|Home | About | Journals | Submit | Contact Us | Français|
In a future warmer subarctic climate, the soil temperatures experienced by dispersed seeds are likely to increase during summer but may decrease during winter due to expected changes in snow depth, duration and quality. Because little is known about the dormancy-breaking and germination requirements of subarctic species, how warming may influence the timing and level of germination in these species was examined.
Under controlled conditions, how colder winter and warmer summer soil temperatures influenced germination was tested in 23 subarctic species. The cold stratification and warm incubation temperatures were derived from real soil temperature measurements in subarctic tundra and the temperatures were gradually changed over time to simulate different months of the year.
Moderate summer warming (+2·5 °C) substantially accelerated germination in all but four species but did not affect germination percentages. Optimum germination temperatures (20/10°C) further decreased germination time and increased germination percentages in three species. Colder winter soil temperatures delayed the germination in ten species and decreased the germination percentage in four species, whereas the opposite was found in Silene acaulis. In most species, the combined effect of a reduced snow cover and summer warming resulted in earlier germination and thus a longer first growing season, which improves the chance of seedling survival. In particular the recruitment of (dwarf) shrubs (Vaccinium myrtillus, V. vitis-idaea, Betula nana), trees (Alnus incana, Betula pubescens) and grasses (Calamagrostis lapponica, C. purpurea) is likely to benefit from a warmer subarctic climate.
Seedling establishment is expected to improve in a future warmer subarctic climate, mainly by considerably earlier germination. The magnitudes of the responses are species-specific, which should be taken into account when modelling population growth and migration of subarctic species.
In Arctic and subarctic regions, which are expected to be affected more by anthropogenic climate change than other regions in the world (ACIA, 2004), the effects of climate warming on germination have hardly been studied (but see Wookey et al., 1995; Molau and Shaver, 1997; Graae et al., 2008). However, the timing and level of germination strongly affect a plant's recruitment success and may consequently have implications for species migration. Therefore, research on how a warmer climate affects germination is important for our ability to predict population dynamics and future distributions of subarctic species (Higgins et al., 2003; Neilson et al., 2005; Pearson, 2006; Thuiller et al., 2008).
In northern ecosystems, climate warming is not only expected to result in significant increases in temperature, but also changes in the timing and reduction in the depth and duration of the snow cover are projected to occur (ACIA, 2004; Phoenix and Lee, 2004; Keller et al., 2005; IPCC, 2007). The extent of snow cover over Arctic land areas has declined by about 10 % over the past 30 years and model projections suggest that it will decrease an additional 10–20 % before the end of the century (ACIA, 2004). Snow is an excellent insulator (Sturm et al., 2005); therefore a warmer climate is likely to result in colder soils and deeper soil frost during the winter when the soil is not insulated by snow (Groffman et al., 2001; Venäläinen et al., 2001; Stieglitz et al., 2003; Öquist and Laudon, 2008). Moreover, important snow quality changes are projected; for instance, the development of hard packed snow due to changing wind patterns and ice layer formation, due to an increase in thawing and freezing in winter (ACIA, 2004), resulting in less insulation and thus colder soils (Körner, 2003). This may have important implications for recruitment from seed, because in subarctic environments, where there is hardly any current season (pre-winter) germination (Körner, 2003), winter temperatures possibly influence seed germination (Vleeshouwers et al., 1995; Baskin and Baskin, 1998).
In cold climate species, experiencing winter (cold stratification) usually increases germination percentages (Probert, 2000; Körner, 2003; Giménez-Benavides et al., 2005) and often contributes to reducing the temperature needed for later germination (Reynolds, 1984; Densmore, 1997; Shimono and Kudo, 2005). Cold stratification studies of Arctic and alpine species have, however, mainly focused on the duration of the stratification period (e.g. Baskin et al., 2000; Cavieres and Arroyo, 2000; Schütz, 2002) rather than on the stratification temperature. This is usually selected close to 5 °C because this is supposed to be optimal for dormancy-breaking in many species (Bewley and Black, 1994). Some recent studies have used more realistic stratification temperatures close to 0 °C for cold climate species (e.g. Baskin et al., 2000; Schütz, 2002; Shimono and Kudo, 2005; Graae et al., 2008), but this only reflects situations where a thick snow cover provides insulation. The effect of colder winter soil temperatures, reached in the absence of snow, has to our knowledge only been tested in two experiments with stratification under field conditions (Baskin et al., 2002; Graae et al., 2008). However, these included only two species, Empetrum hermaphroditum and Vaccinium uliginosum, and no effects of stratification temperature were observed.
