Contrasting sensitivity to extreme winter warming events of dominant sub-Arctic heathland bryophyte and lichen species


  • Jarle W. Bjerke,

    Corresponding author
    1. Norwegian Institute for Nature Research (NINA), FRAM, High North Research Centre on Climate and the Environment, NO-9296 Tromsø, Norway
    2. Tromsø University Museum, University of Tromsø, NO-9037 Tromsø, Norway
    Search for more papers by this author
  • Stef Bokhorst,

    1. Department of Animal and Plant Sciences, University of Sheffield, Western Bank, Sheffield S10 2TN, UK
    Search for more papers by this author
  • Matthias Zielke,

    1. Division for Arctic Agriculture and Land Use, Norwegian Institute for Agricultural and Environmental Research, NO-9269 Tromsø, Norway
    Search for more papers by this author
  • Terry V. Callaghan,

    1. Department of Animal and Plant Sciences, University of Sheffield, Western Bank, Sheffield S10 2TN, UK
    2. Royal Swedish Academy of Sciences, Lilla Frescativägen 4A, SE-114 18 Stockholm, Sweden
    Search for more papers by this author
  • Francis W. Bowles,

    1. The Ecosystems Center, Marine Biological Laboratory, 7 MBL Street, Woods Hole, MA 02543, USA
    Search for more papers by this author
  • Gareth K. Phoenix

    1. Department of Animal and Plant Sciences, University of Sheffield, Western Bank, Sheffield S10 2TN, UK
    Search for more papers by this author

Correspondence author. E-mail:


1.  Climate change in northern high latitudes is predicted to be greater in winter rather than summer, yet little is known about the effects of winter climate change on northern ecosystems. Among the unknowns are the effects of an increasing frequency of acute, short-lasting winter warming events. Such events can damage higher plants exposed to warm, then returning cold, temperatures after snow melt, and it is not known how bryophytes and lichens, which are of considerable ecological importance in high-latitude ecosystems, are affected by such warming events. However, even physiological adaptations of these cryptogams to winter environments in general are poorly understood.

2.  Here we describe findings from a novel field experiment that uses heating from infrared lamps and soil warming cables to simulate acute mid-winter warming events in a sub-Arctic heath. In particular, we report the growing season responses of the dominant lichen, Peltigera aphthosa, and bryophyte, Hylocomium splendens, to warming events in three consecutive winters.

3.  While summertime photosynthetic performance of P. aphthosa was unaffected by the winter warming treatments, H. splendens showed significant reductions in net photosynthetic rates and growth rates (of up to 48% and 52%, respectively). Negative effects were evident already during the summer following the first winter warming event.

4.  While the lichen develops without going through critical phenological stages during which vulnerable organs are produced, the moss has a seasonal rhythm, which includes initiation of growth of young, freeze-susceptible shoot apices in the early growing season; these might be damaged by breaking of dormancy during warm winter events.

5. Synthesis. Different sensitivities of the bryophyte and lichen species were unexpected, and illustrate that very little is known about the winter ecology of bryophytes and lichens from cold biomes in general. In sharp contrast to summer warming experiments that show increased vascular plant biomass and reduced lichen biomass, these results demonstrate that acute climate events in mid-winter may be readily tolerated by lichens, in contrast to previously observed sensitivity of co-occurring dwarf shrubs, suggesting winter climate change may compensate for (or even reverse) predicted lichen declines resulting from summer warming.


