Introduction

The Mediterranean Basin is one of the world’s major hotspots of plant biodiversity1,2,3, which, at the same time, faces a climate and habitat crisis due to ever-increasing warming, drying, deforestation and landscape fragmentation4,5,6,7. It is thus of the utmost urgency to assess the germination processes and the environmental threats that regard keystone plant species. One of these is the monospecific taxon Poterium spinosum L. (Rosaceae family), called “thorny burnet”, which dominates eastern Mediterranean shrublands known as “bathah” in Israel and “phrygana” in Greece8. Because of its severe fragmented distribution and scarce renewal capacity, P. spinosum is considered as “endangered” (EN) on a regional scale9, and as “vulnerable” (VU) on Italy’s scale10. Over the last few decades, massive landscape fragmentation further disrupted the range size of P. spinosum, causing local extinction and making surviving populations more and more scattered and isolated11,12.

Ex situ conservation of existing genetic resources is emerging as a valuable tool to create seed reserves for programs of biodiversity protection and restoration13. Regarding P. spinosum, several studies investigated its ecology and seed germination14,15,16,17,18,19,20. However, to our knowledge, no study has ever investigated the germination behavior of P. spinosum at an intrapopulation level. In this study, indeed, we compared the different germination processes of five distinct subpopulations of P. spinosum, growing on the south-eastern coast of Sicily. Although analyzing intrapopulation variability of seed germination could improve our understanding of how plants adapt and survive, especially in unpredictable or changing environments, studies about intrapopulation germination are generally very poor21,22. For an exhaustive picture of the conservation state, germination studies should be also integrated with the investigation of environmental threats such as climate changes and habitat fragmentation, which are primary drivers of biodiversity loss23,24,25,26. The aims of this study were to investigate (a) the intrapopulation germination variability of P. spinosum at different temperatures, and (b) the trends of climate changes and (c) land-cover transformation affecting two different populations of P. spinosum over the periods 1931–2020 and 1958–2018, respectively. The ultimate end of this study was to corroborate the conservation efforts of P. spinosum by shedding further light on its germination ecology poorly known at intrapopulation level, and by providing a scenario that shows the magnitude of the threats posed by climate changes and habitat fragmentation.

Materials and methods

Study area

The distributional range of Poterium spinosum is restricted to the south-eastern part of Sicily, from Portopalo to Villasmundo localities (Fig. 1; Table 1). This area lies mainly on the Hyblean Plateau, which is crossed by numerous deep and narrow torrential canyons (in Italian, the so-called “cave”). Across the Hyblean Plateau, two distinct sectors can be recognized: the eastern sector mainly composed of a Cretaceous-Late Tortonian carbonate succession, with intercalations of vulcanites, and the western sector marked by an Oligocene–Miocene carbonate succession27,28. P. spinosum occurs along the coastline and in inner hilly areas (Fig. 1), both in protected areas (e.g., Natura 2000 sites) and in unprotected sites29,30,31. However, on the basis of direct field surveys, this species has almost disappeared from several localities such as “Costa Reitani” and “Marzamemi”. This study, in particular, analysed climate trends and land-cover changes affecting the populations of P. spinosum distributed across the coastal protected areas of “Plemmirio” (Fig. 1, site A) and “Vendicari” (Fig. 1, site B). The germination experiments focused instead on the subpopulations of P. poterium growing within the “Vendicari” reserve (Fig. 2). The “Plemmirio” area mainly consists of Pleistocene limestone rocks and calcarenites. In turn, the “Vendicari” area is characterized by rocks mixed with sandy dunes, which are inwards and often replaced by large brackish lagoons, such as the so called “Pantano Piccolo” (small quagmire) and “Pantano Grande” (large quagmire). The sedimentary sequence of the “Vendicari” site consists of Pliocene–Quaternary deposits. The climate of the distributional area of P. spinosum is typically Mediterranean, with irregular rainfall events, harsh dry summer (up to 5-month dry periods), and mild winter periods. According to the classification of Rivas-Martínez32, the study area lies in the Macrobioclimate Mediterranean, with thermo-mediterranean thermotype and semiarid ombrotype. The software ArcGIS version 10.2 was used to make Figs. 1 and 2, whose images are orthophotos publicly available from the official website of the Sicilian Regional Government.