The germination temperature in Arctic and alpine plants is relatively high in comparison with ambient temperatures (Cavieres and Arroyo, 2000; Körner, 2003), which has been thought to be an adaptation to avoid germination in early spring or autumn when the probability of frost is high (Billings and Mooney, 1968; Cavieres and Arroyo, 2000). Further, germination percentages are often positively correlated with the temperature under which seeds are incubated (Baskin and Baskin, 1998; Graae et al., 2008). However, apart from these general trends, detailed knowledge on the germination requirements of (sub)arctic species is lacking (Baskin and Baskin, 1998). This is mainly because most studies use optimal incubation temperatures for testing germinability, often 20/10 °C or 25/15 °C (for a review, see Baskin and Baskin; 1998), independent of the species' habitat. The use of more realistic temperatures adapted to the environment of the species or the application of gradually changing temperature regimes simulating each month of the growing season is still scarce (but see Baskin et al., 1995; Thompson and Naeem, 1996), although essential to apply germination data from laboratory studies in an ecological context. Moreover, the importance of the timing of germination has often been overlooked, despite its significance for plants from harsh environments where the growing season is highly limited (Schütz, 2002; Deines et al., 2007).
To improve our knowledge of how climate warming may affect germination in subarctic ecosystems, an experiment was set up in which the effects of two realistic cold stratification treatments and three warm incubation treatments were tested. The temperature regimes were derived from real soil temperature measurements in subarctic tundra and were gradually changed over time to simulate different months of the year. As stratification temperatures, winter soil temperatures were used from snow-covered and exposed habitats, the latter to simulate a reduced snow cover in a future warmer climate. The incubation temperatures corresponded to (a) current spring and summer temperatures in subartic tundra, (b) temperatures that are 2·5 °C higher than that, and (c) optimum germination temperatures (20/10 °C). To unravel general patterns, seeds from 20 subarctic species belonging to different functional types were used. Also three boreal tree species were added to test if their current expansion is limited by unsuccessful germination under present subarctic temperatures.
Specific research questions were: (a) How are germination percentage and germination time affected by colder winter soil temperatures, occurring under a reduced snow cover? (b) Does a higher germination temperature (+2·5 °C) affect germination of subarctic species, and if so, how? (c) Are the effects general or species specific? (d) Is the range expansion of boreal tree species currently limited by too-low germination temperatures in subarctic tundra?
Seeds of 20 species were collected in the surroundings of the Abisko Scientific Research Station (68°21′N, 18°49′E) in subarctic northern Sweden (Table 1). The species were selected to belong to different functional types and were important components of the plant communities in the study area. Additionally, seeds of Alnus incana and Betula pubescens, tree species with a boreal distribution, were collected close to Umeå (63°50′N, 20°20′E) and seeds of Pinus sylvestris were ordered from a seed company and originated from Karesuando (68°20′N, 21°53′E). All seeds (except from P. sylvestris) were collected between 15 August and 15 September 2007 and stored dry at room temperature until sowing (6 November 2007). Seeds of fleshy fruits were extracted from the fruits immediately after collection. The seed mass of each species, without any attachments, was determined by weighing four times 30 air-dry seeds (ten seeds for Pinus sylvestris and Vicia cracca because of their large seed size).
The seeds were subjected to two contrasting cold stratification treatments for 20 weeks. The stratification temperatures (Table 2A) were based on winter soil temperature measurements (3 cm depth) during 2006 and 2007 in subarctic tundra sites with either a thick insulating snow cover throughout winter (meadow vegetation) or a strongly reduced snow cover (wind exposed poor dwarf shrub heath). The measurements were done approx. 8 km from the Abisko Scientific Research Station at 850 m a.s.l. (B. J. Graae, unpubl. res.). Cold stratification occurred in complete darkness.