The circumpolar region is characterized by vegetation rich in lichens and bryophytes (Callaghan et al. 2005). In addition to their significant contribution to biodiversity in high-latitude ecosystems (e.g. Elvebakk & Bjerke 2006), bryophytes, lichens and their associated cyanobacteria play major functional roles by contributing considerably to ecosystem carbon and nitrogen sequestration, soil insulation, soil stability and preservation of permafrost (Oechel & van Cleve 1986; Longton 1988; Chapin & Bledsoe 1992; Heijmans, Arp & Chapin 2004). However, bryophyte- and lichen-rich ecosystems in the circumpolar region are under threat. Large areas of heath vegetation rich in bryophytes and lichens have deteriorated because of overgrazing by reindeer and caribou, increasing land use and air pollution (Tømmervik et al. 2004; Rees et al. 2008; Myking et al. 2009). In addition, enhanced plant growth as a result of increasing growing season temperatures and precipitation rates (Cornelissen et al. 2001; Tømmervik et al. 2004; van Wijk et al. 2004; Walker et al. 2006) and increasing deposition rates of reactive nitrogen (van Wijk et al. 2004; Fremstad, Paal & Möls 2005) induce shifts in higher plant–cryptogam interactions, leading to reduced lichen and bryophyte cover. These factors are predicted to increase in impact in the future (Callaghan et al.2005), thus causing declines in cryptogam biomass and biodiversity in circumpolar regions, with potentially important ecological and socioeconomic impacts. Effects on reindeer husbandry, for example, could be considerable, given that lichens constitute a major component of the winter diet of reindeer and caribou. Incorporating the loss of lichen biomass in their modelling, Rees et al. (2008) estimate that reindeer numbers in Norway and Sweden must be reduced by 50–60% by 2080 to maintain self-sustainable populations.

Most currently used climate-related models for predicting future vegetation changes (e.g. Kaplan & New 2006; Sitch et al. 2007; Wolf, Larsson & Callaghan 2008) are based on directional climate trends, e.g. of regional warming and humidification. However, there is growing evidence that discrete or extreme climate events play an important role in shaping and changing ecosystems world-wide (e.g. Jentsch, Kreyling & Beierkuhnlein 2007; Barrett et al. 2008; Bokhorst et al. 2008, 2009; Post et al. 2009). For instance, acute climate events, e.g. a week-long period with above-zero temperatures during mid-winter or summer ‘heat waves’ in the Arctic, can have disproportional influence over ecosystems relative to the short temporal scale over which they occur (Barrett et al. 2008).

Detection of acute climate events in climate records and prediction of the future frequency of such events are challenging, particularly because many regions do not have satisfactory monitoring systems for carrying out extreme-value climate statistics, and because standard outlier eliminations tend to hide rare events from climate time series (Jentsch, Kreyling & Beierkuhnlein 2007). Despite these challenges, several studies suggest that discrete wintertime warming events have become and will become even more frequent in circumpolar regions (Visbeck et al. 2001; Callaghan et al. 2010). In these events, warm spells can cause partial or complete snow melt over large areas (Bonsal et al. 2001; Putkonen & Roe 2003). The loss of the insulating snow cover combined with the return of subzero temperatures after such warming events may lead to extensive plant damage. Simulations of 1-week-long winter warming events in the sub-Arctic, using infrared heating lamps suspended above the vegetation run with and without soil warming cables (to further investigate the impacts of soil thaw), had major negative impacts on survival, phenology and reproduction of ericoid dwarf shrubs, suggesting that such events may potentially severely impact the biodiversity and productivity of these systems (Bokhorst et al. 2008, 2009, 2010). Also, a natural event in December 2007 in the sub-Arctic confirmed these damaging impacts in some higher plants (Bokhorst et al. 2009). Currently, though, the impacts on cryptogams remain unknown.

Given the importance of lichens, bryophytes and their associated cyanobacteria for sub-Arctic ecosystems, there is a clear need to determine how these cryptogams are affected by such winter warming events. Using the same field experimental facility as described by Bokhorst et al. (2008, 2009), we aimed to quantify the effects of three discrete warming events (winters of 2007, 2008 and 2009) on the most abundant lichen, Peltigera aphthosa, and bryophyte, Hylocomium splendens, in this area. While long periods of mild temperatures are required to break the winter dormancy of many sub-Arctic vascular plants, past work suggests that cold-adapted lichens and bryophytes are reactivated relatively rapidly when temperature, light and humidity conditions rise above threshold limits (Kallio & Saarnio 1986; Lange 2003; Glime 2007; Kappen & Valladares 2007; Bjerke 2010).

Given the ability of bryophytes and lichens to switch rapidly between states of metabolic rest and activity as dictated by fluctuations in the external environment, we hypothesize that photosynthetic performance and growth rates during growing seasons following winter warming events would not be negatively affected, and that these cryptogams can therefore tolerate winter warming events better than the dwarf shrubs with which they co-occur (cf. Bokhorst et al. 2008, 2009). We also hypothesize that the impacts of canopy only warming and canopy and soil warming will be the same, as the cryptogams are not in direct contact with the soil.

