Features of Natural Succession of Ex-Arable Forest Steppe Grassland (from Western Romania) under the Influence of Climate

Important land surfaces from hill and mountain areas from the northern hemisphere formerly used for cropping were abandoned. Often, the abandoned land evolved by natural succession to grassland, shrubland or even to forest. The main goal of this paper is to bring new datasets necessary for the understanding of the evolution of ex-arable grassland vegetation from the forest steppe area into relationship with climate. The researches were performed in the locality of Grădinari (Caraş-Severin County, Western Romania) on an ex-arable plot abandoned since 1995. The vegetation data were collected for 19 years (time interval 2003–2021). The analyzed vegetation features were floristic composition, biodiversity and pastoral value. The climate data considered were air temperature and rainfall amount. The vegetation and climate data were correlated statistically, with a view to highlighting the potential impact of the temperature and rainfalls during the evolution of succession process on the grassland’s floristic composition, biodiversity and pastoral value. The pressure of the increased temperatures on the natural restoration process of the biodiversity and pastoral value of ex-arable forest steppe grassland could, at least partially, be mitigated by random grazing and mulching works.


Introduction
The reclamation of every abandoned arable land plot has great importance, because the intensity of arable land loss is very dynamic [1]. Some assumptions claiming that at worldwide level in about the last 40 years already lost about a third from the total arable land, mainly due to soil erosion and pollution [1,2]. Other reasons for the abandonment of the arable land in the northern hemisphere are the location of the fields in hill and mountain areas, fragmentation and transition of the ownership from public to private (e.g., in the former communist countries) [3][4][5], population migration from rural to urban areas [6][7][8], etc. Thus, the restoration of ex-arable land as successional secondary grassland [4,[9][10][11][12][13][14][15] is one of the ways of successful recovery of the economical, ecological and even social functions [3,10,13,[15][16][17][18][19] of these land surfaces. The great importance of the ex-arable land reclamation by turning it in grassland consists in numerous advantages. The restored ex-arable grassland provides conservation of biodiversity at the landscape

Climate
The climate dynamics during the research period were characterized by exceptional values compared with the multiannual average values. The most dramatic evolutions were noticed in the case of air temperatures ( Figure 1a). The exceptional temperatures were above the average value from the experimental period more frequently in the last decade, namely in 9 years, and were above the 1955-2012 multiannual average value (11.21 • C)  Thus, in general the values of air temperature had an increasing trend. In the case of rainfalls, there was a decrease near to the end of the studied time interval, but in the same time they were characterized by an irregular accentuated variation from one year to the next. In this context, it was considered that the conection between the analyzed climate parameters can be expressed with the Lang Rainfall Index (R). For the experimental site, The situation regarding the rainfalls was not so dramatic as it was for the temperatures ( Figure 1b). Thus, the rainfall amounts were even greater, with 9 years overpassing the The wettest months registered during the research were July 2011 (238.2 mm) and July 2020 (234.2 mm), while the multiannual average value for July is 101.4 mm. The driest months were July 2015 (1 mm) and October 2013 (4 mm).
Thus, in general the values of air temperature had an increasing trend. In the case of rainfalls, there was a decrease near to the end of the studied time interval, but in the same time they were characterized by an irregular accentuated variation from one year to the next. In this context, it was considered that the conection between the analyzed climate parameters can be expressed with the Lang Rainfall Index (R). For the experimental site, the R values obtained during 2003-2021 show a high variation, as they fell between 41.82-119.31 ( Figure 2). According to the classification of the R values as characteristic climate types, the lowest values falling in the interval 40-60 were obtained toward the end of the analyzed interval (years 2013, 2015, 2019 and 2021), these values corresponding to the semiarid climate. In 74% from the analyzed cases, covering 14 years from the study interval, the R values were framed between 60-100, corresponding to the warm temperate climate. Values of R greater than 100, characteristic of the humid temperate climate, were obtained in a single case, namely in the year 2005. Thus, in general the values of air temperature had an increasing rainfalls, there was a decrease near to the end of the studied time inter time they were characterized by an irregular accentuated variation fr next. In this context, it was considered that the conection between th parameters can be expressed with the Lang Rainfall Index (R). For th the R values obtained during 2003-2021 show a high variation, as they 119.31 ( Figure 2). According to the classification of the R values as ch types, the lowest values falling in the interval 40-60 were obtained to analyzed interval (years 2013, 2015, 2019 and 2021), these values co semiarid climate. In 74% from the analyzed cases, covering 14 ye interval, the R values were framed between 60-100, corresponding to t climate. Values of R greater than 100, characteristic of the humid temp obtained in a single case, namely in the year 2005.