After stratification, the seeds were allowed to germinate under three different incubation treatments: optimum germination temperatures (20/10 °C) (Baskin and Baskin, 1998), current spring and summer soil temperatures in subarctic tundra (control), and temperatures that are 2·5 °C warmer than that (+2·5 °C; Table 2B). The temperature increase of 2·5 °C is in accordance with the projected temperature rise in Arctic regions (60–90°N) by the year 2050 relative to 1981–2000 (ACIA, 2004). The control incubation temperatures were derived from soil temperature measurements during the spring and summer of 2006 and 2007 in dwarf shrub heath, the dominant vegetation type at the earlier-mentioned location. To simulate the subarctic summer, seeds were kept in 24 h of daylight during incubation, but light intensities were alternated every 12 h coinciding with the temperature intervals. During ‘night’, light was provided with a photosynthetic photon flux density (PPFD) of 25 µmol m−2 s−1 (400–700 nm), whereas during daytime a PPFD of 45 or 110 µmol m−2 s−1 was provided every other day, to simulate overcast and clear days. In the ‘optimum’ treatment, the daytime PPFD was always 110 µmol m−2 s−1.
For each combination of stratification and incubation treatment, four replications of 30 seeds (ten seeds for P. sylvestris and V. cracca) were sown on moist commercial pot soil in 90-mm-diameter Petri dishes on 6 November 2007. After sowing, the dishes were wrapped with parafilm to reduce loss of water and they were subsequently put in incubators for the cold stratification treatments.
During incubation, the seeds were checked weekly for germination (protrusion of the radicle) and the germinated seeds were removed to reduce counting time in subsequent weeks. After 10 weeks, most species had completed germination but one final germination counting was done after 13 weeks of incubation. The final germination percentages were arcsine transformed to improve normality and stabilize variances. Species with <5 % germination (Carex rostrata, C. saxatilis and Empetrum hermaphroditum) were excluded from further analyses (Table 1).
In addition to percentage germination, the mean germination time (MGT) for each of the species was determined in order to distinguish fast from slowly germinating species and to determine how germination speed was affected by the treatments. It was calculated as
where ni is the number of seeds that germinated within consecutive intervals of time, ti the time between the beginning of the test and the end of a particular interval of measurement, and N the total number of seeds that germinated (Deines et al., 2007).
Significant differences in germination percentages were observed between the functional types (ANOVA, F5,474 = 16·0, P < 0·001), with the poorest germination in the grasses (20 %) and the highest germination in the forbs (56 %; Fig. 1A). Germination percentage slightly increased with seed mass (linear regression, F1,478 = 30·83, P < 0·001, R2 = 0·06).
Germination percentages were significantly affected by the incubation and stratification treatments. In general (all species combined), the ‘thick snow cover’ stratification treatment resulted in more germination (43 %) than the ‘thin snow cover’ treatment (39 %; Fig. 2A and Table 3). In addition, germination percentages were higher in the ‘optimum’ (45 %) incubation treatment than in the ‘ + 2·5 °C’ (39 %) and ‘control’ (39 %) treatments (no difference between +2·5 °C and control; Fig. 2A). The lack of a significant interaction effect between incubation and stratification (Table 3) suggests that the effects of both are additive, which is supported by the highest germination percentage when ‘thick snow cover’ stratification was followed by ‘optimum’ incubation (Fig. 2A).
Because of strong species × incubation and species × stratification interaction effects (Table 3) species-specific analyses were also required. These demonstrated that stratification treatment significantly affected germination percentage in five out of the 20 species, either forbs or dwarf shrubs (Fig. 3). In E. angustifolium, Silene dioica, Vaccinium myrtillus and V. uliginosum germination percentages were higher in the ‘thick snow cover’ than the ‘thin snow cover’ treatment. Silene acaulis, on the other hand, performed better after ‘thin snow cover’ stratification (see Table 4 for mean differences in germination percentages).
Significant effects of incubation treatment on germination percentage were found in S. acaulis (+15 %), V. myrtillus (+67 %), V. uliginosum (– 15 %) and V. vitis-idaea (+14 %) when the ‘optimum’ incubation treatment was compared with the ‘control’ incubation treatment (Fig. 3). Contrary to the other species, germination percentage decreased with increasing incubation temperature in V. uliginosum, but only in the ‘thin snow cover’ stratification treatment (significant incubation × stratification interaction; Fig. 3). In none of the species did a temperature increase of 2·5 °C (‘ + 2·5 °C’ versus ‘control’ incubation treatment) result in higher germination percentages (Fig. 3 and Table 4).