Fig. 1
figure 1

Occurrence sites of Poterium spinosum across Sicily, and within the protected areas of “Plemmirio” (A) and “Vendicari” (B) (see Table 1 for the correspondence between number and site name).

Full size image
Table 1 Occurrence sites of Poterium spinosum across Sicily, their names and geographical information (see Fig. 1 for the spatial distribution of these sites).
Full size table
Fig. 2
figure 2

Geographical distribution of subpopulations A, B, C, D, E of Poterium spinosum across the “Vendicari” protected area (site B as shown in Fig. 1).

Full size image

Biology and ecology of Poterium spinosum

The genus Poterium consists only of the species Poterium spinosum L., called “prickly burnet” or “thorny burnet”. Locally known as “spinaporci”, it is a nanophanerophyte belonging to the Rosaceae family (Fig. 3). P. spinosum is a very branchy thorny shrub (60–70 cm high, max 1 m), with young tomentose shoots; the lateral branches are leafless ending in dichotomous, whitish thorns; the leaves are imparipinnate and pubescent. Inflorescences (short spikes) with unisexual flowers are dense; the female flowers, with red–purple feathery stigmas, are in the upper portion, whereas the male flowers (with 10–30 stamens) are in the lower one16,33. Flowering season occurs from March to May, and fruits mature in the period June–August; the fruit is a spongy berry. P. spinosum reproduces by means of anemophily pollination, and autochory or endozoochory dispersal34. The annual fruit production is normally very high, but the rate of seed germination varies among different countries19. Although fire may promote germination and seed dispersal35,36, seedling survival is generally low37. The plant propagates both sexually and vegetatively. The vegetative reproduction occurs via rooting of branches (“ramets”); the production of new ramets can be thought of as growth leading to clonal expansion11.

Fig. 3
figure 3

(A) Garigue vegetation characterized by Poterium spinosum shrubs; (B) habit of P. spinosum; (C) detail of the fruits; (D) fruit sampling; (E) collected fruits ready for germination tests; (F) germination of P. spinosum from fruits; (G) germination of P. spinosum from seeds.

Full size image

P. spinosum is a pioneer species typical of shrublands known as “bathah” in Israel, “tomillares” in Spain, and “phrygana” in Greece8. In the eastern part of its distributional range, P. spinosum grows in a wide range of habitats38,39. The species shows a halo-tolerant trait interpreted as adaptation to coastal habitats where it generally grows19. In Italy this plant occurs in few regions33, and grows on a narrow range of habitats (garrigue/scrubland), in isolated and generally small zones near the shoreline, on calcareous rocks or on littoral gravelly areas31,40. In south-eastern Sicily, P. spinosum colonizes areas near the coastline as well as inner areas up to c. 600 m a.s.l., and usually grows on calcareous rocks and sandstones or sandy gravelly substrate.

Germination experiments

The laboratory tests investigated (1) the germination behavior of fruits and seeds of P. spinosum to shed further light on the effects of fruit spongy tissue on germinability; (2) how the germination behavior of P. spinosum varies at intrapopulation level. Regarding the second point, we compared, under four different conditions of temperature, several germination parameters of five different subpopulations of P. spinosum growing within a protected area (Fig. 2), which is also a Natura 2000 site (“Vendicari” ITA090002). To our knowledge, there are no studies that have ever investigated the germination variability of P. spinosum at intrapopulation level by using a comparative approach among subpopulations. Specifically, all fruits of P. spinosum were collected in July 2022 from five different subpopulations distributed across the study site “B” (“Vendicari” reserve, Fig. 2), along 8-km coastline at 0–30 m a.s.l. Large, robust and healthy subpopulations of P. spinosum were selected, with the highest density and quantity of reproductive output. P. spinosum is an evergreen dwarf shrub that dominates the “Vendicari” coastal garigue together with other species like Thymbra capitata, Cytisus infestus, Teucrium fruticans, Thymelaea irsuta, and Chamaerops humilis. In particular, P. spinosum garigue grows between Pistacia lentiscus and Juniperus macrocarpa scrubland, and rocky halophilous vegetation characterized by Limonium syracusanum.

