetation with decreased naturalness and minor signs of degradation, not typical association of the indicator species, low proportion of generalists.
Group 4: degraded vegetation of the association of the indicator species with high representation of generalists.
Group 5: strongly altered vegetation not assignable to any target association, generalists predominate. riteria into five groups of naturalness based on the proportion of characteristic, accompanying st species (for supporting data see S1).

m Plot
Group No. and Its Symbol in Figures 2a, 3a    originating from CaCO3, NaHCO3, etc. In order to exclusively measure Cox concentration, inorganic carbon must be removed by decalcification using a dilute HCl solution.
The analytical database (Supplementary Material S3) was checked for consistency according to the suggestions of Richards [46].

Data Processing
The plot dataset of 134 relevés (cover scale) and the subplot dataset of 3350 relevés (presence/absence) were stored in the TURBOVEG database [47] and exported for analysis in the JUICE program [48]. The recorded species were divided into characteristic species (including indicator species), accompanying halophytic species, and generalists according to the list of halophytes compiled by V. Krist [49] from the northern part of the Pannonian Basin. Details of classification criteria are contained in Tables S1-S3 in Supplementary Material S1. Characteristic species have a percentage frequency in reference plots of at least 5%. For this purpose, the scale of 0.01 m 2 was used. Due to the higher number of records, this scale captures the typical species composition more exactly and suppresses random occurrences of species with low cover (frequency below 5%), which would otherwise be recorded at the 0.25 m 2 scale (for details, see Figures S1-S3 in Supplementary Material S1).
The recorded plots were divided into five groups of naturalness, based on the species composition of the three surveyed plant communities of Pannonic salt steppes (Artemiso-Festucetum pseudovinae, Camphorosmetum annuae, and Puccinellietum limosae) represented by the three indicator species. Classification criteria are shown in Table 1.
Group 1: vegetation of the highest naturalness, typical association of the indicator species, high representation of specialists, no occurrence of species indicating degradation (generalists).
Group 2: vegetation of high naturalness, typical association of the indicator species, only random occurrence of generalists.
Group 3: vegetation with decreased naturalness and minor signs of degradation, not typical association of the indicator species, low proportion of generalists.
Group 4: degraded vegetation of the association of the indicator species with high representation of generalists.
Group 5: strongly altered vegetation not assignable to any target association, generalists predominate. sification criteria into five groups of naturalness based on the proportion of characteristic, accompanying d generalist species (for supporting data see S1).

Data Analysis
For each plot with a percentage cover scale, indicator values for light, temperature, moisture, soil reaction, and nutrients were calculated and weighted by species cover. The species number and land use type in the groups of naturalness and analyzed soil properties were displayed by boxplots, for species number and land use with connected median points; all were generated in the program Statistica 7 [52]. The type of land use (surveyed at the time of vegetation sampling) was scaled according to its strength of positive effect on the vegetation based on general studies about management of seminatural grasslands [17,53]: 3-extensive grazing by grey cattle/sheep/mixed herd (land use type at moderate intensity, positive effect), 2-one-year mowing (low intensity, mixed effect), 1-abandonment (zero intensity, negative effect). Statistically significant differences among groups displayed in boxplots were tested using ANOVA and for species number and land use also Tukey post-hoc test (α = 0.05).
For the ordination (DCA and data attribute plots with the LOESS visualization regression modelling method), we used the program CANOCO for Windows 4.5 [54]. The map was created using the program QGIS, version 3.2 [55] with the QuickMapServices plugin and a terrain background layer from Stamen Design with data by OpenStreetMap. In group 1 the average species number was three; in the rest of plots there were eight (Figure 2b, left). The highest number of species was found in groups 2, 3, and 4; in the most degraded stands (group 5) the expansion of generalists hindered the occurrence of other species. The difference in absolute species number in the case of ART was outstanding: 68 species were recorded on the Slovak plots and 7 in Hortobágy (Table S1 in Supplementary Material S1). In stands of reduced naturalness (groups 2-5), the most frequent generalists were Elytrigia repens (30%), Poa angustifolia (26%), then Achillea millefolium, Inula britannica, and Bromus hordeaceus.