The speed of germination differed significantly between the functional types (ANOVA, F5,447 = 29·0, P < 0·001). The dwarf shrubs (41 d) and the legumes (40 d) germinated significantly slower than the other functional types, and the forbs (19 d) were the faster germinating group (Fig. 1B). Large-seeded species germinated a little slower than species with smaller seeds (linear regression, F1,451 = 4·53, P = 0·034, R2 = 0·01).
MGT was significantly affected by the incubation and stratification treatments (Table 3). In general (all species combined), the germination time was shorter after ‘thick snow cover’ (27 d) than after ‘thin snow cover’ (30 d) stratification (Fig. 2B). Further, germination was fastest in the ‘optimum’ incubation treatment (16 d), followed by the ‘ + 2·5 °C’ (31 d) and the ‘control’ (39 d) incubation treatments (Fig. 2B). A lack of interaction between incubation and stratification (Table 3) resulted in the shortest MGT when ‘thick snow cover’ stratification was followed by ‘optimum’ incubation (14 d) and the longest MGT when ‘thin snow cover’ stratification was followed by ‘control’ incubation (40 d).
Analyses on individual species level revealed significant stratification effects on MGT in eleven species, belonging to all functional types except the grasses (Fig. 4). Whereas ‘thin snow cover’ stratification usually delayed germination, the opposite was observed in S. acaulis. The strongest effects occurred in S. dioica and V. cracca, in which germination was on average delayed by 10 d and 12 d, respectively, after ‘thin snow cover’ stratification (Table 4).
In all species except the legumes (Astragalus frigidus and V. cracca), incubation temperature significantly affected MGT (Fig. 4). Germination was slowest in the ‘control’ treatment and was faster in the ‘ + 2·5 °C’ and the ‘optimum’ incubation treatments, but in Deschampsia flexuosa and Salix glauca no difference was seen between the ‘control’ and the ‘ + 2·5 °C’ treatment. In D. octopetala, E. angustifolium and V. uliginosum, the strength of the stratification effect depended on the incubation treatment (significant incubation × stratification interaction), whereas for the other species the effects of incubation and stratification were additive (Fig. 4). In ten species, germination occurred at least 1 week earlier in the ‘ + 2·5 °C’ than in the ‘control’ treatment; in V. myrtillus and V. vitis-idaea MGT was even reduced by 2 weeks (Table 4).
In the majority of species, the positive effect of warming (+2·5 °C) on germination time was diminished if the seeds were first stratified in the ‘thin snow cover’ treatment (Table 4). Eight species, though, still showed a reduction in MGT by at least 1 week. In E. angustifolium, V. myrtilllus and V. uliginosum, the advantage of faster germination was counterbalanced by reduced germination percentages after ‘thin snow cover’ stratification. Silene acaulis, on the other hand, showed both an increased germination percentage and faster germination in the scenario of warming combined with a reduced snow cover. Silene dioica and V. cracca suffered most from the future scenario: V. cracca with a strongly delayed germination (+12 d) and S. dioica with delayed germination (+3 d) combined with a reduced germination percentage (−15 %).
To examine how germination was influenced by the different temperature steps in each of the incubation treatments, the accumulated germination percentages (averages of the two stratification treatments) were plotted over time for two species (Fig. 5). In B. nana, germination started during the first week of incubation and almost immediately reached its maximum percentage in the ‘optimum’ incubation treatment (Fig. 5A). In the ‘control’ and ‘ + 2·5 °C’ treatments, germination started under June temperatures and reached maximum values under summer temperatures. Despite the strong delays in germination under colder incubation temperatures, final germination percentages were similar in all treatments. This pattern was observed in the majority of species (not shown). A different pattern was observed in V. myrtillus (Fig. 5B) and V. vitis-idaea (not shown), in which germination occurred only after summer temperatures had been reached in the ‘control’ and ‘ + 2·5 °C’ treatments. In the ‘optimum’ treatment, germination started under June temperatures and reached a higher final germination percentage than in the colder treatments, but percentages in the ‘ + 2·5 °C’ and ‘control’ treatments were similar, after an initial delay. Figure 5 clearly shows that germination is strongly temperature regulated, with no germination until a certain temperature threshold has been reached.