The strategy for seed collection and post-harvest treatments followed the recommendations of national and international protocols41,42,43,44. Fruit clusters were collected, in each subpopulation, from 25–30 randomly chosen different plants of P. spinosum, 3–5 m far from each other to ensure that they were distinct individuals within a relatively large area (minimum 800 m2), thus obtaining an adequate representation of genetic diversity. Fruits were bulked into labelled paper bags and taken to the laboratory on the same day of the collection. After that, some fruits were pressed, and then a cutter was used to manually extract seeds; the other fruits were left untouched. The average mass of fruits and seeds (g ± SD) was calculated by weighing five replicates of 20 fruits and 20 seeds from each subpopulation (Table 2). The average weight (of 100 fruits and 100 seeds) was determined by using a balance with the accuracy of 0.001 g (Mettler AE 50). The dimensions (length and width) of ten random fruits and seeds were measured through a stereoscopic microscope (Olympus SZX 12 stereo microscope). Regarding the kind of embryo, on the basis of Martin’s46 key for the types of seeds47, the embryo may be considered as spatulate fully developed.

Table 2 Mass, length and width of fruits and seeds in the five subpopulations (subpops) of Poterium spinosum.
Full size table

Germination tests were performed in the laboratories of the Germplasm Bank (BGS-CT) of the Department of Biological, Geological and Environmental Sciences, Catania University. The following pretreatments and experiments were carried out according to the germination tests of the plant species Muscari gussonei45. Before starting the germination tests, all fruits and seeds were cleaned and stored at controlled conditions (20 °C, 40% relative humidity) for about two weeks before use; fruits and seeds that were not used in the germination tests were stored at −18/−20 °C for long-term ex situ conservation; fruits and seeds, used in the experiments, were preliminarily washed with sterile distilled water. All germination tests were carried out in growth chambers under controlled temperature and light conditions; the light in each growth chamber was provided by white fluorescent tubes (Osram FL 40 SS W/37), with photosynthetic photon flux density of 40 μmol m-2 s-1. These experiments considered various temperature regimes, including three fixed (10 °C, 15 °C, 20 °C) and one alternating (10/20 °C) temperatures, and a photoperiod of 12-h light and 12-h dark (12/12 h light–dark); for each experimental condition and for each of the five subpopulations, four replicates of 25 fruits and of 25 seeds were sown on three layers of filter papers, moistened with distilled water in plastic Petri dishes, 9 cm in diameter. Distilled water was added as needed, and parafilm M® was used for wrapping (sealing) the Petri dishes to limit any moisture loss, and reduce contamination. Fruit/seed germination was checked daily, and the criterion to establish germination was the emergence of a 1-mm long radicle43. The experiments were considered finished after 60 days since the start of the germination tests. At the end of the incubation period, the viability of each remaining fruit/seed was estimated by the cut test, and each fruit/seed was classified as viable/fresh, empty or dead 44. Fruit/seed with white and firm embryos were considered viable; empty and deteriorated (dead) fruits/seeds were excluded from the calculation of the final germination percentages (FGP); viable fruits/seeds, which did not germinate at the end of each test, were not further investigated.

Germination parameters

The following parameters were used to investigate the germination process of P. spinosum:

(a) Final germination percentage (FGP)48,

$$FGP= \left(\frac{\sum_{i=1}^{k}{n}_{i}}{N}\right)\times 100$$

where

ni = number of seeds germinated on the ith day;

k = 60, number of days of experiment duration;

N = 25, total number of seeds put in a Petri dish;

(b) Mean germination time (MGT)49,

$$MGT= \frac{\sum_{i=1}^{k}{n}_{i}{t}_{i}}{\sum_{i=1}^{K}{n}_{i}}$$

where

ti = number of days between the start of the experiment and the ith day;

(c) Mean germination rate (MGR), calculated as the reciprocal of MGT50,

$$MGR = {1 \mathord{\left/ {\vphantom {1 {MGT}}} \right. \kern-0pt} {MGT}}$$

(d) First day of germination (FDG), expressed as the day on which the first germination event occurs51;

(e) Last day of germination (LDG), expressed as the day on which the last germination event occurs51;

(f) Median germination time (T50)52,

$${T}_{50}={t}_{i}+\frac{\left(\frac{N+1}{2}-{n}_{i}\right)\times \left({t}_{j}-{t}_{i}\right)}{{n}_{j}-{n}_{i}}$$

where

N = final number of germinated seeds;

ni and nj are the total number of seeds germinated by adjacent counts at time ti and tj , with ni < N/2 < nj;