Ecological Range Limits of Indicator Species of Pannonic Salt Steppes-Vegetation Composition and Land Use Approach
Regarding land use (Figure 2b, right), the vegetation degradation was related to poor management. Most of the group 5 plots were recorded on abandoned sites, and only a few of them were grazed (Supplementary Material S3).
A substantial aspect in Artemisia steppes is the proportion of the two main association species (A. santonicum and Festuca pseudovina). Preserved sodic steppes are characterized by high cover of F. pseudovina, generally 40-70% [56]. While the percentage cover of A. santonicum in both areas was similar (65% versus 64%, see Table S1 in Supplementary Material S1), F. pseudovina showed great differences: in Hungary 97% and in Slovakia 55%. This suggests that F. pseudovina is suppressed by expansive grasses in degraded salt steppes. A. santonicum persists in unfavorable conditions longer than F. pseudovina and the former vegetation of Artemisia steppes is indicated only by the remnant populations of A. santonicum, and the characteristic subordinated species are missing. Plots with lower naturalness revealed that the survival chance of this species is high. During our field survey one-meter high specimens of A. santonicum have been observed occupying a 40-year-old poplar plantation without any characteristic salt steppe species (site no. 11 in Supplementary Material S3). A. santonicum turns out to be a stress tolerant competitor (C-S strategy) [57]: on strongly salt-affected soils it uses mechanisms of resistance to high salt content, while on weaker saline soils its perennial life form (hemicryptophyte) as well as absolute height of individuals is successfully used in competition with glycophytes. It tends to form pure stands on soils with high salinity after destruction of the vegetation cover (e.g., former ploughing). Due to the deep rooting its great regeneration ability was confirmed also by Galvánek et al. [58]. After all, the recent populations' vitality did not correspond with the habitat quality. We recorded the most vital populations of the species in groups 2, 3, and 4. These are previously plowed (about 20 years ago) and subsequently abandoned salt steppes. High abundance of the indicator species cannot be associated with good habitat condition.

Camphorosma Annua (CAM)
Nine plots (from the 26) of Slovakia had the highest naturalness similar to the 13 reference ones sampled in Hortobágy (Supplementary Material S3), which are open, extremely dry, nutrient-poor soils with high soil reaction and active land use (grazing). Those plots are located on the left side of the DCA (Figure 3a). In the lower part of the ordination space are plots with higher moisture; these are regarded as intact vegetation of Puccinellietum limosae. In the right section are plots of drier habitats (transition to Artemiso-Festucetum pseudovinae) which suggest that plots of naturalness groups 4 and 5 cannot be regarded as association of Camphorosmetum annuae.
In reference plots, C. annua had altogether four accompanying species, mostly Plantago tenuiflora and Matricaria chamomilla (Table S2 in Supplementary Material S1). They are typical association species with similar ecological preferences as C. annua [27], but in plots of group 1 in Slovakia were not recorded at all. On the DCA (Figure 3a) these plots do not overlap with the Hungarian reference ones. The perennial halophytes in groups 2-5 (Puccinellia distans agg. 63%, A. santonicum 21%, and F. pseudovina 21%, see Table S2 in Supplementary Material S1) indicate the succession of the pioneer association towards closed stands of Artemisia steppes as a result of decreasing habitat extremity. It is also associated with the presence of generalists; 10 different species were recorded, most frequently Poa angustifolia and Cynodon dactylon. The absolute species number was 17. The lowest mean species number (three) was detected in stands of group 1; the highest (six) was detected in group 5 (Figure 3b, left), which gives a different pattern as in the case of ART (Figure 2b, left). Here species number is increasing with decreasing naturalness and the most degraded plots are not colonized by a certain generalist species.
Habitat quality is associated with the intensity of land use, and plots with the highest naturalness were all grazed (Figure 3b, right); however, in the case of CAM, specific ecological conditions were more crucial. As the influence of land use was not significant (p > 0.05), abiotic stress (high soil salinity and an unstable water regime) acted here as a stabilizing mechanism, maintaining the survival of this species. Stands with C. annua have managed to persist through long-term abandonment in Slovakia due to extreme edaphic conditions. Sites where C. annua stands occur refer to an overall higher habitat quality. The absence suggests worsened habitat quality of Pannonic salt steppes, since a certain vegetation zone of salt steppes indicates the most extreme habitats are missing.
Long term survival chances of C. annua are therefore limited compared to A. santonicum due to the higher preference for habitat extremity and very low competition ability. A higher salt demand by halophytes results in lower competitiveness in non-saline environments as it was formerly described by Ungar [59] and observed also by Grigore et al. [60]. Species with high bioindicator potential are only able to survive in habitats with narrowly defined physicochemical characteristics [61].