The results demonstrate strong effects of stratification and incubation temperatures on the timing and level of germination in a large set of subarctic plant species. Although the experiment carried out did not include all abiotic and biotic changes that go along with a warmer climate in a natural environment, we believe it is justified to focus on temperature because this factor has several times been proven to be the most important environmental variable regulating the dormancy state and germination of seeds (e.g. Roberts, 1988; Vleeshouwers et al.; 1995; Probert, 2000).
In earlier studies on the influence of snow cover and freezing temperatures on subsequent germination, no significant effects were found. In a study by Graae et al. (2008), there was no difference in germination percentage between seeds of E. hermaphroditum and V. uliginosum stratified at 0·5 °C in incubators and seeds experiencing colder winter temperatures outdoors at boreal (−0·5 to −2 °C) and Arctic (mean winter temperature of −6·9 °C) sites. Similarly, Baskin et al. (2002) could not detect any difference between full- and half-snow-cover treatments on the germination of E. hermaphroditum. However, it was found that colder winter soil temperatures delayed the germination in ten species and decreased the germination percentage in four species, whereas the opposite was found in S. acaulis. The results are thus the first to demonstrate that the projected reductions in snow depth and duration in (sub)arctic and alpine regions (ACIA, 2004; Keller et al., 2005) may have important implications for seed germination, both by affecting the number of germinating seeds and the timing of germination.
Regarding summer warming, it was observed that even relatively small increases in temperature (+2·5 °C) strongly reduced the germination time in all but four species. The lack of a response in the legumes is possibly due to their physical dormancy, which is less temperature dependent than physiological dormancy (Baskin and Baskin, 1998; Probert, 2000). Contrary to expectations (Baskin and Baskin 1998; Graae et al., 2008), a positive relationship between incubation temperature and germination percentage was not commonly observed. This occurred only in three species when the optimum treatment was compared with the control treatment; otherwise final germination percentages were similar. Lower incubation temperatures thus mainly resulted in delayed germination, rather than in lower percentages. Because germination studies are usually finished after 4 weeks (for a review, see Baskin and Baskin, 1998), we think that the frequently reported lower germination percentages at lower incubation temperatures may often be an artefact of the experimental procedure. Indeed, if the present experiment had finished after 4 weeks, a similar report on the different germination percentages between the incubation treatments would have been made (see Fig. 5A). Therefore, we stress the importance of continuing germination trials until germination percentages stop increasing.
Most species were able to germinate under June temperatures in the control treatment, suggesting that a temperature of 10·5/2 °C (during several weeks) is generally high enough for germination in subarctic species, although germination occurred much faster at slightly higher temperatures. As suggested by Körner (2003), a great functional variability in the germination behaviour of the species was found. For instance, the high-altitude species S. acaulis and D. octopetala started germinating during incubation at 6·5/3 °C, whereas V. myrtillus and V. vitis-idaea needed 12·5/4·5 °C for several weeks before the onset of germination, suggesting that species from higher elevations may be better adapted to germinate under low-temperature conditions. This contradicts the idea that seeds from higher elevations need higher germination temperatures than those from lower altitudes to avoid germination in early spring when the probability of frost is still high (Billings and Mooney, 1968; Cavieres and Arroyo, 2000). In addition, S. acaulis was the only species that benefited from the ‘thin snow cover’ stratification treatment, both with a higher germination percentage and a shorter germination time. Apparently it is especially well adapted to perform well in the harsh wind-swept habitats where it mostly grows. The protective nature of the dense cushions in which it occurs (Körner, 2003) may enable its seedlings to overcome frost and drought without much damage. The other high-altitude species, D. octopetala, did not show the same adaptation, which may explain why this species reproduces predominantly by means of clonal growth (Wookey et al., 1995).