(g) Coefficient of variation of the mean germination time (CVt)50,

$$CV_{t} = \left( {{{S_{t} } \mathord{\left/ {\vphantom {{S_{t} } {MGT}}} \right. \kern-0pt} {MGT}}} \right) \times 100$$

where the variance of the mean germination time is expressed as:

$${s}_{t}^{2}= \frac{\sum_{i=1}^{k}{{n}_{i}({t}_{i}-MGT)}^{2}}{\sum_{i=1}^{k}{n}_{i}-1}$$

(h) Coefficient of velocity of germination (CVG)53,

$$CVG= \frac{\sum_{i=1}^{k}{n}_{i}}{\sum_{i=1}^{k}{n}_{i}{t}_{i}}\times 100$$

Germination rate index (GR)54,

$$GRI = {{G_{1} } \mathord{\left/ {\vphantom {{G_{1} } 1}} \right. \kern-0pt} 1} + {{G_{2} } \mathord{\left/ {\vphantom {{G_{2} } 2}} \right. \kern-0pt} 2} + \ldots + \left( {1 \times n_{60} } \right)$$

where G1 is the final germination percentage (FGP1) on day 1, G2 is the final germination percentage (FGP2) on day 2, and so on;

(j) Germination index (GI)55,

$$GI = \left( {60 \times n_{1} } \right) + \left( {59 \times n_{2} } \right) + \ldots + \left( {1 \times n_{60} } \right)$$

where n1, n2,…, n60, are respectively the number of seeds germinated on the first, second,…ith day.

Climate patterns: multi-temporal analysis

Air temperature and rainfall are key parameters to describe climate, which is one of the main elements used to define plant niches and species distribution56,57. In particular, this study investigated the multi-temporal trends of temperature and rainfall in the two protected areas examined. Climate analysis followed the methodology of Bonanno and Veneziano45. The climate oscillations of these areas were analysed over a 90-year period, by considering three 30-year time intervals, namely 1931–1960, 1961–1990, 1991–2020, and the whole investigated period of 90 years (1931–2020) (Figs. 4, 5). The raw data used for the processing were recorded by the weather stations located in the towns and villages near the two protected areas. Such data were obtained from the official website of the Sicilian Regional Government58. In particular, a total of 20 weather stations was selected within a radius of 30 km from the sites “Plemmirio” (site A) and “Vendicari” (site B). Each annual mean value of temperature and rainfall—shown in Figs. 4, 5—was obtained through the kriging interpolation of the annual mean values of the selected weather stations. The annual mean values of the weather stations were first arranged in an excel spreadsheet, and then processed with the software ArcGIS version 10.2.

Fig. 4
figure 4

Temperature trends in the protected areas of “Plemmirio” and “Vendicari” during the 30-year periods of 1931–1960, 1961–1990, 1991–2020, and the 90-year period of 1931–2020.

Full size image
Fig. 5
figure 5

Rainfall trends in the protected areas of “Plemmirio” and “Vendicari” during the 30-year periods of 1931–1960, 1961–1990, 1991–2020, and the 90-year period of 1931–2020.

Full size image

CORINE Land Cover: multi-temporal analysis

The CORINE Land Cover (CLC) methodology was applied to investigate the multi-temporal variations of soil-use across the two study areas where P. spinosum is mainly distributed45. CLC III level was specifically used to determine land-cover classes in 1990, 2000, 2006, 2012, 2018 (Figs. 6, 7). The CLC GIS data were provided by SINANET59. As regards land-cover data in 1958, the information was collected from the “map of Italy’s soil-use”60. The correspondence between CNR-TCI classes and CLC classes was shown in Figs. 6, 7. All soil-use data were preliminarily transformed into shapefiles, and then processed through the software ArcGIS version 10.2.

Fig. 6
figure 6

Land-cover changes in the “Plemmirio” protected area (site A) according to CNR-TCI in 1958, and CORINE land cover (CLC) III level in 1990, 2000, 2006, 2012 and 2018 (CLC classes correspondent to CNR-TCI classes are reported in brackets).

Full size image
Fig. 7
figure 7

Land-cover changes in the “Vendicari” protected area (site B) according to CNR-TCI in 1958, and CORINE land cover (CLC) III level in 1990, 2000, 2006, 2012 and 2018 (CLC classes correspondent to CNR-TCI classes are reported in brackets).