Biodiversity Dynamics and Relationship with Air Temperature and Rai
Biodiversity was analyzed during the experiment by the speci Shannon index (H') and the Simpson index (D). The species rich increasing trend during the research ( Figure 3). The lowest species rich

Biodiversity Dynamics and Relationship with Air Temperature and Rainfall
Biodiversity was analyzed during the experiment by the species richness (S), the Shannon index (H ) and the Simpson index (D). The species richness had a general increasing trend during the research ( Figure 3). The lowest species richness was registered in 2004 (20 species) and the highest in 2020 (92 species). The evolution of the species richness from the ex-arable grassland from Grădinari proves that the number of taxa increases throughout the research period.
The species richness corresponds to the expected yearly distribution according to the multinomial test outputs (p < 0.001). The number of taxa appeared to be influenced by the year (Figure 4). Climate is one of the independent features of interest for our investigation that can be characterized yearly in relationship with the vegetation features and presented in the following chapters of the work.
Plants 2023, 12, 1204 5 of 24 in 2004 (20 species) and the highest in 2020 (92 species). The evolu richness from the ex-arable grassland from Grădinari proves that t increases throughout the research period. The species richness corresponds to the expected yearly distribut multinomial test outputs (p < 0.001). The number of taxa appeared to b year ( Figure 4). Climate is one of the independent features of interest f that can be characterized yearly in relationship with the vegetation fea in the following chapters of the work.    The species richness corresponds to the expected yearly distr multinomial test outputs (p < 0.001). The number of taxa appeared year ( Figure 4). Climate is one of the independent features of inter that can be characterized yearly in relationship with the vegetation in the following chapters of the work.    In general, the Shannon indexes obtained during the research proved the increase in the biodiversity in a great measure from low to high and even with respect to very high biodiversity (values of H above 3.5).
The Simpson index (D) has a little different pattern as can be noticed in Figure 6, but the meaning is similar, with the greatest biodiversity being in 2016 (D = 0.03) and the lowest in 2003, 2004 and 2006 (D = 0.16), respectively. Simpson's dominance index is inverse proportional to the Shannon index, the results obtained being near 0 and characteristic for very high diversity (D = 0 represents infinite diversity; D = 1 represents no diversity).  The results regarding the correlation coefficients between the tem biodiversity indexes are presented in Table 1. It was noticed that the air tem February (TA II) was positively correlated with the Shannon index (r = 0.47 correlation identified was a negative one and showed that the increa temperature in September determined the decrease in the Simpson index (r Table 21. Correlations between the averages of the temperature (I-XII) and species H' and D) in ex-arable grassland from Grădinari (period 2003-2021).   The results regarding the correlation coefficients between the tem biodiversity indexes are presented in Table 1. It was noticed that the air tem February (TA II) was positively correlated with the Shannon index (r = 0.47 correlation identified was a negative one and showed that the increa temperature in September determined the decrease in the Simpson index (   The results regarding the correlation coefficients between the temperatures and biodiversity indexes are presented in Table 1. It was noticed that the air temperature from February (TA II) was positively correlated with the Shannon index (r = 0.479 *). The other correlation identified was a negative one and showed that the increase in the air temperature in September determined the decrease in the Simpson index (r = −0.498 *). After the analysis referring to the existence of the correlation between the rainfall amount and the considered biodiversity indexes (Table 2), no statistically significant correlation was identified.