Materials and methods

Site description and warming treatment

The study was performed in sub-Arctic heathland vegetation close to the Abisko Scientific Research Station in northern Sweden (68°21′ N, 18°49′ E). The field layer is dominated by dwarf shrubs, in particular the deciduous Vaccinium myrtillus L. and the evergreen V. vitis-idaea L. and Empetrum nigrum L. s.l. The ground layer is dominated by feather mosses, especially Hylocomium splendens (Hedw.) Schimp., and by foliose cyanolichens, especially Peltigera aphthosa (L.) Willd. Other bryophytes and lichens are more scattered.

The experiment – which simulates warming based on real events occurring previously in the Abisko region – is described in the study by Bokhorst et al. (2008, 2009). In brief, it consists of 18 plots (2.1 × 1.0 m) representing six control plots and six each of two warming treatments, canopy warming and canopy and soil warming. For canopy warming plots, four infrared heating lamps were suspended (70 cm apart) in parallel from wooden frames. To further investigate the effects of increased soil thaw, the same set-up was used in six additional plots, but these were further warmed by soil heating cables producing 120 W m−2 at 5 cm soil depth and running parallel at 20 cm distance from each other (canopy and soil warming). Soil warming cables were switched on 2 days after the lamps to simulate the delay in soil thaw during a real event. Control plots received no warming treatment and remained insulated under the natural winter snow cover. To monitor temperature, thermistors were placed in each plot at canopy height and at the soil surface and 5 cm depth. Temperatures were logged at 6-h intervals and recorded on a data logger.

Simulation of discrete winter warming events started at the beginning of March in 2007, 2008 and 2009. Each event lasted 7 days during which the lamps were kept at 50 cm distance from the snow surface and lowered accordingly as the snow depth decreased. This approach ensured gradual snow thaw, taking 2–3 days to thaw the full depth of snow in each plot. As vegetation became exposed, lamps were kept at 50–70 cm above the soil surface to maintain canopy warming (the lower lamp heights were needed during higher wind speeds and lower ambient temperatures). Temperatures from the thermistors were monitored to ensure warming was realistic and within the bounds of temperatures recorded for real events. Thermocouple measurements of vegetation surface temperatures were also made. At the end of the warming treatment, heating lamps were removed from the frames to avoid shading effects of the lamps during the following growing season. Plots were then left untouched for the remaining period of winter before data collection in spring and summer (late May–August).


Hylocomium splendens and Peltigera aphthosa are the dominant cryptogams in the ground flora. These two species are most abundant in mesic heath vegetation, which, under normal winter conditions, is covered by snow for 5–7 months. Hylocomium splendens occurred in all 18 plots, but in some plots with very low abundance, while P. aphthosa was absent from three plots, reducing the number of replicates to 4 and 5 for canopy and canopy-and-soil warming, respectively. Cyanobacteria grow as epiphytes on the stems of H. splendens and in well-defined colonies, called cephalodia, in the thallus of P. aphthosa, which is tripartite, i.e. it consists of three major bionts; the fungus, the green alga and the cyanobacterium. Measurements of the cyanobacteria’s nitrogen fixation rates are presented in Appendix S1 in Supporting Information.

Photosynthesis and chlorophyll fluorescence

Gas exchange measurements were taken in June or July during the growing seasons after winter warming events. Measurements were made with the portable gas exchange fluorescence system GFS-3000 (Heinz Walz GmbH, Effeltrich, Germany). As photosynthesis in poikilohydric organisms is strongly affected by water status, repeated measurements were made on each sample, starting with very humid samples that experienced photosynthetic depression. Samples for gas exchange were randomly selected and carefully removed from the plots, placed on paper towels in small containers and sprayed with water to achieve full hydration. The samples were left overnight to fully recover from desiccation. Samples were re-hydrated in the morning before initiation of measurements. Only first-year and second-year segments of H. splendens were used, as older segments have reduced photosynthetic rates (Oechel & van Cleve 1986). Each moss sample consisted of c. 5 cut shoots. Samples of P. aphthosa consisted of one ellipsoid lobe without apothecia, c. 2.5 cm wide and 4 cm long. Samples were allowed to slowly air dry between measurements, which were made until assimilation rates showed that optimal water content was passed. Immediately after each measurement, samples were dried completely and weighed. Assimilation rates were then calculated on a per-dry-weight basis. One sample each of P. aphthosa and H. splendens was analysed from each plot at each occasion.