Tripolium pannonicum Subsp. pannonicum (TRI)
Seven plots (from the 34) in Slovakia were classified as having the highest naturalness and there was a partial overlap with the 10 reference plots recorded in Kiskunság (Hungary). Land use vector on Figure 4a points in the direction of the group-1 plots. On the Hungarian plots, six species were recorded (Table S3 in Supplementary Material S1) each typical for the target association Puccinellietum limosae. High frequency was observed for T. pannonicum and P. distans agg. (99.6% and 89.6%). On the Slovak plots, T. pannonicum had only 68% frequency and P. distans 35%; moreover, 47 species were found altogether, with a strong prevalence of generalists (Table S3 in Supplementary Material S1). Mean species number increased significantly in vegetation of lower naturalness (Figure 4b, left). Similar to that observed for ART, the maximum species number was found in moderately degraded vegetation; on the most degraded plots it decreased due to expansion of generalists like E. repens (26%), P. angustifolia (15%), and, to a lesser extent, Agrostis stolonifera and I. britannica (Table S3 in Supplementary Material S1).
The relation between naturalness and land use is the same for TRI as for ART and CAM: reduced management intensity resulted in greater degradation of the vegetation of the studied species (Figure 4b, right). In Slovakia, the vegetation most similar to the reference plots (high cover of characteristic species and low species number) is of secondary origin, regenerated after shallow tillage of flat waterlogged depressions. Although abandoned, the soil disturbance has inhibited secondary succession. In the case of TRI, the most vital populations were also found on ploughed plots, similar to ART. The species has good competition ability on disturbed saline soils although its dominance is only temporary as was also observed by Melečková et al. [62]. The characteristic subordinated species typical in reference area Kiskunság were missing there.
In spite of the statistically insignificant difference between the five groups of naturalness from the aspect of land use, Figure 2b

Soil Chemistry of Indicator Species of Pannonic Salt Steppes (ART, CAM, TRI)
Soil samples of vegetation with Camphorosma annua had the most extreme edaphic conditions: the highest measured EC, SAR, Na_exch, and pH which indicate strongly sodic soils ( Figure 5). Out of the total 77 soil samples, 40 had EC higher than 4 mS/cm and pH higher than 8.5, which are classified as sodic and saline soils according to [63]. From there, 18 samples were detected in the C. annua plots (groups of naturalness 1 to 4), 18 samples were detected in the plots of Tripolium pannonicum in each group of naturalness, and four samples were detected in the plots of Artemisia santonicum (groups of naturalness 1 to 3) (Supplementary Material S3). It follows that the lowest sodicity had plots of ART (Figures 5 and 6). It refers to less extreme edaphic conditions in the Artemisia steppe zone which is in accordance with the general knowledge about the ecology and zonation of Pannonic salt habitats [7]. The highest nutrient content (Ntot and Cox) was detected in ART and TRI, the lowest in CAM (Figures 5 and 6), which is related to limited nutrient availability in saline soils [64]. Soils of TRI plots had a high variance of Ntot and Cox, and the species was able to grow on nutrient-poor and nitrogen-rich soils as well. On several sites with degraded vegetation we detected higher sodicity (higher Na_exch, SAR, and pH, see Figure 5) than in the reference salt steppes of Hungary. We explain this unexpected finding by the impact of past ploughing. In pristine Solonetz steppes, the topsoil layer, where salinity and pH are usually lower than at deeper levels, is leached [65], as Figure 6 similarly demonstrates in plots of ART, group 1, pH. When such a soil type is rotated by ploughing, the deeper layer is exposed to the surface. As it is more clayey and sodic than the topsoil, the salts will leach more slowly and the new secondary sodic soil surface is again occupied by halophytes. Simultaneously, the disturbed surface favors the colonization of opportunistic weeds [58] which in case of abandonment remain a constant component of the regenerated vegetation. The plots in Slovakia demonstrate this phenomenon. Every site of Pannonic salt steppes has there undergone an alteration of the soil due to attempts to fertilize the salt-affected areas. They have been ploughed at least once, which caused an overturn of the soil profile and changes in soil properties. For this reason, the typical plant-soil correlations detected, for example by Tóth and Rajkai [66], have not been confirmed recently in Slovakia. Our samples taken from the topsoil (at a depth of 12 cm) are not deep enough to demonstrate the effect of soil rotation. Soil analyses were therefore not straightforward in evaluating the current condition of the target habitat, compared to the vegetation or land use approach. However, they revealed the whole range of key soil properties in which the studied indicator species and their target vegetation persist from favorable (optimal) to marginal (suboptimal) ecological conditions.