Full size image

Statistical analysis

The Wilcoxon signed ranks test was used when two related samples were compared. In case of unrelated pairs of samples, the Mann–Whitney U-test was carried out. The Friedman test was instead used when the comparison of multiple related samples was needed. The Kruskal–Wallis H-test was carried out when multiple unrelated samples were compared. To detect significant differences between pairs, contrasts were carried out with the Wilcoxon signed ranks test (related pairs) and the Mann–Whitney U-test (unrelated pairs). We specifically used Kruskal–Wallis H-test and Mann–Whitney U-test to check whether each morphometric parameter of P. spinosum (seeds and fruits) varied significantly among the five studied subpopulations (Table 2). These same tests were also used to verify whether each germination parameter, at the same temperature conditions, was statistically different among the five subpopulations of P. spinosum (Tables 3 and 4). The Friedman test and the Wilcoxon signed ranks test were instead carried out to ascertain whether each germination parameter changed significantly under different temperature conditions and within the same subpopulation (Tables 3 and 4). Similarly, we conducted Friedman and Wilcoxon tests to show whether each CORINE Land Cover class varied significantly over different time intervals (Table 5). Conducting multiple contrasts may increase the Type I error rate. The initial level of risk (α = 0.05) was therefore adapted according to the Bonferroni formula αB = α/k, where αB is the adapted level of risk, and k is the number of comparisons45. The Kendall rank correlation coefficient was applied to check for significant monotonic correlations. The degree of significance was set at 0.05. Statistical processing was performed with the statistical package IBM SPSS Version 27.0.

Table 3 Germination parameters of Poterium spinosum fruits.
Full size table
Table 4 Germination parameters of Poterium spinosum seeds.
Full size table
Table 5 Surfaces (ha) of CORINE land-cover classes in the areas of “Plemmirio” (site A) and “Vendicari” (site B) in 1958, 1990, 2000, 2006, 2012 and 2018.
Full size table

Results and discussion

Germination response of Poterium spinosum

The germination parameters of P. spinosum showed a general pattern according to which fruits and seeds from distinct subpopulations (subpops) respond differently to diverse temperatures (Tables 3, 4). Specifically, the highest values of FGP were found at 15 °C for the fruits of subpops A, C, D, E, whereas the fruits of subpop B reported their greatest FGP at 10 °C. In turn, the seeds of subpops B, C, D, E reached the highest FGP at 20 °C, except for subpop A, which did so at 15 °C. Seeds showed generally higher values of FGP compared to fruits, and at higher temperatures, namely 20 °C in seeds versus 15 °C in fruits. In particular, the highest FGP in seeds was 70% (subpop C at 20 °C), whereas FGP = 58.2% was the highest value in fruits (subpop C at 15 °C). The results of the other germination parameters further corroborated the significant intrapopulation variability of P. spinosum. For instance, the lowest mean values of MGT in fruits were found in subpops A (10 days) and C (15.9) at 20° C, B at 10/20 °C (12.7), D at 10 °C (16.3), and E at 15 °C (17.4). In seeds, the trend of MGT was (lowest mean values): 14.4 days in subpop A, 12.3 in B and 13 in C at 15 °C, and 14.3 in D and 14.1 in E at 20 °C. Similarly, T50 showed significant variability in fruits and seeds among distinct subpopulations at different temperatures. In particular, the lowest mean values of T50 in fruits were 11 days in subpop A, 12 in C, 13.5 in D and 16 in E at 15 °C, whereas 13.5 days in subpop B at 10/20 °C. In seeds, the lowest mean values of T50 were instead 9.8 days in subpop A at 15 °C, 11.2 in B, 10.5 in C and 11.5 in E at 20 °C, as well as 12.8 days in subpop D at 10/20 °C.