A saturating photosynthetic photon flux density (PPFD) of 800 μmol m−2 s−1 was used during measurements of net photosynthesis. CO2 concentration was set to 380 p.p.m. and cuvette humidity to 9000 p.p.m. H2O. Cuvette temperatures were adjusted to ambient noon temperatures at the time of measurement. Samples were stored at the same temperatures as the one at which they were dried between the measurements. They were equilibrated in the cuvette for 20 min each time, and exchange rates were measured repeatedly during the last 10 min of this period.

Using the PAM-fluorometer of the GFS-3000, maximal quantum efficiency of PS II, i.e. Fv/Fm (Maxwell & Johnson 2000), was measured on dark-adapted, naturally moist or wetted samples during the growing seasons after the winter warming events. Measurements on H. splendens were taken on second-year plus first-year developing segments (Callaghan, Collins & Callaghan 1978). Measurements on P. aphthosa were taken on young lobe parts without any, or with very few, visible cephalodia.

Segment growth

Shoots of H. splendens collected at one point in time during the growing seasons following the winter warming events of 2008 and 2009, and also during the first recovery growing season (first season following a winter without an event simulation – 2010), were dried, and lengths of first-year and second-year segments were measured with digital callipers.

Data analyses

Differences between treatments at any one measurement event were tested with a fixed one-way anova design. Time differences and overall treatment differences, and their interactions, were tested with repeated-measures anova. The Tukey HSD test was used for multiple comparisons of treatment effects. F- and P-values presented in the text are from the repeated-measures anova unless otherwise stated.

Data sets containing more than one randomly chosen data point per replicate (plot) were first surveyed using a nested anova design to check whether the variation within the replicates was larger than between replicates. As the variation within plots in all cases was very low as compared to the variation between plots, the mean value per replicate was used in further anova analyses. Data sets were tested for normality and heterogeneity. No transformation of data was necessary. No significant or near-significant (< 0.10) interactions between treatment and time were detected for any of the analysed parameters. Therefore, time–treatment interaction terms are not referred to in the Results. All tests were performed using the spss 16 and 17 statistical packages for Windows (SPSS Inc., Chicago, IL, USA).


Treatment effects on canopy air and cryptogam surface temperatures

The initiation of the canopy warming and canopy and soil warming treatments caused an immediate rise in soil and canopy air temperatures (see Bokhorst et al. 2008, 2009, 2010 for detailed figures). In all years, both types of warming treatment resulted in complete snow melt and full exposure of the ground vegetation after 3 days, as typically observed during a natural warming event (Bokhorst et al. 2009). Leaf canopy temperatures were on average between 5 and 10 °C during the warming simulations. Canopy and soil surface temperatures did not differ between the canopy-only and the canopy-with-soil warming treatments, whereas the soil warming in the canopy and soil warming treatment increased soil temperatures by 4 °C at 5 cm depth compared with control plots (see fig 2b in Bokhorst et al. 2008). Surface temperatures of H. splendens and P. aphthosa, measured by leaf thermocouples, were on average 8.6 for both species ±0.6 and 0.7 °C standard error, respectively.

After the warming periods (following switching off of the lamps and cables), temperatures in the warmed plots immediately declined and returned to ambient. However, the small amount of snow that accumulated in the warmed plots following the warming periods was not sufficient to re-insulate the ground cover, resulting in much colder canopy temperatures (with a minimum at −18 °C) than in plots covered by snow (controls, minimum temperature of −8 °C) for the remainder of winter until spring.

Peltigera aphthosa

The warming treatments of 2007, 2008 and 2009 did not affect net photosynthetic rates and PS II efficiency in the following growing seasons (Fig. 1). There was some variation in net photosynthetic rates between growing seasons, but this variation was also seen in the controls (Fig. 1a). There was no time effect on PS II efficiency. The warming treatments did not affect the nitrogen fixation rates of P. aphthosa and the associated lichen Nephroma arcticum (Fig. S1).

Figure 1.