Current Condition of Indicator Species of Pannonic Salt Steppes
A small proportion of plots (19%) in Slovakia overlapped with the Hungarian reference ones, having the highest naturalness (group 1). The majority (57%) of plots had decreased naturalness and 24% of plots were detected in the lowest naturalness (group 5). Their small area within each site points to reduced habitat quality compared to Hungarian sites where each vegetation type was found in larger areas in a relatively high habitat quality [19]. In Slovakia, salt steppes of high naturalness are limited only to a few sites and the long-term survival chance is limited due to fragmentation and accelerated vegetation degradation.
The most threatened ones are barren spots of Camphorosmetum annuae and its indicator species, C. annua. In its eight known localities at its absolute northern distribution limit, there is no population where its area exceeds 50 m 2 . On the other hand, the disappearance of the community in its central area is local, for example in Kisalföld (northwestern Hungary) close to the study area Podunajská nížina where an overall decreased habitat quality has been reported [19,67]. Being a border floristic element is therefore a hindering factor for species survival.
Regarding A. santonicum and T. pannonicum, their vital populations in Slovakia are mono-dominant stands lacking specialists typical for their associations. They are able to co-exist with generalists (mainly A. santonicum, described also by Bölöni et al. [37]), but their seemingly good condition is temporary. Abundance of these indicator species therefore cannot be linked with habitat naturalness.

Conservation Perspectives of Pannonic Salt Steppes on Their Border Area
The ecological circumstances (vegetation composition and land use) of A. santonicum, C. annua, and T. pannonicum reflect well the current habitat quality of Pannonic salt steppes. Worsened environmental conditions (enrichment of generalist species in the vegetation and lack of habitat management) pointed to decreased habitat quality at the northern distribution limit of the target habitat.
Natural habitats of halophytes deteriorate faster in abandoned vegetation due to the high pressure of spreading generalists than on sites where soil was formerly ploughed [62]. Therefore, the introduction of grazing is the most desired restoration measure for their conservation as phytomass consumption and seed dispersal promote the emergence/survival of specialists. A positive example is the rediscovery of Triglochin maritima in salt marshes of Slovakia, a rare species in the Pannonian Basin which was found on trampled salt depressions after 60 years [68]. An intended consequence of grazing was to improve the survival conditions of A. santonicum, C. annua, and T. pannonicum in Slovakia. Those stands are, however, not in the optimal condition seen in the center of their distribution. Populations of these species are well maintained, but restoration of their characteristic plant communities is unfeasible even at the most appropriate grazing intensity. The prospect of reversing the degradation processes is more promising in contiguous areas than in the highly fragmented salt habitats on their distribution border.

Conclusions and Practical Outcomes
The habitat quality of Pannonian salt steppes at their northern geographical limit is decreased compared to their central occurrence. Only 19 out of the 101 plots were detected at the highest naturalness. The recent sites are scattered islands inside the intensively used landscape and halophytic specialists are substituted by generalists (Elytria repens, Agrostis stolonifera). Suboptimal ecological conditions which increase their risk of extinction may be mitigated by appropriate management in the case of Artemisia santonicum and Tripolium pannonicum subsp. pannonicum, where wider niche and competition ability was detected. The occurrence of the therophytic, low competitor Camphorosma annua was driven largely by abiotic stress; grazing alone is insufficient for its long-term persistence in decreased habitat extremity.
With respect to the groups of naturalness, soil analyses brought disparate results. Contrary to expectations, higher habitat naturalness did not correspond with habitat extremity (high soil sodicity). Each indicator species in its target vegetation was detected on less sodic soils, too, regardless of the level of naturalness. We assume that C. annua, A. santonicum, and T. pannonicum are able to grow on soils of greatly decreased salinity/sodicity if competitor species (generalists) are absent. Even on the reference vegetation in Hungary were detected low rates of SAR, Na_exch, EC, and pH (see Figure 5 for minimal values). This refers also to the high stability of salt steppes there thanks to the continuous non-intensive grazing. On Slovak sites the abiotic stress takes the major role in the persisting of abandoned or short-time grazed salt steppes.
The plot-based vegetation analyses (beyond presenting the optimal and marginal synecological range limits of indicator species) can serve as a quick tool for assessing the overall condition of the priority habitat 1530 Pannonic inland salt steppes. We proceeded from the definition of the Habitats Directive [69] which states that a site is of good quality if it allows a good chance of survival for a range of species that are typical for the habitat. The presented method may be used as a basic material in conservation practice, possibly to apply to different non-forest habitats in any other geographic area. Accurate habitat assessment based on remote sensing on a micro-regional scale [70] and the increasing efficiency of drones [71] has high accuracy, but their time and technological demand show some limitations which are difficult to overcome when the main goal is to obtain complete data in a short time. Evaluating habitat conservation status still relies on traditional fieldwork.