Florisic Composition Dynamics and Relationship with Air Temperature and Rainfall
Floristic composition analysis of the successional ex-arable forest steppe grassland grouped the taxa in three functional groups, namely grasses, legumes and forbs. They were characterized according to species number (no.) and specific contribution rate (CS%).
The less numerous grass species were determined in 2004, namely three; and the greatest number of grass taxa was eighteen in the year 2020 ( Figure 7). From the perspective of the specific contribution rate of the three functional groups, the differences were relatively small, because the lowest value was CS% = 43.06 in 2014 and the highest CS% = 62.77 in 2021 ( Figure 8).
After the analysis referring to the existence of the correlation between the rainfall amount and the considered biodiversity indexes (Table 2), no statistically significant correlation was identified. Table 2. Correlations between the rainfall amounts (I-XII) and species biodiversity (S, H' and D) in ex-arable grassland from Grădinari (period 2003-2021). I  II  III  IV  V  VI  VII  VIII  IX  X  XI  XII  I-XII III-

Florisic Composition Dynamics and Relationship with Air Temperature and Rainfall
Floristic composition analysis of the successional ex-arable forest steppe grassland grouped the taxa in three functional groups, namely grasses, legumes and forbs. They were characterized according to species number (no.) and specific contribution rate (CS%).
The less numerous grass species were determined in 2004, namely three; and the greatest number of grass taxa was eighteen in the year 2020 ( Figure 7). From the perspective of the specific contribution rate of the three functional groups, the differences were relatively small, because the lowest value was CS% = 43.06 in 2014 and the highest CS% = 62.77 in 2021 ( Figure 8).  After the analysis referring to the existence of the correlation between the rainfall amount and the considered biodiversity indexes (Table 2), no statistically significant correlation was identified. Table 2. Correlations between the rainfall amounts (I-XII) and species biodiversity (S, H' and D) in ex-arable grassland from Grădinari (period 2003-2021). I  II  III  IV  V  VI  VII  VIII  IX  X  XI  XII  I-XII III-

Florisic Composition Dynamics and Relationship with Air Temperature and Rainfall
Floristic composition analysis of the successional ex-arable forest steppe grassland grouped the taxa in three functional groups, namely grasses, legumes and forbs. They were characterized according to species number (no.) and specific contribution rate (CS%).
The less numerous grass species were determined in 2004, namely three; and the greatest number of grass taxa was eighteen in the year 2020 ( Figure 7). From the perspective of the specific contribution rate of the three functional groups, the differences were relatively small, because the lowest value was CS% = 43.06 in 2014 and the highest CS% = 62.77 in 2021 ( Figure 8).  Air temperature correlated with floristic composition (Table 3) highlighted several significant results, namely some positive correlations in the case of grasses species number and the temperatures from March (r = 0.456 *) and from September (r = 0.505 *); and a negative one between the temperature from February and grasses CS% (r = −0.482 *). The other functional group correlated positively with the air temperature from February (r = 0.503 *) was forbs as species number. Between air temperature and legumes, functional group, correlations in the timeframe of the present research were not identified. The species number of the main functional groups was not correlated with the rainfall amount during the experiment (Table 4), but the CS % values proved to be influenced by this climatic factor. Thus, the grasses' contribution increases with the increase in the rainfalls from January (r = 0.549 *) and decreases with the increase in the rainfalls from March (r = −0.565 *). The rainfall amount proved to have a positive impact on the contribution rate of the legumes group (Table 4), namely for the rainfalls from April (r = 0.622 **) and from December (r = 0.592 **), in addition to the total rainfalls from the entire year (r = 0.531 *) and from the vegetation season (r = 0.637 **). Regarding the analysis of the potential influence of the rainfalls on the specific contribution of forbs, there was identified a positive correlation with the rainfalls from March (r = 0.553 *) and a negative correlation coefficient with the rainfalls from December (r = −0.465 *) ( Table 4).