This study showed that the optimum germination temperatures of P. spinosum are mainly in the range of 15–20 °C, in general agreement with previous authors. Santo et al.19 found that P. spinosum seeds reach the highest FGP in the range of 10–20 °C, in different populations from several Mediterranean countries. Similarly, Meloni61 found that 20 °C is the most favorable temperature for P. spinosum germination in populations from Sardinia, Greece and Malta. Lantieri et al.18 investigated instead populations from Sicily, and identified the range of 15–20 °C as the optimum condition for germination. The results of this study corroborated previous research according to which P. spinosum shows the “Mediterranean germination syndrome” that is typical of Mediterranean coastal species62,63, whose germination occurs at low and moderate temperatures (10–20 °C). All these findings suggest that the field germination of P. spinosum is especially favored between autumn and early spring62,64, when the higher water availability may further stimulate seed germination. This study, in particular, showed that germination optimum temperatures are different, namely 15 °C in fruits and 20 °C in seeds. Moreover, seeds showed higher values of germinability compared to fruits (seeds+spongy tissues). These results suggest that spongy tissues may reduce both temperature optimum and germinability. Vahl14 noted that seeds stripped of fruit spongy tissues germinated faster and with greater percentages, suggesting the presence of an inhibitor in the spongy material. Similarly,16 detected a slower germination time in fruits than seeds. Santo et al.19, however, found that the influence of spongy tissues may vary among different populations: for some of them it caused a remarkable reduction in germinating seeds, but for others the effects were not significant. Further light should be shed on the possible presence of an inhibitor in the spongy tissue of P. spinosum fruits, and the associated effects.

Previous studies found a great germination variability at interpopulation level in P. spinosum, thus suggesting that Mediterranean populations of P. spinosum are considerably different among them and highly adapted to local habitat conditions19,61. Latitude, altitude, temperature, rainfall, light, moisture, soil nutrients, proximity to the sea and habitat disturbance are among the numerous factors linked to interpopulation variability of germination 47. This study, however, found that P. spinosum can show significantly different germination patterns also within the same population. If, on the one hand, different abiotic conditions may explain different germinability among populations of P. spinosum from ecologically diverse Mediterranean regions, on the other hand, as shown in this study, a different germination behavior can be also found among subpopulations of P. spinosum distributed along a narrow coastal strip, environmentally homogeneous. The significant germination variability at subpopulation level of this Sicilian population (“Vendicari”, site B) may be associated with the peripheral position of Sicily in the western geographical distribution of P. spinosum. Peripheral populations are usually influenced by evolutionary divergence processes31, which could differentiate them not only from other populations but also within the same population. In particular, several factors such as maternal reproductive phenology and microenvironment, as well as genotype, can be responsible for the variation of germination response among individuals of the same population65,66,67. Germination variability may also exist at intrapopulation level as a bet-hedging strategy for enabling the species to produce numerous seeds optimized for different climatic conditions68. As a result, this intrapopulation variation increases the probability of generational survival in an unpredictable or changing environment21,69. High intrapopulation variability can be thus considered as an adaptation strategy that can increase the reproductive success of P. spinosum under climate and land-use changes. The results of this study shed further light on P. spinosum seed ecology, and therefore contributed to develop new germination protocols than can be used for conservation and restoration programs across a global climate and biodiversity hotspot: the Mediterranean region.

Multi-temporal trends of temperature and rainfall in the study areas of P. spinosum

Climate patterns reported rising temperature and declining rainfall over a 90-year period in site A (Plemmirio reserve) and site B (Vendicari reserve) (Figs. 4 and 5). The values of temperature and rainfall, in particular, showed oscillating values over the three 30-year subperiods. Regarding temperature, in the period 1931–1960, the average values were relatively constant, within the range of 17.1–17.2 °C, whereas in the following period of 1961–1990, the average temperature increased significantly by 0.9 °C (17–17.9 °C). Over the last interval of 1991–2020, the average values declined moderately by 0.3 °C (18.3–18 °C). Overall, in the whole period of 1931–2020, the average temperature increased significantly by 1.5 °C, specifically from 16.8 to 18.3 °C. Regarding rainfall, the oscillations were instead more moderate. In the first period of 1931–1960, the average values ranged from 710 to 790 mm, marking a significant increase of 80 mm. However, in the following two 30-year periods, rainfall showed declining trends: during 1961–1990, the average values decreased from 620 to 580 mm, whereas during 1991–2020, from 700 to 690 mm. In the whole 90-year study period, from 1931–2020, the average rainfall decreased from 710 to 650 mm.