 Ecophysiological performance of Peltigera aphthosa in the growing seasons following winter warming events. (a) Net photosynthetic rates at optimal water content (treatment: F2, 9 = 2.9, = 0.11, time: F2, 18 = 31.8, = 0.000); (b) photosystem II efficiency of dark-adapted samples (treatment: F2, 10 = 0.2, = 0.85, time: F2, 20 = 4.3, = 0.028). = 4–5 per treatment. Error bars are ±1 SE.

Hylocomium splendens

Growing season net photosynthetic rates in H. splendens were reduced by an average of 44% over the 3 years in the canopy warming treatment and 35% in the canopy and soil warming treatment (Fig. 2a). In 2007, net photosynthetic rates were significantly lower in the canopy warming treatment than in the control (= 0.03). While the canopy and soil warming treatment was not statistically different from the canopy warming treatment or the control, it none-the-less had a mean value much closer to that of the canopy warming treatment. In 2008 and 2009, both warming treatments had significantly lower rates than the controls, while there were no significant differences between the two treatments (Fig. 2a). The relative reductions and corresponding one-way P-values for 2007, 2008 and 2009 were 38% (0.03), 48% (0.01) and 47% (0.008) in the canopy warming treatment, and 32% (0.15), 42% (0.04) and 44% (0.02) in the canopy and soil warming treatment. There were no treatment or time effects on PS II efficiency (Fig. 2b).

Figure 2.

 Ecophysiological performance of Hylocomium splendens in the growing seasons following winter warming events. (a) Net photosynthetic rates at optimal water content (treatment: F2, 15 = 15.5, = 0.000, time: F2, 30 = 7.8, = 0.002); (b) photosystem II efficiency of dark-adapted samples (treatment: F2, 13 = 1.2, P = 0.34, time: F2, 26 = 4.1, P = 0.03) n = 4–6 per treatment and time combination (some plots with low abundance and hence not sampled in all occasions). Error bars are ±1 SE. Different letters indicate significant differences (< 0.05) between treatments.

Growth of shoot segments in the warmed plots was severely reduced following the second and third winter warming events, while there were no significant differences between the two warming treatments (Fig. 3). Following the second event (2008), second-year segment growth was reduced by 33–35% in the treatment plots (Fig. 3b, left panel). However, first-year segment growth in 2008 was more variable and only marginally affected (Fig. 3a, left panel; warming treatments analysed separately: F2, 15 = 2.68, = 0.101; warming treatments pooled: F1, 16 = 4.54, = 0.049). In 2009, both first-year and second-year segments suffered from a significant, serious growth reduction in winter-warmed plots (50–52% and 40–45% reduction, respectively; Fig. 3, middle panels). These effects persisted during the first recovery growing season, i.e. the growing season of 2010, with a 34–38% reduction in first-year segments, and a 41–42% reduction in second-year segments in treated plots compared with controls (Fig. 3, right panels).

Figure 3.

 Lengths of shoot segments of Hylocomium splendens measured in the growing seasons following the winter warming events of 2008 and 2009, and during the first recovery season of 2010. (a) First-year segments, which were under development. Left scale is for 2008 and 2009, right scale is for 2010 (samples were collected later in the growing season in 2010, hence the different scales). (b) Second-year segments. First-year segments: treatment: F2, 15 = 11.8, P = 0.001, time: F2, 30 = 75.5, P = 0.000. Second-year segments: treatment: F2, 15 = 41.0, P = 0.000, time: F2, 30 = 7.71, P = 0.002. = 6 for each treatment. Error bars are ±1 SE. Different letters indicate significant differences (< 0.05) between treatments.

The nitrogen fixation rate of epiphytic cyanobacteria growing on H. splendens was not affected by the warming treatments (Fig. S1).


Very little is known about the winter ecology of bryophytes and lichens in general, and their responses to winter climate change in particular are more or less unknown (Glime 2007; Bjerke 2009, 2010) with most cold- and frost-related research on bryophytes and lichens being performed either in the laboratory or during the snow-free season in cold biomes. Whereas some winter climate change studies have tested the effects of altered snow conditions on bryophyte- and lichen-containing vegetation (Dorrepaal et al. 2003; Wahren, Walker & Bret-Harte 2005; Scott et al. 2007), this study is the first to test the effects of winter warming events on these important functional groups.