Pastoral Value Dynamics and Relationship with Air Temperature and Rainfall
The pastoral value (PV) dynamic is displayed in Figure 9. The lowest pastoral value determined was PV = 19.09 in the year 2003, namely the first year of the investigation. The highest pastoral value was identified in 2005, namely PV = 43.31, closely followed by the PV from 2020, which was 41.6. At the beginning of the research, the PV was characteristic of the medium quality grassland (VP between 14 and 25). The highest values of the PV Plants 2023, 12, 1204 9 of 24 determined in the ex-arable successional grassland were characteristic for good pasture (PV greater than 25).
The pastoral value (PV) dynamic is displayed in Figure 9. The lowest pastor determined was PV = 19.09 in the year 2003, namely the first year of the investigat highest pastoral value was identified in 2005, namely PV = 43.31, closely followe PV from 2020, which was 41.6. At the beginning of the research, the PV was chara of the medium quality grassland (VP between 14 and 25). The highest values o determined in the ex-arable successional grassland were characteristic for good (PV greater than 25). The only correlation identified between the air temperature analyzed values pastoral value was a negative one, namely for the temperature from June (r = − showing that the increase in the temperature from June determined the decreas pastoral value (Table 5). The rainfalls have a greater impact on the pastoral value (Table 6) than temperatures (Table 5). There were identified a series of positive correlations: on significant in the case of the rainfall amount from the vegetation season (r = 0.716 very significant for the total annual rainfall amount (r = 0.692 **), and other sig correlations for some monthly rainfall amounts. Thus, the significant correlation pastoral value were with the rainfalls from the months of February (r = 0.475 *), J 0.559 *), July (r = 0.469 *) and December (r = 0.475 *).  The only correlation identified between the air temperature analyzed values and the pastoral value was a negative one, namely for the temperature from June (r = −0.564 *), showing that the increase in the temperature from June determined the decrease in the pastoral value (Table 5). The rainfalls have a greater impact on the pastoral value (Table 6) than do the temperatures (Table 5). There were identified a series of positive correlations: one highly significant in the case of the rainfall amount from the vegetation season (r = 0.716 ***), one very significant for the total annual rainfall amount (r = 0.692 **), and other significant correlations for some monthly rainfall amounts. Thus, the significant correlations of the pastoral value were with the rainfalls from the months of February (r = 0.475 *), June (r = 0.559 *), July (r = 0.469 *) and December (r = 0.475 *). * p < 0.05, ** p < 0.01, *** p < 0.001.

Vegetation
Vegetation of a successional grassland has a dynamic structure [45,[47][48][49][50]. Thus, the characterization of the sward with the phytosociological "association" concept is not always proper, because it characterizes the ecosystems found in the climax evolutional stage [51]. Secondary grasslands are managed agro-ecosystems, which can be characterized most properly [52] from a typological point of view [3,53] that covers more complex datasets relating to the vegetation with the stational conditions and with the management [54,55]. The random intervention with grazing cattle [35,49,56,57] and minimal maintenance works [3,58] influences intensively the grassland vegetation. The extensive grazing is a major source of propagules of grassland species, useful for the intensification of the succession process of the abandoned arable land to grassland.
The ex-arable grassland from Grădinari was extensively grazed by sheep since the abandonment, with the mention that in about the last 10 years it was grazed randomly. According with the literature, grazing is beneficial for ex-arable land succession to grassland [35]. Thus, the establishment of the grassland species during the succession seems to be mediated by the soil microbiota too [59], grazing being beneficial for the proliferation of the soil microorganisms specific for grassland [59][60][61], as they favor the establishment of the grazing tolerant species [61].
Random grazing with sheep [57,60,62] and mulching works [3,58] applied late in autumn could have an important contribution to the great restoration and high presence of late successional grassland species. The certain effect of the grazing can be considered the increase in the number of taxa (see Figure 2) in a significant manner.
Forrest steppe grassland is a notion referring to a vegetation type spread across Eurasia, fulfilling the criteria of a biome, with particularities that divide it in nine specific regions [45], our research area being framed in Region A: Southeast Europe from the geographical point of view. Some of the species identified in the experimental field characteristic for Region A were Festuca valesiaca, Stipa capillata, Fragaria viridis, Brachypodium pinnatum, etc. According to the classification on the regions proposed by Erdős et al. [45], some of the forest steppe grassland species identified in our experimental field are characteristic for other regions of this biome, namely for Region B: East Europe (Filipnedula vulgaris, Salvia pratensis), for Region D: West Siberia and North Kazahstan (Lathyrus pratensis, Vicia cracca) and for Region F: far East (Botriochloa ischaemum, Poa pratensis).
In the first year of vegetation records (2003) on the analyzed ex-arable grassland, numerous annual weed species were dominant, above all Bromus hordeaceus. However, at the same time, some perennial grasses such as Lolium perenne, Calamagrostis epygeios and Cynodon dactylon were identified (see Table A1). In 2006, Festuca valesiaca appeared in the sward, and since 2007 it became the most dominant species together with Agrostis tenuis as codominant, these species remaining the most abundant in all the years of the research. In other researches referring to the successional forest steppe grassland evolution from Hungary, Calamagrostis epygeios was an abundant species at the beginning of abandonment, but its cover decreased in the later abandonment phase [41,63] This species was not abundant in our analysed plot in general; even it was present there. This fact could be explained by a faster transition of the vegetation from the studied abandoned arable land to forest steppe grassland.