In the majority of species, the combined effect of a reduced snow cover and summer warming resulted in faster germination, but four species were not affected and two species showed slower germination. The effect on germination percentage was positive in one species, negative in four species and not significant in all others. In subarctic ecosystems, the short growing season (approx. 3 months in the tundra near Abisko; Molau et al., 2005) is a major barrier for the survival of seedlings because it constrains the period during which seedlings need to attain a critical biomass and acquire resistance to freezing to withstand the harsh and long-lasting winter conditions (Maruta, 1994; Stocklin and Baumler, 1996; Schütz, 2002). Seedling mortalities in cold climates are usually high and 12-month losses often exceed 50 % in large-seeded and 99 % in small-seeded species (Jolls and Bock, 1983; Körner, 2003). This implies that even relatively small reductions in germination time may have substantial effects for recruitment from seed by improving the chance of seedlings to survive the following winter (Chambers, 1995). Consequently, the present results suggest that a warmer subarctic climate is likely to be beneficial for seedling establishment in most species, even though colder soil temperatures during winter may diminish the positive effect.
The species that may benefit most are trees (A. incana and B. pubescens), (dwarf) shrubs (V. myrtillus, V. vitis-idaea and B. nana) and grasses (C. lapponica and C. purpurea), in addition to the forb S. acaulis. These species showed considerably earlier germination (>1 week), sometimes combined with an increased germination percentage (especially in S. acaulis) when stratified and allowed to germinate under future conditions. Vaccinium myrtillus, however, may only improve its recruitment in more protected, snow-rich places (strong negative effect of colder winter temperatures), but its germination percentage is expected to increase substantially if summer temperatures increase >2·5 °C. It has already been documented that dwarf shrubs, grasses and trees will expand in Arctic and subarctic areas (Sturm et al., 2001; Kullman, 2002; Dullinger et al., 2004; Tape et al., 2006) and that they are predicted to perform better in terms of productivity in a warmer climate (Parsons et al., 1995; Arft et al., 1999; van Wijk et al., 2004; Walker et al., 2006). Thus, the functional types that are expected to benefit most from warming in the adult stage may also benefit most in the recruitment phase, possibly resulting in a positive feedback.
The species in which germination was most negatively affected by a warmer climate were three forbs that currently occur in subarctic birch forest (E. angustifolium, S. dioica and V. cracca) and the dwarf shrub V. uliginosum. Seeds of these species apparently suffer greatly from below zero temperatures during winter, resulting in less and slower germination. However, they performed well after stratification under mild winter temperatures, so, similarly to V. myrtillus, their expansion is expected to be restricted to snow-rich places, for instance close to shrubs which trap and hold snow during winter (Sturm et al., 2005). We can, however, not explain why the combination of ‘thin snow cover’ stratification and ‘optimum’ incubation was extremely unfavourable in V. uliginosum.
The observed species-specific responses may have consequences for the patterns of species migration due to warming. In some species, seedling recruitment will improve more than in others in a warmer climate; therefore not all species will be able to shift northward or upslope at the same pace. This may result in plant communities without previous analogues (Davis, 1989; Kullman, 2002) or in shifts in dominance within existing communities (Kelly and Goulden, 2008). In any case, the importance of species-specific reactions, also during the recruitment stage, should be taken into account in models that forecast population dynamics and future species migrations.
Seed germination does not appear to limit northward range expansion in the three boreal tree species of the present experiment. Alnus incana, B. pubescens and P. sylvestris all germinated equally well (similar germination percentages) in the colder stratification and incubation treatments as in the warmer ones. Moreover, they did not germinate significantly slower under current subarctic conditions than the dominant tree species in the Abisko region, B. pubescens ssp. czerapanovii. The current distribution of these species is thus most likely limited by other processes than seedling recruitment.
To conclude, it was found that colder winter soil temperatures are likely to result in lower germination percentages and slower germination in about half of the subarctic species. Warmer summers, on the other hand, reduce the germination time and may sometimes increase the number of germinating seeds. The combined effect of both is beneficial in most species and therefore seedling establishment was expected to improve in a future warmer subarctic climate.
We thank all people who helped us with collecting and counting seeds and also the staff at the Abisko Scientific Research Station for their support. This work was supported by the Research Foundation – Flanders (postdoctoral fellowship to A.M., visiting postdoctoral fellowship to A.S.); and The Kempe Foundation (B.J.G.).