• birch (Betula pubescens);
  • dehardening;
  • ecotype;
  • frost hardiness;
  • global warming;
  • lipid peroxidation


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • • 
    The aim was to elucidate the effects of elevated winter temperatures on the dehardening process of mountain birch (Betula pubescens ssp. czerepanovii) ecotypes and to evaluate their susceptibility to frost damage under warming climate conditions.
  • • 
    Ecotypes from 60 to 71° N latitudes and 20–750 m altitudes were grown in northern Norway (70° N) and subjected to simulation of the photoperiod in southern Norway (60° N) by artificial illumination from September onwards. In November, the seedlings were transported to the south (60° N) to overwinter at ambient or 4°C above ambient temperatures. Frost hardiness and lipid peroxidation were determined during January–April.
  • • 
    The higher winter temperature accelerated dehardening, and there were significant differences between the ecotypes. Among tree individuals of southern origin, the alpine ecotype exhibited the most rapid rate of dehardening, whereas the oceanic type showed the slowest rate. Lipid peroxidation supported the above findings.
  • • 
    Since temperature elevation was unequal for the ecotypes with respect to climatic change, the frost hardiness results were normalized to obtain an equal +4°C temperature rise. The risk of frost injury seemed to be lowest in the northernmost ecotypes under a temperature elevation of +4°C, obviously due to their adaptation to a wider temperature range.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The current atmospheric CO2 concentration of c. 370 ppm is expected to double by the end of 2100. Consequently, the globally averaged surface temperature is projected to increase by 1.4–5.8°C. The most pronounced warming is predicted to occur at high latitudes, especially in winters, with an anticipated increase of c. 4–7°C for the boreal zone (Bach, 1988; Maxwell, 1992; Murray, 1995; Houghton et al., 2001). In perennial plant species, frost hardiness is strongly controlled and determined by the prevailing ambient temperatures, in that warm winter temperatures reduce frost hardiness (e.g. Sakai & Larcher, 1987). The increasing effect of climatic warming on the incidence of the risk of frost damage has been debated during the past two decades. Theoretical approaches have emphasized the possible risk, especially in late spring (e.g. Cannell & Smith, 1986; Hänninen, 1991), and this finding has been supported by some experimental and field studies (Repo et al., 1996; Taulavuori et al., 1997a; Gansert et al., 1999). However, the risk that the climate change will cause premature dehardening has been thought to be small for native species in Scandinavia (Ögren, 2001).

Mountain birch (Betula pubescens ssp. czerepanovii) is the most important tree-line species in Scandinavian forests (Skre, 1993). The mountain birch forest zone constitutes the northernmost boundary of boreal forest in the Scandinavian mountain chain, and the forest is thus exposed to severe climatic restraints (Tenow, 1996). There was a horizontal zone at 410–430 m a.s.l. in Abisko, northern Sweden (68° N), where the mountain birch did not leaf in the summer of 1991. It was hypothesized that the damage was caused by a cold backlash in mid-April after a very warm first half of the month. Similar hazards have also been recorded in 1962, 1963 and 1985 (Tenow, 1996). Therefore, in addition to theoretical and experimental documentation, there is also some natural evidence to support the assumption that springtime frost damage evolves after unseasonably warm spells.

Adaptation to locally variable climatic and meteorological conditions has resulted in the evolution of genetic ecotypes. Ecotypes in cold regions evolve to derive maximum benefit from the short period that enables photosynthesis, growth and reproduction. On the other hand, correct timing of growth cessation and the length of the dormant state in the overwintering period are necessary for the survival of plants at high northern latitudes. The dormant state is broken by two phases: accumulation of air temperatures near +5°C in winter (i.e. the chilling requirement for completion of rest) and the consequent long-term exposure to temperatures above a given threshold required for growth (i.e. the high-temperature requirement for onset of growth) (Hänninen, 1995). The latter requirement for growth resumption is probably more important than the chilling requirement for dormancy release (Häkkinen et al., 1998; Pop et al., 2000). The lower threshold air temperature for zero growth in birch is 7–7.8°C (Kullman, 1993).

The chilling requirement consists of two components, the temperature and the duration of chilling, and the latter requirement for northern deciduous trees varies considerably both between and within species (Heide, 1993; Myking & Heide, 1995). The chilling requirement of birch species decreases significantly along with increasing latitude of origin (Myking & Heide, 1995). In addition, high altitude (e.g. Murray et al., 1989) and continental climate (Leinonen, 1996b) are factors that minimize the chilling requirement of origin, although the relative importance of these aspects is unknown. Thus, at northern high latitudes, the milder the climate, the higher is the chilling requirement, with a longer winter-to-spring time period of low risk of frost damage under warming climate.

The present study consisted of experimental warming of mountain birch ecotypes (Table 1). The aim was to elucidate the effects of elevated winter temperatures on the dehardening process of these ecotypes. Dormancy and frost hardiness are separate adaptive traits, and frost hardiness does not directly reflect the state of dormancy at any time of the year. However, the rate of dehardening or the capacity to maintain frost hardiness depends indirectly on the state of dormancy, since frost hardiness depends on the ontogenetic cycle of a plant (e.g. Fuchigami et al., 1982) in addition to environmental fluctuations (Repo et al., 1990). There is evidence that dehardening of northern woody species starts gradually much before bud break, although frost hardiness is eventually lost in parallel with the resumption of growth (see Repo, 1992; Taulavuori et al., 1997a,b). The buds maintain a high degree of frost hardiness until opening, while the cambial activity of stem tissue arises due to warm spells in the spring, making the plant vulnerable to frost damage (e.g. Zalasky, 1976). In the present study the frost hardiness in stem tissue (e.g. cambium) was determined and followed in late winter and spring. Because of their shared dependency on ontogenetic development, we decided to test if the rate of dehardening is linked to their chilling requirement, and which of the following hypotheses best fits the observed pattern of the dehardening rate.