Climate
The evidence of the extreme climatic average values shows a great climate pressure, mainly in the case of the temperature increase that nowadays is a worrying phenomenon. The same trend was noticed in the studied area, with a multiannual normal value of the air temperature of 11.21 • C, where it registered the overpassing above the normal value in 18 years from all the 19 years of the research time interval. More than that, in some years, the yearly average value of the air temperature was even greater than 13 • C (in 2013, 2015, 2018 and 2019), the greatest registered average value being 13.70 • C in 2018, which represents a deviation of 2.49 • C, such extremes being mentioned in the literature referring to our region [36,64]. Thus, such extreme heat events were also registered in other neighbouring countries, such as Bulgaria [46].
The rainfall regime was characterized in general by extremes regarding the monthly distribution of the rainfalls during the year, with high amounts of rain in a very short time [65,66] or long time intervals of severe drought [67,68]. In the experimental area, the average rainfall amounts were in general near or above the multiannual averages, the situation being better in comparison with the temperatures that have strongly deviated from the average values.
In the analyzed time interval (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019)(2020)(2021), the value of the main meteorological parameters implied in the development of the vegetation structure and typology had great variations. The general trend of the temperature, with few exceptions, was an increasing one since the beginning of the research time interval. In the case of rainfall amount, a general decrease in them was noticed toward to the end of the research interval, and an irregular variation from one year to the next. The Lang Rainfall Index multiannual mean value for the study area is 79.44, characteristic for warm temperate climate. The values of R had a great variation during 2003-2021 (R from 41.82 to 119.31). According to the characteristic classification of R [69], four years corresponded to the semiarid climate (years 2013, 2015, 2019 and 2021), fourteen years corresponded to the warm temperate climate and one year (2005) to the temperate humid climate. The Lang Rainfall Index will be analyzed in following research to see if the arid climate continues to occur in the studied area. The good transition of the analysed abandoned arable land to the secondary successional forest steppe grassland was probably influenced in a great measure by the relatively good rainfall regime at our experimental time scale.

Biodiversity Dynamics and Relationship with Air Temperature and Rainfall
The general evolution trend of the biodiversity analyzed as species richness, Shannon index and Simpson index highlighted a certain increase in the specific diversity of the analyzed ex-arable forest steppe grassland during the research interval. The evolution of species richness seemed to be influenced by year (see Figure 3). Such results regarding the increase in the species richness [4,5,30] in ex-arable forest steppe successional grassland and the Shannon and Simpson diversity indices [41,63] were found in the literature. Thus, often in practice the secondary succession of ex-arable grassland proved to determine the increase in the vegetation biodiversity [3][4][5]30,41,63,70], our results confirming this hypothesis for the analysed ex-arable forest steppe grassland in conditions of natural succession.
Climate influences grassland vegetation typology and structure [49,54] and the evolution by secondary succession of the abandoned arable land to grassland [42,45,47,71,72]. Air temperature proved to have an impact on biodiversity, there being determined the existence of some correlations, namely a positive one between the air temperature from February and the Shannon index (r = 0.479 *) and a negative one between the air temperature from September and the Simpson index (r = −0.498 *). There were noticed some other values close to the significant ones, which suggest the potential for the future investigation of this hypothesis. Extrapolating, the air temperature from February, March and September could also influence the increase in the biodiversity in the analyzed successional grassland, because the obtained r calc values were very close to the r critical values. The rainfall amounts from the experimental area seemed to not be very influential for biodiversity, from statistical point of view, during the succession at the time scale of our experiment. Some of the researches regarding the impact of climate change on grassland vegetation biodiversity assumes the existence of at least a partial influence of the climatic extremes [67,[73][74][75]. Aside from the analyzed climatic factors, other factors were implied probably too, with stronger influence on the biodiversity, such as random grazing with sheep and mulching works applied late in autumn. They could have an important contribution to the great natural restoration process and even an increase in the biodiversity, such results being confirmed by the literature [28,49,58,60,70].

