Present address: Department of Ecology and Environmental Science, Umeå University, Box 62, SE-98107 Abisko, Sweden.
Disentangling effects of an experimentally imposed extreme temperature event and naturally associated desiccation on Arctic tundra
Article first published online: 16 OCT 2006
Volume 20, Issue 6, pages 917–928, December 2006
How to Cite
MARCHAND, F. L., VERLINDEN, M., KOCKELBERGH, F., GRAAE, B. J., BEYENS, L. and NIJS, I. (2006), Disentangling effects of an experimentally imposed extreme temperature event and naturally associated desiccation on Arctic tundra. Functional Ecology, 20: 917–928. doi: 10.1111/j.1365-2435.2006.01203.x
- Issue published online: 16 OCT 2006
- Article first published online: 16 OCT 2006
- Received 24 May 2006; revised 1 August 2006; accepted 15 August 2006 Editor: O. Atkin
- arctic tundra;
- chlorophyll fluorescence;
- global change;
- heat stress;
- stomatal conductance
- 1Climate projections suggest that extreme events will increase in frequency during this century. As tundra is recognized to be among the most vulnerable biomes, we exposed patches of arctic tundra vegetation to an experimental heatwave (by infrared irradiation), followed by a recovery period. The heating increased the surface temperature with an average of 7·6 °C during 13 days, which slightly exceeded the longest climatic episode with such a temperature deviation since 1961.
- 2The heatwave decreased stomatal conductance (gs) and PSII maximum efficiency (Fv/Fm), although there were differences in response among the four target species. Salix arctica Pall. (shrub) was affected during the heatwave and could not recover. In Carex bigelowii Tor. ex Schwein (sedge) and Pyrola grandiflora Radius (forb), on the other hand, the effects on gs and Fv/Fm became clear, particularly in the aftermath of the heatwave, whereas Polygonum viviparum L. (forb) was never stressed.
- 3Effects of the heat on gs were mainly indirect, through increased desiccation, whereas effects on Fv/Fm were more related to leaf temperature (although not in all species). The observed changes can therefore probably be ascribed to a combination of heat and drought causing dysfunctions that ultimately led to senescence.
- 4Two conclusions of this study, species-specific responses and increased leaf mortality, indicate that more frequent extreme temperature events accompanied by desiccation might alter/endanger tundra communities in a future climate. Predictions of global change effects on arctic ecosystems should therefore take into account the impact of extremes.
Climate warming has been a major preoccupation of ecologists during past decades. As it is widely recognized that polar regions are more vulnerable to warming than is the rest of the Earth (Przybylak 2002), arctic tundra ecosystems are expected to undergo the earliest and potentially most devastating impacts. In addition to increases in mean air and surface temperature, projections suggest that extreme climate events (e.g. hot days, heatwaves) will also increase in frequency during the 21st century (Wigley 1985; Easterling et al. 2000b; Houghton et al. 2001; Weisheimer & Palmer 2005). For example, in Britain by 2060–2100 the probability of hot spells longer than 7 days with daily maximum temperature >25 °C is predicted to rise from its present value of 2% per summer to 10% per summer for summers with high soil moisture, and to 80% per summer when summers with low soil moisture are included (Brabson et al. 2005). Changes in heat extremes are likely to occur first, and to be greatest at high latitudes (Maxwell 1992; Przybylak 2002). Trend changes in extreme high-temperature events have been demonstrated in some parts of the world by historical analysis, for example Gruza et al. (1999) found a significant increase in the number of days with exceptionally high temperatures in the course of the 20th century in Russia (anomalies being defined by a 10% or less frequency of occurrence). Plummer et al. (1999) observed that warming from 1941–91 resulted in two more days per year exceeding 30 °C in New Zealand, and similar increases in Australia. Despite growing awareness of the importance of climate variability, only a few studies have investigated effects of extremes on natural systems. Field experiments on grasslands and other herbaceous plant communities at temperate latitudes have revealed that extreme heat and drought can affect a range of community processes and characteristics, including competition, biological invasions and plant diversity (MacGillivray et al. 1995; White et al. 2000, 2001; Van Peer et al. 2001). Laboratory studies have confirmed that realistic high-temperature events can induce mortality (e.g. in several alpine plant species, Gauslaa 1984; and in temperate tree seedlings, Bassow, McConnaughay & Bazzaz 1994). Although few data exist on tundra, it appears that this biome will also be strongly influenced by extremes of weather and climate (Easterling et al. 2000b). By applying an artifically induced extreme heat event in High-Arctic Greenlandic tundra, Marchand et al. (2005) showed that plant responses first improved, but deteriorated after exposure. Also in Greenland, but in an experiment at a Low-Arctic site with two consecutive heatwaves and in a different year, Marchand et al. (2006) observed much more species-specific responses, with only some species showing the reverse pattern. This would confirm plant responses to manipulations of mean air temperature, which tend to vary with species (Henry & Molau 1997; Callaghan et al. 2004). The heatwave(s) did not substantially modify soil water availability in either of the Marchand et al. studies. Here we report a third experiment on arctic tundra, in which the added infrared irradiation did induce appreciable additional summer water shortage.