Our findings are consistent with the hypothesis that lichens and their associated cyanobacteria are tolerant of short-lived winter warming events as seen from the general absence of winter warming impacts on chlorophyll fluorescence, net photosynthetic rates and nitrogen fixation rates in the dominant P. aphthosa. In contrast, the dominant bryophyte H. splendens was clearly not tolerant of winter warming events, showing significant reductions in both net photosynthetic rates and growth rates. The reasons for the contrasting sensitivity are not fully understood, although such differences may be driven by contrasting growth form and seasonal rhythm between lichens and bryophytes from mesic habitats.

While radial growth in lichens is undertaken by an almost indefinite number of hyphal cells (e.g. Armstrong 2003), growth in mosses is restricted to a few regions of meristematic cells (e.g. La Farge-England 1996). The huge difference in number of growing points suggests that the buffering capacity of lichens is much higher than that of mosses against damage from external stress. This may explain why only the moss was affected by the winter warming treatments.

Many bryophytes, H. splendens included, go through phenological stages with initiation of growth of new segments early in the growing season, completion of growth during the growing season and instigation of frost resistance mechanisms in autumn (Callaghan, Collins & Callaghan 1978; Longton 1988; Rütten & Santarius 1992). There is, to our knowledge, no evidence of active growth of sub-Arctic or alpine heath mosses during the cold season, despite the potential for slightly positive net photosynthetic rates under rather low light intensities in autumn and after snow melt in spring (Kallio & Saarnio 1986; Larsen et al. 2007). Breaking of dormancy during the winter warming events is the likely driver for the observed mortality of the vascular plants (Bokhorst et al. 2010). Mosses probably also experience breaking of dormancy, and reactivation in mid-winter may disturb the seasonal timing in H. splendens, initiating development of new segments. Young shoot apices appear to be especially susceptible to freezing damage (Clausen 1964; Hudson & Brustkern 1965), and indeed, this was also observed in bryophytes living in cold Antarctic environments (Longton & Holdgate 1967; Collins 1976; Kennedy 1993). Temperatures at −7.6 °C within 24 h after warming was switched off in our study, followed by temperatures as low as −18 °C 2 weeks later (Bokhorst et al. 2008, 2009, 2010) probably caused freezing damage to the moss shoot apices, which was manifested as reductions in growing season photosynthesis and growth. The recorded reductions in H. splendens photosynthesis and growth seem to be far beyond the year-to-year variation found in populations with stable winter climates (cf. Callaghan, Collins & Callaghan 1978; Callaghan et al. 1997; Økland 1997).

Lichens, on the other hand, do not have any clear phenological stages that need to be completed before the onset of winter (Benedict 1990; Hahn et al. 1993; Lange 2003; Kappen & Valladares 2007). When reactivated, lichens apparently continue growth processes at the point they were stopped before anabiosis, and there are no vulnerable organs in lichens which are susceptible to freezing damage, like the young, frost-sensitive moss shoot apices. Hence, in contrast to H. splendens, the lack of clear phenological stages contributes to explaining why P. aphthosa was unaffected by the mid-winter warming.

We have interpreted the reduced vitality of H. splendens as being a result of the freezing stress that they were exposed to after the warming events. However, as winter-adapted cryptogams have reduced heat tolerance (Tegler & Kershaw 1981; Longton 1988), one may consider that the warming per se also had negative effects. There are, however, to our knowledge, no reports that might suggest that H. splendens would be less tolerant than P. aphthosa to winter heat. Moreover, the warming temperatures they were exposed to during the events are similar to the temperature regimes immediately after snow melt in spring (and not followed by severe freezing), and these natural conditions do not seem to harm the cryptogams in any way, suggesting that the damage to H. splendens was largely caused by the severe freezing temperatures following the warming events. The responses seen here for H. splendens may be restricted to continental regions in contrast to more oceanic regions where extreme winter warming events are unlikely to be followed by very low temperatures (<−15 °C). It is currently unclear whether the bryophyte response seen here is driven by the disruption of the developmental program or simply by deep-freezing because of the absence of snow cover. Further studies incorporating different regimes of mid-winter snow melting and re-covering with snow across oceanic to continental gradients should elucidate the mechanisms behind the susceptibility of H. splendens to these winter warming events.