Florisic Composition Dynamics and Relationship with Air Temperature and Rainfall
Floristic composition is a feature of the grassland vegetation community influenced by the cumulated actions of the biotic and abiotic factors with the management [39,45,47,49]. The dominance of the perennial grasses in the sward of the analyzed ex-arable successional grassland is a good sign of the vegetation recovery [3,4,76]. Even the statistical results regarding the influence of the air temperature on grassland species were less statistically significant at our experiment time scale. The appearance in 2015 of the thermophilic and xerophytic grass species Stipa capillata and its constant presence in the following years of the research suggest the adaptation of the vegetation sward to dryer and hotter summers. The negative correlation of the grasses' contribution rate with the rainfalls from March could be explained by the low soil temperatures and nebulosity, which could affect the grasses' growth at the beginning of spring.
The establishment of the perennial grasses in the abandoned arable land contributes to the succession of the secondary grassland vegetation, and the climate can be a driver in this way together with the management [28,60,73,76].
Legumes are also an important element of the quality of the grassland vegetation [13,43,60]. In the present experiment, it was noticed that they were the most sensitive to the rainfall regime. An interesting situation was noticed in the case of Trifolium alexandrinum, a Mediterranean species that occurred in a relatively high rate (CS% = 9.71) in the year 2005. This species' propagules were probably brought by the sheep that have probably grazed previously on a plot abundant in T. alexandrinum from the farm, where it was seeded in the 1960s and is still abundant and self-seeding since then. The presence of T. alexandrinum is not constant in the experimental plot; the next appearances were registered in rates comprised between 0.38 and 1.44% in the years 2010-2012, 2016-2017 and 2020. The high rate of T. alexandrinum from 2005 could be explained by the numerous empty ecological niches of the ex-arable grassland in the early successional phase that are decreasing while the succession process is advancing.
Thus, the air temperature seemed to not be correlated with the contribution rate of the legumes at the timescale of the present research. However, in the case of the rainfall correlation with the specific contribution of the legumes, the results obtained highlighted that this functional group was positively influenced by the yearly rainfall regime (RA I-XII) and by the rainfall regime from the vegetation period (RA III-IX). Results highlighting the impact of climate on the functional groups were investigated by the modeling and creating of algorithms, which can have future application on the future research of the temperate grasslands [73].
The presence of the forbs is an important element of the grassland floristic composition [43,76,77]. They are bringing added value by certain valuable species from other botanical families rich in nutritive substances and active compounds (e.g., Achillea millefolium, Cichorium intybus, Plantago sp., Taraxacum officinale, etc.) and are enriching the biodiversity [54,78]. A balanced floristic composition is essential for the stability of the grassland ecosystem and for its resilience in the situation of unfavorable conditions such as extreme weather or inappropriate management (e.g., overgrazing, under-grazing, abandonment etc.) [28,45,63]. In a recent research, the hypothesis regarding the interaction between the grassland functional groups of plants and climate, together with other biotic and abiotic data, was used in the development of a modeling code in Norway for alpine grassland with the purpose to be used in the future as a tool in grassland ecosystem analysis [77].