Table 1.  Seed origins representing the mountain birch (Betula pubescens ssp. czerepanovii) ecotypes investigated in the experiment
Seed originLatitudeAltitudeDist. to oceanTmean*Tminabs**
  • *

    Mean winter (Dec–Mar) temperature based on 30-year data (1931–1960).

  • **

    Absolute minimum temperature in Mar based on 30-year data (1931–1960).

NB (Blefjell, N)60° N750 m100 km −4.3°C−12.6°C
IC (Hafnaskogur, IC)63° N 50 m < 5 km ±0.0°C −9.3°C
NMe (Melbu, N)68° N 20 m < 5 km +0.6°C −9.7°C
FJ (Kevo/FIN)69° N250 m200 km−11.9°C−33.3°C
NHa (Hammerfest/N)71° N 50 m < 5 km −4.2°C−17.2°C

Hypothesis 1

Dehardening of the mountain birch ecotypes follows the pattern derived from their chilling requirement. This assumes that the chilling requirement of ecotypes decreases towards: (i) northern high latitudes; (ii) high altitudes; and (iii) continental sites (Fig. 1).


Figure 1. Three-dimensional plot to illustrate the basis for Hypothesis 1. E denotes the hypothetical ecotype with the highest chilling requirement and a low risk of spring frost damage. E′ stands for the opposite ecotype with the lowest chilling requirement and a consequent high risk of spring frost damage. This is based on the following assumptions: chilling requirement of ecotypes decreases towards (1) northern high latitudes (2) high altitudes and (3) continental sites. The NB, FJ, NHa, NMe and IC are the investigated ecotypes (see Table 1 for their spatial parameters).

Download figure to PowerPoint

According to Hypothesis 1, birch ecotypes of different climatic origin are to be arranged from low to high chilling requirement (and low to high capacity to maintain frost hardiness under warming winter climate, respectively) given as:

  • Order 1 (Hypothesis 1): NB; FJ < NHa; Nme < IC

Hypothesis 2

The assumptions given in the context of Hypothesis 1 may be generalized as: The chilling requirement of ecotypes decreases towards harsh winter climate. Thus, if ‘harshness’ of winter is a reflection of local temperature conditions, the order derived from Hypothesis 1 can be modified as:

Order 2 (Hypothesis 2): FJ < NHa; NB < Nme; IC

According to this hypothesis, the ecotypes are arranged from low to high temperature climate (see Table 1), which reflects the capacity to maintain frost hardiness on a similar basis as in Hypothesis 1.

In addition, there is evidence that light climate plays a role in the start of ontogenetic development from dormant to active growth phase (Linkosalo et al., 2000). Therefore, we performed the present experiment in a common garden-like system: the seedlings experienced the same photoperiod over frost hardening to dehardening phases irrespective of their ecotype. Once light has an effect on the ontogenetic development during dehardening, the ambient daylength in Bergen (60° N) should favour the dormancy release of the ecotypes from higher latitudes. The hypothetical order to maintain frost hardiness is thus:

  • Order 3 (Hypothesis 3): NHa; FJ; Nme < IC < NB

In order to follow the dehardenining process we determined the frost hardiness (LT50) of stems of the mountain birch seedlings. In addition, we analysed lipid peroxidation (malondialdehyde) to investigate stress response in cell membranes, which are the primary site of freezing injury (Steponkus, 1990). Lipid peroxidation is a widely used stress indicator of plasma membranes (e.g. Taulavuori et al., 2001), which undergo both biochemical and biophysical changes during cold acclimation. The purpose of lipid peroxidation analysis in the present investigation was to provide additional information about the physiological state of stress attributed to cold hardiness. The final aim of this work was to evaluate the risk of the studied mountain birch ecotypes to frost damage under warming climate conditions.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Experimental design

Seeds from 6 to 8 trees per population from five different mountain birch (B. pubescens Ehrh. ssp. czerepanovii (Orlova)) ecotypes (Table 1, Fig. 1) were sown into the glasshouse of the University of Tromsø during 2 yrs, 2000 and 2001. In 2000 the seeds were stratified for 20 d and sown on 30 June, and in 2001 they were stratified for 28 d and sown 1 month later (6 August). The material in 2000 was kept at 18°C and in the ambient light conditions with a 24 h photoperiod until September. The material in 2001 was also kept at 18°C. The 24 h photoperiod was supplemented with an artificial lamp system (Osram 58 W/30; 200 µmol m−2 s−1) because of the late season for growth. On the 7 and 18 of September (2000 and 2001, respectively), the seedlings were potted in 10 cm containers with fertilized peat (Skre, 1993) and watered regularly once a week with SUBERBA nutrient solution equivalent to 10 g N m−2 yr−1 and placed at two autumnal temperatures (+9°C and +15°C) for hardening in the ambient daylight of Bergen (60° N).