There were no significant treatment effects on the PS II efficiency of any of the tested species. It is interesting that PS II efficiency in H. splendens was not affected by the warming treatments, while net photosynthesis was. Damage to PS II is often the first manifestation of stress (Maxwell & Johnson 2000). However, PS II efficiency is not always a true indicator of plant stress. Several stress studies, both on vascular plants and on bryophytes, have found strong effects on growth and photosynthetic rates, but no or only small effects on PS II efficiency (e.g. Taulavuori et al. 2000; Nabe et al. 2007; Granath, Wiedermann & Strengbom 2009). Although the mechanisms are not fully understood, these reports and our results suggest that gas exchange and growth rates are more precise indicators of plant performance under various types of stress conditions.

Activity of organisms underneath winter snow cover is probably less dormant than generally assumed. Indeed, the microbial community composition is known to change during winter (Schadt et al. 2003), considerable ecosystem carbon fluxes occur (Larsen et al. 2007; Liptzin et al. 2009) and some plants actively grow new roots underneath the snow layer to access nutrients (Onipchenko et al. 2009). Bryophytes and lichens are also among the organisms which can be active during favourable winter conditions. However, the contrasting sensitivities of the dominant bryophyte and lichen in our study suggest that this strategy may not be successful under all winter snow conditions. The contrasting sensitivities further suggest that many more species-specific studies on cryptogam winter activity are required if we are to understand their roles in winter ecosystem processes, which also affect growing season processes (Cornelissen et al. 2007).

Concluding remarks

We have shown here that acute, extreme warming events in the sub-Arctic can strongly affect the ecophysiology and growth of the dominant bryophyte H. splendens. This bryophyte therefore appears sensitive to winter warming, as are the dwarf shrubs in this widely distributed Arctic/sub-Arctic ecosystem (Bokhorst et al. 2008, 2009, 2010). If such events become more frequent, as a result of climate change, this may have large consequences for productivity of the affected species and may ultimately induce community shifts (Hobbs, Yates & Mooney 2007; Bokhorst et al. 2008). Only one lichen species, Peltigera aphthosa, has so far been investigated in detail. The contrasting sensitivities of the dominant bryophyte and lichen species were unexpected, and highlight that the general understanding of cryptogams’ adaptive mechanisms to winter climate and snow cover is still poor. If most lichen species are tolerant to winter warming events, then lichens may be favoured by such events at the expense of competing dwarf shrubs and bryophytes. If so, some of the trends for declines in lichen-dominated ecosystems seen in recent decades (Cornelissen et al. 2001; Tømmervik et al. 2004; van Wijk et al. 2004; Walker et al. 2006; Rees et al. 2008; Myking et al. 2009) may be compensated for, which is in sharp contrast to the current view that global change will enhance vascular plant biomass at the expense of lichens (Cornelissen et al. 2001; van Wijk et al. 2004). Furthermore, the consequences for reindeer husbandry – where lichens are a major food source – may be much less negative than hitherto predicted (Rees et al. 2008). Whereas climate-change-driven extension of the growing season can affect lichens negatively, climate-change-driven discrete winter warming events may have positive effects on lichens. The former of these two climate change elements have, to date, received far more attention (Jentsch, Kreyling & Beierkuhnlein 2007; Post et al. 2009). Our current study shows that acute climate events in mid-winter may be equally important as summer warming for regulation of high-latitude ecosystem functioning and community composition. Overall, the changes place a considerable challenge to predicting vegetation change in a future Arctic where winter will warm more than summer


This research was supported by a grant from the Research Council of Norway (project no. 171542/V10) awarded to J.W.B., by a Leverhulme Trust (UK) grant to G.K.P. and T.V.C. (grant F/00 118/AV), by ATANS grants (EU Transnational Access Program, FP6 Contract no. 506004) to J.W.B., S.B., M.Z. and G.K.P., and by the Norwegian Institute for Nature Research. J.W.B.’s position at the Tromsø University Museum was financed by the Norwegian-Swedish Research School in Biosystematics, which received funding from the Research Council of Norway and the Norwegian Biodiversity Information Centre. We would like to thank staff of the Royal Swedish Academy of Sciences’ Abisko Scientific Research Station for their assistance. Infrastructure and equipment support were supplied by the Royal Swedish Academy of Sciences and Jerry Melillo from the Ecosystems Center at the Marine Biological Laboratory, USA. We also thank the University Centre in Svalbard for access to gas chromatography facilities.