Pastoral Value Dynamics and Relationship with Air Temperature and Rainfall
The evolution of the PV during the research was very irregular, but the greatest increase was noticed in the third year of research, namely the eleventh year after the land abandonment. According to the literature, in Alta Murgia National Park from Southern Italy, the pastoral value of the successional grasslands from arable land abandoned by more than 50 years was higher even in comparison with the semi-natural grasslands [79].
The impact of the air temperature on the pastoral value was not concluded during the experiment; the only correlation identified showed the decrease in the PV due to the increase in the air temperature from June (r = −0.564 *). The rainfall regime during the entire year and during the vegetation season seemed to have the most significant impact on the pastoral value of the studied successional forest steppe grassland from the experiment (r = 0.692 **; r = 0.716 ***). In the literature, there are not numerous the references regarding the influence of the temperature and rainfall amount on the pastoral value of successional grasslands for such timeframes, but some were focused on the influence of rainfall on dry matter yield [47] and fresh fodder yield [73]. Thus, the results obtained by Tripolskaya et al. [28] highlighted the small impact of the climate change on the temperate grassland yield due to the early spring water supply that mitigates the later rainfall deficit. Our monthly results suggested the same thing, namely the significant positive correlations between the rainfall amount from December (r = 0.475 *) and from February (r = 0.475 *) on the pastoral value (see Table 6), but in our case the rainfalls from June (r = 0.559 *) and July (r = 0.469 *) had a positive influence on the PV too.

Description of the Study Area
The background of the present research concerns the investigation of the climate implications for the natural recovery of an ex-arable forest steppe grassland. The vegetation data were collected from the Mezdraia area belonging to the locality of Grădinari (Caraş-Severin County, Western Romania) from an ex-arable grassland plot ( Figure 10). experiment (r = 0.692 **; r = 0.716 ***). In the literature, there are not numer references regarding the influence of the temperature and rainfall amount on the value of successional grasslands for such timeframes, but some were focused influence of rainfall on dry matter yield [47] and fresh fodder yield [73]. Thus, th obtained by Tripolskaya et al. [28] highlighted the small impact of the climate ch the temperate grassland yield due to the early spring water supply that mitigates rainfall deficit. Our monthly results suggested the same thing, namely the sig positive correlations between the rainfall amount from December (r = 0.475 *) a February (r = 0.475 *) on the pastoral value (see Table 6), but in our case the rainfa June (r = 0.559 *) and July (r = 0.469 *) had a positive influence on the PV too.

Description of the Study Area
The background of the present research concerns the investigation of the implications for the natural recovery of an ex-arable forest steppe grassla vegetation data were collected from the Mezdraia area belonging to the loc Grădinari (Caraş-Severin County, Western Romania) from an ex-arable grassla ( Figure 10).  The study area is located in the southwest part of Romania, on the territory of Caraş-Severin County, at a distance of 2 km from the border with the Republic of Serbia. The experimental site (45 • 8 40" N, 21 • 33 20" E) has a surface of 10 hectares and is located in the Tirolului Hills. According to the data obtained by the processing of the Digital Elevation Model (DEM) [82] using methods and techniques specific for the Geographic Informatic Systems (GIS) [83], it was noticed that the experimental site is placed at elevations comprising 127-191 m, with the average altitude of 170 m a.s.l. Because the area is hilly (see Figure 10), the slope varies between 2.4 and 16.3%, with an average value of 9.4%. The experimental site is placed on the eastern and southeastern slope, which influences the vegetation type. From the biogeographical point of view, the experimental site belongs to the forest steppe area [84,85]. Figure 11 presents some views from the analyzed field at different moments during the research. experimental site (45°8′40″ N, 21°33′20″ E) has a surface of 10 hectares and is located i Tirolului Hills. According to the data obtained by the processing of the Digital Elev Model (DEM)[82] using methods and techniques specific for the Geographic Inform Systems (GIS) [83], it was noticed that the experimental site is placed at eleva comprising 127-191 m, with the average altitude of 170 m a.s.l. Because the area is (see Figure 10), the slope varies between 2.4 and 16.3%, with an average value of 9 The experimental site is placed on the eastern and southeastern slope, which influe the vegetation type. From the biogeographical point of view, the experimental site bel to the forest steppe area [84,85]. Figure 11 presents some views from the analyzed fie different moments during the research.  The biogeographical vegetation framing of the studied area is forest steppe and the soil type is brown forest soil [86]. The arable land field has been abandoned since the year 1995 [86] and the vegetation has evolved gradually by natural succession. Since the abandonment, the analyzed successional grassland surface has been used by sheep for random extensive grazing. The maintenance work applied was only mulching, which occurred late in autumn at the ending of grazing season, but it is also applied randomly.