On 1 and 22 November (in 2000 and 2001, respectively), the seedlings were transferred for overwintering to the Norwegian Forest Research Institute in Bergen. They were subdivided between two glasshouse compartments with different temperatures: either ambient (unheated) or approx. 4°C above ambient (heated) with an ambient day length of 6 h. The daylight intensity in the glasshouse was reduced by c. 20–50% relative to the outside. The daily minimum temperatures during the overwintering period are shown in Fig. 2.


Figure 2. Daily minimum temperatures in Bergen during the overwintering periods used in the experiments in (a) 2000–2001 and (b) 2001–2002. Day 1 denotes November 1. The lower curve stands for ambient and the upper curve for elevated temperatures.

Download figure to PowerPoint

To sum up, the experimental design consists of dehardening determinations during two springs. During both springs (2001 and 2002), a sample set containing 40 seedlings was transported three times to the University of Oulu (Finland) for further analysis. The sampling dates were: 18 January, 1 February and 22 March in 2001, and 24 January, 21 February and 4 April in 2002.

Frost hardiness

The frost hardiness test was similar to that described by Gansert et al. (1999) for another mountain birch species (Betula ermanii), with the exception of propanol infiltration, which was substituted by a standard shaking procedure in Millipore water. Main shoots were taken from below the top crown and exposed to the following temperatures in the following order: −10, −20, −40, −70 and −196°C (= LN) (Month 1 and 2), and −2, −10, −20, −40 and −196°C (Month 3). Stem pieces of 10–15 mm were moistened, wrapped in aluminium foil and frozen to the test temperature. The freeze-thaw procedure, the consequent electrolyte leakage and the calculation of LT50 were done as described earlier for Pinus sylvestris (Taulavuori et al., 1997b, 2000), Vaccinium myrtillus (Taulavuori et al., 1997, 1997a) and V. vitis-idaea (Taulavuori et al., 2001). The special adaptation of the method (preliminary results, unpublished data) to the stem material of birch species required: (1) dissection of buds after the freeze-thaw procedure to obtain a homogeneous organ (stem); (2) the remaining internode was cut into 3 segments of approx. 3–5 mm to accelerate the rate of electrolyte leakage out of the tissue.

Normalized hardiness at +4°C temperature elevation

Because of the climatic adaptations, the ecotypes experienced unequal temperature elevations in Bergen. Therefore, frost hardiness was normalized at the +4°C temperature elevation corresponding to the mean winter temperatures of each ecotype, in order to allow evaluation and comparison of the susceptibility of the ecotypes to frost injuries under conditions of global climatic warming. The normalization was performed as follows (Fig. 3):


Figure 3. An example of normalized hardiness (NH) at +4°C. The numerals in the parentheses refer to the text.

Download figure to PowerPoint

  • 1
    The ‘Ambient’ and ‘Elevated’ treatments were calculated to obtain the temperature elevation experienced by a given ecotype. The calculated temperature elevation was based on the difference between the mean temperatures in winter adapted for seed origin (Table 1) and the mean temperatures experienced in Bergen. The mean temperatures in Bergen were averages of the weekly mean values from the same period (December–March) as the reference values in Table 1. The long-term data from vicinity of origins of all the ecotypes concern the period 1931–1960, and provide an adequate basis to evaluate environmental adaptation (genetic) because of the long life span of trees. The resulting mean temperatures in Bergen were +5.5 and +2.1°C for the elevated and ambient temperature treatments in the winter of 2000–2001 and +6.9 and +3.5°C in the following winter, respectively.
  • 2
    The determined LT50 values of the last sampling for each year (March and April in 2001 and 2002, respectively) were plotted against the true temperature elevation obtained by this method.
  • 3
    Normalized hardiness (NH) was obtained through extrapolation of the response line to intercept the true temperature elevation at +4°C.
  • 4
    Subsequent interpolation to the Y-axis.

Lipid peroxidation

Lipid peroxidation was analysed by the malondialdehyde (MDA) method (Hodges et al., 1999, Taulavuori et al., 2001), with the exception of minor modifications during homogenization. A 0.4 g sample was homogenized in liquid nitrogen with a mortar and pestle. The homogenized tissue powder was suspended in 5 ml of 0.1% TCA on ice, after addition of 0.6 g of PVPP (poly vinyl-poly pyrrolidone) in 4 ml of 0.1% TCA (trichloro acetic acid). Extraction was continued as described by Taulavuori et al. (2001), including the following steps: centrifugation, TCA/TBA (thiobarbituric acid) addition, heat/cool cycle and another round of centrifugation. The absorbance of supernatant was read at 440, 532 and 600 nm, and each sample had a reference without TBA. The MDA equivalents were calculated according to Hodges et al. (1999).

Statistical analyses

The statistical analyses were performed with the anova General Linear Model (SPSS software) stepwise as follows. The main effects of the factors: (1) year, (2) month, (3) seed origin (ecotype), (4) autumn temperature, and (5) winter temperature were tested first (Table 2). Since the temporal variation (yearly, monthly) in ambient temperature determines frost hardiness, and the temperature differed markedly between the years, further analysis of the data was performed using successive years as separate cases. Each month was analysed step by step. Since the autumn temperature (Table 2) showed no main effect on frost hardiness, the data based on the two autumn temperatures were pooled, thus resulting in a replicate number of four (n = 4). The data on lipid peroxidation were analysed similarly.