Climate change is already impacting biodiversity and is likely to intensify over the next few decades unless practical mitigation efforts are implemented70. The Mediterranean region, in particular, is not only a hotspot of climate changes7, but also a hotspot of biodiversity, characterized by an endemic flora expected to experience severe global warming71,72. Rising temperatures and declining rainfall can indeed seriously compromise the germinability of most Mediterranean species, especially those with a narrow germination optimum such as P. spinosum, whose best conditions for germination are moderate temperatures at 15–20 °C. In this study, no mitigation trends were found for the climate conditions affecting P. spinosum range, where temperature increase by 1.5 °C during 1931–2020 is equivalent to 0.17 °C/decade, in line with current global warming increase at about 0.2 °C/decade73. According to IPCC70, during 2016–2035, the average global temperature is expected to increase between 0.3 and 0.7 °C, and global warming is likely to reach 1.5 °C between 2030 and 2052. Similarly to this study, in most areas of the Mediterranean region, precipitation is predicted to decrease (− 12%), particularly in summer (− 24%) 74.

Concerns associated with variations in global temperature, rainfall and other climate variables encouraged research addressing the impact of climate changes on species distribution57,75,76. Geographic range shifts, expansions and contractions are indeed some of the main responses of species to climate changes77. In particular, species with wide geographic ranges are expected to be theoretically less vulnerable, as they may find refugia within their distributional area77. However, extreme variability in annual and seasonal patterns of temperature and precipitation can disrupt a wide range of natural processes, especially when changes occur more quickly than species adaptation56,70. As a result, the peculiarities of Mediterranean climate may restrict plant adaptive plasticity. Mediterranean climate can indeed be highly unpredictable, and this can reduce plasticity if environmental signals that trigger adaptive responses are unstable, or maintaining induced phenotypes proves biologically expensive under changeable climatic conditions78,79,80. However, this study found significant differences in the germination behavior of five subpopulations of P. spinosum, thus suggesting both significant within-population genetic diversity, and potential plasticity not only at interpopulation but also at intrapopulation level. Various studies found evidence for significant within-population evolutionary potential for functionally important traits in several Mediterranean species81,82. A highly different germination behavior at intrapopulation level, as found in this study for P. spinosum, may be taken as evidence for the existence of the underlying genetic variation necessary for an effective response to strong selection pressures imposed by climate changes.

Despite the consequences for species survival under global changes83,84, overall, the level of adaptive plasticity and evolutionary potential is poorly known in Mediterranean plants85,86. In particular, the rate at which environmental conditions change can also constrain adaptive evolution if species are not able to track changes, despite the presence of genetic variation84,87. Long-term demographic surveys are also needed to understand how past climatic variations affected population dynamics, and to predict population viability under climate changes88,89,90. Further studies are thus necessary to shed light on how patterns of plasticity vary among and within populations for different traits and ecological factors, and whether current eco-physiological processes will still be adaptive under future environmental conditions.

Land-cover multi-temporal trends in the study areas of P. spinosum

CORINE land-cover (CLC) classes, 3rd level, were reported for the period 1958–2018, in the study areas of “Plemmirio” (site A) and “Vendicari” (site B) (Figs. 6, 7; Table 5). In site A, artificial areas ranged between 18.7% in 1990–2000, and 21.2% in 2012–2018, whereas site B showed a much lower artificial surface, which was only 1.3% in 1990–2000, and 2.0% in 2018. The “discontinuous urban fabric” (1.1.2.) was the dominant CLC class in site A, and the only one in site B. Both site A and site B showed instead a complex and variegated soil-use mosaic mainly characterized by several agricultural classes. Specifically, the total agricultural surface of site A was 93.9% in 1958, and then declined progressively until the stable value of c. 76% during 2012–2018. Similarly, site B showed a very high agricultural surface, which was 98.3% in 1958, and then stabilized at c. 91% in the period 2012–2018. The dominant agricultural classes were “non-irrigated arable land” (class 2.1.1.) and “fruit trees and berry plantations” (class 2.2.2.) in sites A and B, the latter of which was also characterized by “annual crops associated with permanent crops” (class 2.4.1.). Regarding forests and semi-natural areas, although site A showed an appreciable value of 7.7% in the decade 1990–2000, this figure dramatically declined to 2.5% in the period 2012–2018. The trend of site B was similar, with an initial value of 7.3% during 1990–2000, and a final surface of 5.1% in 2012–2018. Sclerophyllous vegetation was the main natural area in both study sites (class 3.2.3.). Overall, salt marshes (4.2.1.) and coastal lagoons (5.2.1.) were negligible respectively in site A (0.5%) and site B (1.9%).