Vegetation Data Collection and Analysis
The vegetation data were collected for 19 years, namely from the year 2003 to 2021 (months May, July and September). The vegetation surveys were conducted with the linear point quadrate method described by . The considered vegetation features in the present research were the following: floristic composition (grasses, legumes and forbs considered as species number and specific contribution-CS%) [3,39,47,48,77,88], biodiversity (as species richness-S, Shannon index-H [3,89,90] and Simpson index (D) [3,48,91]) and pastoral value-VP (0-100 scale) [3,53,87,92]. The pastoral value characterizes the relative forager value of the grassland sward based on the specific contribution in the grassland cover of the species with the forager value expressed as the Specific Forage Quality Index (SFQi) [78,87]. The Specific Forage Quality Index (SFQi) used for the VP calculation is characteristic for the Romanian grassland species [78].

Climate Data Collection and Analysis
The multiannual average temperature in the study area is 11.21 • C, and multiannual average rainfall amount is 890.5 mm. The climate parameters considered in this work were air temperature and rainfall amount registered at the nearest meteorological station (Oraviţa, Caraş-Severin County). The climate data used in statistical analyses were grouped in climatic years (months I-XII), vegetation seasons (months III-IX) and monthly averages.
The Lang Rainfall Index (R) represents a ratio between the rainfall amount and air temperature [69,93,94], namely the "connection" between the two analyzed climate parameters, because temperature is considered one of the main factors implied in the process of evapotranspiration [95]. The Lang Rain Factor is also named the pluvial-thermal index [95] and illustrates the water inputs and outputs from a territory. It was taken into consideration because it is considered to be influential for the structure and typology of vegetation.

Statistical Analysis
Based on the vegetations and climate data, a series of statistical analyses were undertaken. The software programs used for data processing were Excel 2019 [96] and JASP 0.16.4.0; the statistical methods used were Pearson r [97][98][99] and a multinomial test [99].
The aim of this research was to release new outputs and datasets regarding the evolution of the vegetation sward of successional ex-arable grassland from the forest steppe area related to air temperature and rainfall amounts in conditions of natural succession during almost two decades, namely nineteen years (time interval 2003-2021).

Conclusions
Successional grasslands are a potential resource of biomass, biodiversity and habitat for wildlife. Such naturally recovered ex-arable grassland ecosystems are able to restore their natural cycles and provide ecological and socio-economical services as a natural ecosystem.
The resilience of the analyzed ex-arable land was good even in the present climate conditions, with years characterized by higher temperatures and severe droughts or hard rainy time intervals. The rainfall regime less disturbed in comparison with the temperature probably had a good impact on the transition of the abandoned arable land to successional forest steppe grassland at our experiment time scale. However, the appearance of the xerophytic species Stipa capillata since 2015 could be connected with the appearance of the years with a semiarid climate (years 2013, 2015, 2019 and 2021). This trend should be investigated in future research to check if the hypothesis of climate warming is confirmed.
The other element that has probably contributed to the good recovery of the grassland vegetation could be soil seed bank. However, the random grazing with sheep and mulching works applied late in autumn could have a very important contribution to the great natural restoration process and even on the increase in the biodiversity and the high presence of late successional grassland species. The domestic and wild herbivores, namely the grassland maintenance tools and machineries, probably were vectors for some grassland species' propagules that probably have speed up the natural restoration process of the grassland species. Thus, climate change pressure on the natural restoration of the biodiversity and pastoral value of ex-arable forest steppe grassland could be mitigated at least partially by random grazing and mulching works.
The outputs from this work can be helpful for the setting of strategies of natural recovery of great surfaces of abandoned fields from the forest steppe area in the context of climate change. The impact of climate change on the natural restoration of ex-arable successional grassland should be investigated more for a better understanding of the implications at different levels such as vegetation structure, biodiversity, floristic composition, forage value, productivity, carbon sequestration, economic and social value, etc. Acknowledgments: This paper is published from the project 6PFE of the University of Life Sciences "King Mihai I" from Timisoara and Research Institute for Biosecurity and Bioengineering from Timisoara. We are grateful to Doru George Laies , for the great support offered on site for data collection.

Conflicts of Interest:
The authors declare no conflict of interest.