Table 2. anova results from the 2-year data of the main effects on frost hardiness on mountain birch (Betula pubescens ssp. czerepanovii) ecotypes
Source of variationSSd.f.F
  1. The F-values contain indication of significance as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.01; NS, nonsignificant.

Corrected model (R2 = 0.88)113817.32119   6.95***
Intercept217990.17  11583.19***
Y (year) 34767.46  1 252.50***
M (month) 12154.21  2  44.14***
E (ecotype)  8071.95  4  14.66***
AT (autumn temperature)   119.36  1   0.867 NS
WT (winter temperature)  8009.47  1  58.17***
× M   662.43  2   2.41 NS
× E  6940.14  4  12.60***
× AT   507.22  1   3.69 NS
× WT  6881.23  1  49.98***
× M × E  2228.50  8   2.02 NS
× M × AT   770.41  2   2.80 NS
× M × WT   597.34  2   2.17 NS
× AT × WT   125.98  1   0.92 NS
× E × AT  2756.49  4   5.00**
× E × WT  2065.76  4   3.75**
× M × E × AT  1009.30  8   0.92 NS
× M × E × WT  1634.85  8   1.48 NS
× M × AT × WT   389.49  2   1.41 NS
× E × AT × WT  1066.94  4   1.94 NS
× M × E × AT × WT  2751.66  8   2.50*
Month × E  2136.99  8   1.94 NS
× AT   281.48  2   1.02 NS
× WT  2197.90  2   7.98**
× E × AT  1598.54  8   1.45 NS
× E × WT  3689.69  8   3.35**
× AT × WT   805.36  2   2.93 NS
× E × AT × WT  2143.44  8   1.95 NS
× AT  1077.12  4   1.96 NS
× WT  1360.43  4   2.47*
× AT × WT   464.49  4   0.84 NS
Autumn temperature
× WT   517.67  1   3.77 NS
Error 15696.74114 
Corrected total129514.06233 


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Main effects on frost hardiness

The anova table (Table 2) shows the highly significant effects of temporal variation (year, month) on frost hardiness. For example, the averaged frost hardiness based on the complete data of 2001 and 2002 shows LT50 = −44 and −19°C, respectively. The seedlings were thus twice as hardy during the first studied year compared to the second. This reflects the higher winter temperature of approx. 2°C in 2002 and the younger physiological age of the seedlings in that year. In addition, the last sampling in 2002 occurred c. 2 weeks later. Frost hardiness decreased throughout the experiments in both years, indicating the degree of dehardening as follows: LT50 =−49 (January 2001) and −29°C (January 2002), LT50 = −47 (February 2001) and −18.6°C (February 2002), and LT50 =−35 (March 2001) and −9°C (April 2002). The difference between the autumn temperatures (+9 and +15°C) had no effect on frost hardiness, although it interacted with ecotype and year partly due to the above reasons. As a consequence of interaction, the northernmost ecotypes (NHa and FJ) dehardened easier in 2001 when they first experienced the lower growing temperature before overwintering at elevated winter temperatures. However, winter temperature and seed origin (i.e. ecotype) showed a highly significant effect on frost hardiness. The approx. 4°C elevation in winter temperature from ambient levels in Bergen thus reduced the frost hardiness of mountain birch seedlings. The effects of the seed population were variable. The 2- year patterning of the ecotypes (from the most dehardened to hardened) was as shown in Order 4 (differences at P < 0.5).

  • Order 4 NB ≤ NHa, FJ, NMe ≤ IC

The Iceland ecotype (IC) thus maintained its frost hardiness most conservatively, while the NB ecotype from the nearby latitude (63° N), but from a much higher altitude, dehardened most rapidly.

Spring 2001

The elevated winter temperature resulted in reduced frost hardiness at P < 0.01 in January (Fig. 4a), but there were no differences between the ecotypes. Two week later, at the beginning of February, the effect of elevated winter temperature remained significant (P < 0.01) and the effect of seed origin had become visible (P < 0.01) (Fig. 4b). At this stage, the ecotypes NB, FJ and NHa appeared to have dehardened most rapidly. At the end of March (week 26), the effect of winter temperature was most marked (P < 0.001). In addition, there was a significant seed origin–winter temperature interaction (P < 0.01) as the ecotypes IC, NMe and FJ displayed a high level of frost hardiness at ambient temperature (i.e. LT50 > −60°C), while the frost hardiness of each ecotype had gone up to −20°C or more at the elevated temperature (Fig. 4c). The effect of seed origin was also greatest at this time, showing that the ecotypes NB (i.e. the southernmost ecotype from high altitude) and NHa (i.e. the northernmost ecotype near the sea level) had dehardened most rapidly.


Figure 4. Frost hardiness (mean LT50, °C below zero ± se, n = 4) of the mountain birch (Betula pubescens ssp. czerepanovii) ecotypes during the spring of 2001: (a) January (b) February and (c) March. The different small letters indicate the statistical difference between the seed origins (ecotypes) at P < 0.05. The abbreviated seed origins are as in Table 1.