Land-cover transformation as a result of agricultural expansion is one of the most serious drivers of loss and fragmentation of natural habitats, with consequent strong decline in biodiversity91,92,93. Agricultural expansion may especially cause the fragmentation of agricultural landscapes, which become complex, heterogeneous and discontinuous94. Agricultural landscape fragmentation affects biodiversity by impacting on the distribution, migration and movement of species95,96. In particular, by disrupting pollination and diaspore dispersal processes, landscape fragmentation increases genetic drift and inbreeding, thus reducing genetic diversity within plant populations97,98. The consequences of habitat fragmentation for Mediterranean plants are many. Firstly, fragmentation can decrease individual plant fitness and increase population extinction risk due to environmental and demographic stochasticity99,100,101. Secondly, fragmentation can inhibit plant migration if suitable habitat patches are not sufficiently connected to allow gene flow via pollen and seeds102. Thirdly, habitat fragmentation can have interactive and indirect effects on the adaptive potential of plants to other global transformation drivers such as climate changes103,104. This study, in particular, showed that, in Sicily, P. spinosum has a fragmented distribution globally and locally (Figs. 1, 2). This scattered range of P. spinosum may not only reduce genetic variation for fitness-related traits, but also reduce genetic variation for plasticity. Although this hypothesis has been poorly investigated for Mediterranean plants105, the significantly different germination behavior of P. spinosum at subpopulation level suggests that this plant may keep a certain level of plasticity within populations distributed across highly fragmented and impacted agricultural landscapes. Future studies, however, are needed to shed further light on genetic variation and plasticity in P. spinosum along fragmentation gradients, and how these traits can influence the response of Mediterranean plants to climate and land-cover changes.

Landscape changes often lead to habitat fragmentation, affecting both structure and function through loss of original habitat, reduction in habitat patch size, and increasing isolation of patches106,107. Habitat fragmentation is consequently a key conservation concern in many countries, and is strongly associated with loss of biodiversity108,109. In particular, the scattered layout of landscape patches affects the connectivity of the landscape itself, hinders the interaction between different patches, and cause degradation in the quality of natural habitats110. Ecological networks are increasingly accepted as proactive tools for preserving biodiversity by improving landscape connectivity111,112. Conservation on a landscape scale is especially recommended when the populations of a species are highly fragmented across a vast territory such as P. spinosum, whose populations are scattered across south-east Sicily (Fig. 1). Increasing landscape connectivity should be therefore considered as one of the main guiding principles to re-establish the genetic connections within the P. spinosum metapopulation in Sicily.

Conclusions

The highly different germination behavior of P. spinosum at subpopulation level suggests that this species may be characterized by a significant intrapopulation genetic variability. Although this trait is poorly known for Mediterranean plants, it is undeniable that this possible genetic resilience of P. spinosum is encouraging in the face of the ongoing severe climate and land-cover changes. However, the distribution ranges of P. spinosum are generally scattered both locally and across the Mediterranean region. To support genetic fluxes at interpopulation level, at least on a landscape scale, conservation measures should aim to increase the ecological connectivity among highly fragmented natural habitats. P. spinosum communities—the so called “phrygana” or thorny garigue—are distinctive features of Mediterranean landscapes. As a result, their protection and restoration should be a priority for any organization in charge of biodiversity conservation. The germination tests of this study may also contribute to the conservation of P. spinosum. These lab experiments showed indeed that fruits and seeds of P. spinosum not only show different germination optimum but also different germinability. This different germination behavior of fruits and seeds should be considered as an important starting point for any conservation program of P. spinosum, both in situ and ex situ.

Plant materials statement

The permission to collect specimens of Poterium spinosum was granted by the Regional Government of Sicily, which is the managing body of the “Vendicari” Natural Reserve. Plant collection and use were in accordance with all the relevant guidelines. Specifically, in compliance with the collection permission, all specimens of P. spinosum were < 10% of the sampled populations. Voucher specimens were deposited in the public herbarium of the Botanical Garden of Catania University (Italy). Specimens were collected, identified and deposited by the authors of this article.