Download figure to PowerPoint

Lipid peroxidation measured as malondialdehyde content was greatest in Jan (Fig. 5). It differed from the content assessed at the two latter dates at P < 0.05. Interestingly, though no statistical difference emerged, lipid peroxidation tended to be less pronounced in populations with the highest degree of frost hardiness (IC ≤ FJ ≤ NB ≤ NMe ≤ NHa).


Figure 5. Lipid peroxidation (mean ± se, n = 4) of mountain birch (Betula pubescens ssp. czerepanovii) ecotypes during the spring of 2001: (a) January (b) February and (c) March. The abbreviated seed origins are as in Table 1.

Download figure to PowerPoint

Spring 2002

Frost hardiness was unaffected by either the winter temperature or the seed origin in January (Fig. 6a). The elevated winter temperature appeared to have decreased frost hardiness only in April (week 15) (Fig. 6c). In addition, seed origin did not affect frost hardiness except marginally (P < 0.1) in January. As a whole, however, the southernmost population from the high altitude (NB) dehardened most rapidly, in accordance with the dehardening in 2001. The greatest difference between the years due to seed origin was seen in the response of the Iceland population: these seedlings were the most conservative in 2001 and nonconservative in 2002 in maintaining their high level of frost hardiness.


Figure 6. Frost hardiness (mean LT50, °C below zero ± se, n = 4) of the mountain birch (Betula pubescens ssp. czerepanovii) ecotypes during the spring of 2002: (a) January (b) February and (c) April. Abbreviated seed origins are as in Table 1.

Download figure to PowerPoint

Lipid peroxidation content followed a similar pattern as in the spring of 2001. The only statistically significant response was found in response to sampling date (Fig. 7). As in 2001, lipid peroxidation decreased from January (P < 0.05). In addition, a slight recovery in lipid peroxidation, as in 2001, was also observed in February and April. Moreover, the rate of lipid peroxidation was lowest in the most frost-hardy populations (NHa ≤ NMe ≤ NB ≤ FJ ≤ IC), as in the previous year.


Figure 7. Lipid peroxidation (mean ± se, n = 4) of the mountain birch (Betula pubescens ssp. czerepanovii) ecotypes during the spring of 2002: (a) January (b) February and (c) April. The abbreviated seed origins are as in Table 1.

Download figure to PowerPoint

Normalized hardiness

Normalization of the frost hardiness data (Table 3) resulted in the following pattern of responses to a temperature elevation of +4°C. In the winter of 2000–2001, the frost hardiness order of the ecotypes (from the most dehardened to hardened) was as shown in Order 5. In the winter of 2001–2002, the order was almost the same (Order 6). The result indicates that the northernmost ecotypes were most tolerant and the southernmost ecotypes least tolerant against a +4°C elevation in winter temperatures compared to the temperature they were adapted to:

Table 3.  Temperature elevations experienced by the mountain birch (Betula pubescens ssp. czerepanovii) ecotypes in the experiment and the respective normalized hardiness corresponding to +4°C temperature elevations of a given ecotype
EcotypeT elevation (±°C)Normalized hardiness (–°C)
NB 9.8 6.4   14
IC 5.5 2.1   46
NMe 4.9 1.5   32
FJ17.414< 100
NHa 9.7 6.3   55
NB11.2 7.8    1
IC 6.9 3.5   12
NMe 7.5 4.1   18
FJ18.815.4   43
NHa11.3 7.7   13
  • Order 5 NB < Nme < IC < NHa < FJ
  • Order 6 NB < IC; NHa < Nme < FJ


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The present study demonstrates that a 4°C rise in winter temperature may significantly accelerate the dehardening process of mountain birch in spring. In accordance, accelerated dehardening by rising winter temperature was also reported in Scots pine (Repo et al., 1996) and bilberry (Taulavuori et al., 1997a). Elevated winter temperatures may result in either premature or delayed bud break (e.g. Murray et al., 1989). The former is due to accumulation of day degrees after the fulfilled chilling requirement (Hänninen, 1995), while the latter is due to incomplete fulfilment of the chilling requirement at excessively high winter temperatures (Murray et al., 1989). Such a chilling deficit and subsequent delay in bud break is unlikely in Scandinavia, and the likely effects of climatic warming include earlier bud burst, a longer growing season and an increased risk of spring frost injury (Myking & Heide, 1995). It has to be emphasized that the chilling requirement was not measured during the present study. However, there is evidence that two basic mechanisms control frost hardiness: long-term control through the annual ontogenetic cycle (Fuchigami et al., 1982; Leinonen, 1996a) and short-term control through direct environmental stimuli (e.g. Sakai & Larcher, 1987). Because of the long-term control of frost hardiness, it is understandable that the present results are in line with the hypotheses based on the chilling requirement.

A superficial evaluation of the ecotype response to dehardening (Order 4) under the investigated conditions closely follows the pattern presented as Hypothesis 1. The only exception is the ecotype FJ, which was not so susceptible to dehardening as was expected (Orders 1 and 2). However, the IC ecotype maintained its frost hardiness most effectively, as was expected. In turn, the most alpine ecotype NB had the highest rate of dehardening, in accordance with Hypothesis 1. The highest dehardening rate in the NB ecotype is in contradiction with Hypothesis 3. However, this does not exclude that daylength could affect the dehardening. Indeed, the dehardening order (Order 4) of the other ecotypes was as expected according to Hypothesis 3. The result is thus consistent with the stabilized opinion that rising temperature is the dominating factor for dehardening (e.g. Fuchigami et al., 1982), although it suggests that the photoperiod may also slightly modify the process (e.g. Linkosalo et al., 2000).

Because of the ontogenetic control of frost hardiness, the autumn photoperiod may have also affected the obtained result, since the northernmost ecotypes in the photoperiod of Bergen probably first entered into dormancy as a consequence of their longer critical daylength for the induction of dormancy (Cannell, 1990; Myking & Heide, 1995). Consequently, accumulation of chilling temperature and subsequent fulfilment of the chilling requirement started earlier in the northern ecotypes. This explains the significant interaction between the hardening temperature and ecotype in 2001: the northernmost ecotypes (NHa and FJ) dehardened most rapidly in the treatment with low hardening temperature (+9°C) and elevated winter temperature. In other words, the photoperiod of Bergen triggered soon the development of dormancy in the northernmost ecotypes, which started to accumulate chilling temperature. Once the chilling requirement had been met, the effect of elevated winter temperature became obvious. The ecotype NB made an exception from this pattern since it dehardened most effectively, probably as a consequence of the alpine climate it had adapted to, in accordance with Hypothesis 1.

The result reflects the response of each ecotype to given experimental manipulation, but does not provide information for evaluating the consequences of global climate warming. However, long-term temperature data from areas of the studied ecotypes are of great help in this evaluation. The temperature values shown in Table 1 illustrate well the wintertime climatic conditions to which the ecotypes have been adapted. Although the data concern only the period between 1931 and 1960, and do not extend to the present, it is nevertheless valid for two reasons. First, the decades 1930–1940 were warm in Scandinavia, followed by cooler periods until the 1980s, and since the 1990s warming had started again (Kullman, 2001). Cooler and warmer periods undoubtedly equalize the temperature differences of the whole period from 1930 to 2000, and provide a good, though approximate basis for considering environmental adaptation. Secondly, the ecotypes result from genetic adaptation, where 40 year (i.e. the lack of data from the 1960s to the 2000 s) is a short period, especially because of the longevity of trees. When these temperature parameters are compared to temperatures recorded during the experiment (Fig. 3), it is obvious that the ecotypes experienced the elevated temperatures unequally (Table 3). The ecotypes IC and NMe, in the elevated temperature conditions of Bergen, overwintered at temperatures of 4–7°C above the temperature where they have been adapted. The temperature elevation was exactly the same as predicted by the climatic change scenarios for the boreal zone in the near future (Bach, 1988; Maxwell, 1992; Murray, 1995; Houghton et al., 2001). The ambient temperature in Bergen as such provided a temperature elevation within the same range for the ecotypes NB and NHa, while the elevated temperature in Bergen provided conditions warmer than could be expected within the near future. However, the ecotype FJ overwintered at an unrealistic temperature elevation irrespective of the temperature treatment in Bergen.

The normalized hardiness (Fig. 2, Table 3) provides a tool for relating the present results to climatic warming. The ecotype NB seems most susceptible to climatic warming by a +4°C temperature elevation, while the northernmost ecotypes NHa and FJ were relatively well buffered against it, especially the FJ exhibiting the ecotype from the coldest climate within this study (Table 2). This is obviously due to the adaptation of this ecotype to northern and relatively continental conditions at a relatively high altitude, which altogether provide a wide range of annual temperature changes. The significance of the +4°C elevation is thus relatively smaller for an ecotype adapted to a harsh rather than a mild climate. It should be emphasized, however, that climatic warming is most intensive in the northernmost areas (Bach, 1988; Maxwell, 1992; Murray, 1995; Houghton et al., 2001), i.e. equalizing somewhat the magnitude of its effects between the ecotypes.

During the first year of the experiment (i.e. 2000) the growth period was a month longer than it was the next year. Cool periods during the growing season, i.e. unfavourable growing conditions, may disturb development and ripening of the overwintering tissues of Japanese mountain birch (Betula ermanii) (Gansert, 2002). Consistently, the frost hardiness of each studied ecotype was poor (Fig. 6) in January 2002, obviously due to unfavourable growing conditions, although the winter was also approx. 1.5°C milder (Fig. 3). However, the latter cannot explain the low degree of frost hardiness in 2002, since frost hardiness in elevated temperature treatment (approx. +5.5°C) in 2001 was greater than in ambient conditions in 2002 (approx. +3.5°C), for example.

Seasonal changes in lipid peroxidation patterns were high during the winter and tended to decrease in the spring (Cakmak et al., 1995). Consistently, our data show the highest rate of lipid peroxidation in January in both years (Figs 5a and 7a). The purpose of the hardened state, however, is to stabilize and alleviate the increase in lipid peroxidation caused by freezing stress conditions (Wang et al., 1995; Yang et al., 2002). Therefore, a low degree of springtime lipid peroxidation indicates a high degree of frost hardiness in relation to temperature experienced, i.e. low stress. It is thus understandable that the ecotypes IC and FJ with maximum frost hardiness (Fig. 4c) also exhibited the lowest level of lipid peroxidation in 2001 (Fig. 5c). In addition, NHa, being one of the most dehardened ecotypes, showed the highest degree of lipid peroxidation. In accordance with high degree of lipid peroxidation in NHa in March 2001, freezing stress may peak in spring, which is able to induce a temporary increase in lipid peroxidation (Polle et al., 1996; Zhou & Leul, 1998; Benson et al., 1999). On the other hand, in 2002, the ecotype IC belonged to the rapidly dehardened population and showed the highest degree of lipid peroxidation. Lipid peroxidation analysis thus generally supports the dehardening findings based on the determinations of hardiness.

To sum up, an interpretation of the results based on statistical tests shows that the different mountain birch ecotypes dehardened in line with Hypothesis 1 and the presumed chilling requirement. The most alpine and southern ecotype (NB) exhibited the greatest, and the other southern ecotype (IC) from the mild climate the lowest, rate in the dehardening process. However, projection of the results on the difference between the adapted and experienced temperatures indicates that the risk of frost injury is lowest in the northernmost ecotypes under a temperature elevation of +4°C. This indicates adaptability of these ecotypes to a wider range of temperatures, including the absolute minima. Moreover, the most alpine ecotype (NB) was among the populations with the most probable risk of frost injury, regardless of which of the above perspectives was used to interpret the results. The results on all the ecotypes in 2002 demonstrate that a low level of frost hardiness and a respective high risk of frost damage may occur after an unfavourable growing season.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This work has been part of the HIBECO project, which was funded by the EU 5th Framework Programme. We also thank Prof. Heikki Hänninen for valuable comments. Mr Keith Kosola as a native English speaker is thanked for revision of the language.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Bach W. 1988. Development of climatic scenarios: from general to circulation models. In: ParryML, Carter TR, Konjin NT, eds. The impact of climatic variations on agriculture: assessment of cool and temperate regions, Vol. 1. Dortdrecht, The Netherlands: Kluwer Academic Publishers, 125157.
  • Benson EE, Lynch PT, Jones J. 1999. The detection of lipid peroxidation products in cryoprotected and frozen rice cells: consequences to post-thaw survival. Plant Science 85: 107114.
  • Cakmak I, Atli M, Kaya R, Evliya H, Marschner H. 1995. Association of high light and zinc deficiency in cold induced leaf chlorosis in grapefruit and mandarin trees. Journal of Plant Physiology 146: 355360.
  • Cannell MGR. 1990. Modelling the phenology of trees. Silva Genetica 15: 1127.
  • Cannell MGR, Smith RI. 1986. Climatic warming, spring budburst and frost damage on trees. Journal of Applied Ecology 23: 177191.
  • Fuchigami LH, Weiser CJ, Kobayashi K, Timmis R, Gusta LV. 1982. A degree growth stage (°GS) model and cold acclimation in temperate woody plants. In: LiPH, Sakai A, eds. Plant cold hardiness and freezing stress mechanism and crop implications, Vol. 2. New York, USA: Academic Press, 93116.
  • Gansert D. 2002. Betula ermanii, a dominant subalpine and subarctic treeline tree species in Japan: ecological traits of deciduous tree life in winter. Arctic, Antarctic and Alpine Research 34: 5764.
  • Gansert D, Backes K, Kakubari Y. 1999. Altitudinal and seasonal variation of frost resistance of Fagus crenata and Betula ermanii along the Pacific slope of Mt. Fuji, Japan. Journal of Ecology 87: 382390.
  • Häkkinen R, Linkosalo T, Hari P. 1998. Effects of dormancy and environmental factors on timing of bud burst in Betula pendula. Tree Physiology 18: 707712.
  • Hänninen H. 1991. Does climatic warming increase the risk of frost damage in northern trees? Plant, Cell & Environonment 14: 449454.
  • Hänninen H. 1995. Effects of climatic change on trees from cool and temperate regions: an ecophysiological approach to modelling of bud burst phenology. Canadian Journal of Botany 73: 183199.
  • Heide OM. 1993. Daylength and thermal time responses of budburst during dormancy release in some northern deciduous trees. Physiologia Plantarum 88: 531540.
  • Hodges DM, DeLong JM, Forney CF, Prange RK. 1999. Improving the thiobarbituric acid-reactive-substances assay for estimating lipid peroxidation from plant tissues containing anthocyanin and other interfering compounds. Planta 207: 604611.
  • Houghton JT, Ding Y, Griggs DJ, Noguer M, Van Der Linden PJ, Xiaosu D. 2001. Climate change 2001. In: The scientific basis contribution of working group I to the third assessment report of the intergovernmental panel on climate change (IPCC). Cambridge, UK: Cambridge University Press.
  • Kullman L. 1993. Tree limit dynamics of Betula pubescens ssp. tortuosa in relation to climate variability: evidence from central Sweden. Journal of Vegetational Science 4: 764772.
  • Kullman L. 2001. 20th century climate warming and tree-limit rise in the Southern Scandes of Sweden. Ambio 30: 7280.
  • Leinonen I. 1996a. A simulation model for the annual frost hardiness and freeze damage of Scots pine. Annals of Botany 78: 687693.
  • Leinonen I. 1996b. Dependence of dormancy release on temperature in different origins of Pinus sylvestris and Betula pendula seedlings. Scandinavian Journal of Forest Research 11: 122128.
  • Linkosalo T, Carter TR, Häkkinen R, Hari P. 2000. Predicting spring phenology and frost damage risk of Betula spp. under climatic warming: a comparison of two models. Tree Physiology 20: 11751182.
  • Maxwell B. 1992. Arctic Climate: Potential for Change under global warming. In: ChapinFS, Shaver GR, Svoboda J, Chu EW, eds. Arctic ecosystems in changing climate an ecophysiological perspective. San Diego, CA, USA: Academic Press, 1134.
  • Murray RD. 1995. Plant responses to carbon dioxide. American Journal of Botany 82: 690697.
  • Murray MB, Cannell MGR, Smith RI. 1989. Date of budburst of fifteen tree species in Britain following climatic warming. Journal of Applied Ecology 26: 693700.
  • Myking T, Heide O. 1995. Dormancy release and chilling requirement of buds of latitudinal ecotypes of Betula pendula and B. pubescens. Tree Physiology 15: 697704.
  • Ögren E. 2001. Effects of climatic warming on cold hardiness of some northern woody plants assessed from simulation experiments. Physiologia Plantarum 112: 7177.
  • Polle A, Kröniger W, Rennenberg H. 1996. Seasonal fluctuations of ascorbate-related enzymes: acute and delayed effects of late frost in spring on antioxidative systems in needles of Norway spruce (Picea abies L.). Plant Cell Physiology 37: 717725.
  • Pop EW, Oberbauer SF, Starr G. 2000. Predicting vegetative bud break in two arctic deciduous shrub species, Salix pulchra and Betula nana. Oecologia 124: 176184.
  • Repo T. 1992. Seasonal changes of frost hardiness in Picea abies and Pinus sylvestris. Finland. Canadian Journal of Forest Research 22: 19491957.
  • Repo T, Hänninen H, Kellomäki S. 1996. The effects of long-term elevation of air temperature and CO2 on the frost hardiness of Scots pine. Plant, Cell & Environment 19: 209216.
  • Repo T, Mäkelä A, Hänninen H. 1990. Modelling frost resistance of trees. In: JozefekH, ed. Modelling to understand forest functions. Silva Carelica 15: 6174.
  • Sakai A, Larcher W. 1987. Frost Survival of Plants. Responses and Adaptation to Freezing Stress. In: BillingsWD, Golley F, Lange OL, Olson JS, Remmert H, eds. Ecological studies 62. Berlin, Germany: Springer Verlag.
  • Skre O. 1993. Growth of Mountain Birch (Betula pubescens Ehrh.) in Response to Changing Temperature. In: AldenJ, Mastrantonio JL, Ödum S, eds. Forest development in cold climates. London, UK: Plenum Press, 6578.
  • Steponkus PL. 1990. Cold Acclimation and Freezing Injury from a Perspective of the Plasma Membrane. In: KattermanF, ed. Environmetal injury to plants. San Diego, CA, USA: Academic Press, –16.
  • Taulavuori E, Hellström E-K, Taulavuori K, Laine K. 2001. Comparison of two methods used to analyse lipid peroxidation from Vaccinium myrtillus (L.) during snow removal, reacclimation and cold acclimation. Journal of Experimental Botany 52: 23752380.
  • Taulavuori K, Laine K, Taulavuori E, Pakonen T, Saari E. 1997a. Accelerated dehardening in the bilberry (Vaccinium myrtillus L.) induced by a small elevation in air temperature. Environmental Pollution 98: 9195.
  • Taulavuori K, Niinimaa A, Laine K, Taulavuori E, Lähdesmäki P. 1997b. Modelling frost resistance of Scots pine seedlings using temperature, daylength and pH of cell effusate. Plant Ecology 133: 181189.
  • Taulavuori E, Taulavuori K, Laine K, Saari E, Pakonen T. 1997. Winter hardening and glutathione status in the bilberry (Vaccinium myrtillus L.) in response to trace gases (CO2, O3) and nitrogen fertilization. Physiologia Plantarum 101: 192189.
  • Taulavuori K, Taulavuori E, Niinimaa A, Laine K. 2001. Acceleration of frost hardening in Vaccinium vitis-ideae (L.) by nitrogen fertilization. Oecologia 127: 321323.
  • Taulavuori K, Taulavuori E, Sarjala T, Savonen E-M, Pietiläinen P, Lähdesmäki P, Laine K. 2000. In vivo chlorophyll fluorescence is not always a good indicator of cold hardiness. Journal of Plant Physiology 157: 227229.
  • Tenow O. 1996. Hazards to a mountain birch forest – Abisko perspective. Ecological Bulletin 45: 104114.
  • Wang YR, Zeng SX, Liu HX. 1995. Effect of cold hardening on SOD and glutathione reductase activities and the contents of the reduced form of glutathione and ascorbic acid in rice and cucumber seedlings. Acta Botanica Sinica 37: 776780.
  • Yang SC, Xie CT, Zhang P, Jiang XX, Liao QL, Ding YL. 2002. Effects of cold hardening on membrane lipid peroxidation and activities of cell defence enzymes in leaves of Pritchhardia gaudichaudii seedling under low temperature stress. Journal of Plant Resource and Environment 11: 2528.
  • Zalasky H. 1976. Structural changes in tissues of caragana seedlings after frost damage. Canadian Journal of Plant Science 56: 941945.
  • Zhou W, Leul M. 1998. Uniconazole-induced alleviation of freezing injury in relation to changes in hormonal balance, enzyme activities and lipid peroxidation in winter rape. Plant Growth Regulation 26: 4147.