When exposed to air temperatures above their preferred range, plants can maintain tissue integrity either through biochemical protection or by altering radiation exchange (e.g. through strongly reflecting surfaces or orientation of leaves parallel to the sun's rays) (Nilsen & Orcutt 1996). Above particular temperature thresholds, however, biological processes undergo sudden shifts as a consequence of irreversible denaturation of protein complexes or changes in the characteristics of membranes. Photosynthesis is among the processes sensitive to high temperature, for example it is well known that heat-induced structural changes to the thylakoid membranes can block photosystem II (PSII) (Havaux, Ernez & Lannoye 1988). An important aspect of heat extremes is associated excessive drought (Easterling et al. 2000a), which could affect survival indirectly (Callaghan & Carlsson 1997). Whereas mild drought induces temporary closure of stomata, allowing maintenance of leaf relative water content within the limits where photosynthetic capacity is not impaired, extreme drought induces enduring stomatal closure and loss of photosynthetic and growth capacity (Yordanov, Velikova & Tsonev 2003). Moreover, prolonged desiccation can reinforce photoinhibition and photodamage induced by high temperatures (Seddon & Cheshire 2001). This synergistic effect might further impair plant survival during an extreme temperature event.
Here we report on an arctic tundra plant community that was exposed to a single, experimentally induced heatwave in the field in 2003, when prevailing temperature and moisture conditions gave rise to aggravated soil drought associated with the warming. The same site (but different plots) were used as described by Marchand et al. (2005). The treatment recreated one of the most severe extreme temperature events since 1961, and simulated both the intensity (average temperature) and duration of this warm period. Possible stress effects were detected with stomatal conductance and chlorophyll fluorescence measurements. Our goals were: (i) to disentangle the effects of heat and drought, and (ii) to elucidate the species-specific nature of the responses, given the equivocal conclusions of earlier studies (Marchand et al. 2005, 2006).
Materials and methods
site description and set-up
The study was conducted in the summer of 2003 near the Arctic Station of the University of Copenhagen, located on Disko Island in West Greenland (69°15′ N, 53°34′ W). This region of continuous permafrost has an arctic maritime climate with an annual air temperature and precipitation of −3·9 °C and 447 mm, respectively (averages between 1960 and 1991; Nielsen, Humlum & Hansen 2001). The arctic shrub vegetation in the region has a high plant diversity, including 212 of the 513 vascular plant species growing on Greenland. Six plots (40 × 50 cm) were selected with Salix arctica Pall., Polygonum viviparum L., Pyrola grandiflora Radius and Carex bigelowii Tor. ex Schwein as target species for physiological and growth measurements. Other species in the community, not used for measurements due to too small leaves or insufficient abundance, were Vaccinium uliginosum L., Empetrum nigrum L., Betula nana L. and mosses. Cover and species composition in the plots were determined three times, once before and once after the simulated heatwave (day of year, DOY 181 and 195, respectively) and at the end of the subsequent recovery period (DOY 206), using a 500-point frame. On each point (homogeneously distributed across the frame), the species intercepting a vertical laser beam was recorded. The six plots were appointed to two temperature groups to have similar cover and species composition before heating began (manova of cover per species with treatment as fixed factor, P > 0·05, F1,4 0·003–2·283). One group of three plots was heated from 1 July (DOY 182) to 13 July 2003 (DOY 194) with a constant flux density of infrared irradiation (0·8–3 µm), while the other group served as control. To elevate vegetation temperature (Tveg), the free air-temperature increase (FATI) method was used, designed to homogeneously warm limited areas of short vegetation (<30 cm, cf. Nijs et al. 2000 for technical details). This method ensures uniform heating along the vertical profile, in a horizontal plane and with time, in low-stature vegetation (Nijs et al. 1996). Each heated plot had an individual FATI unit placed on the north side, whereas the control plots had dummy units without lamps. During the 13-day heating period, the target increment of Tveg was 8 °C (see Results for justification).
Each 30 min between 13 : 30 local daylight time (LDT) on 29 June (DOY 180) and 10 : 00 LDT on 29 July (DOY 210), dataloggers (16 kb, 12-bit, eight-channel; DL2E, Delta T, Cambridge, UK) recorded seven climate parameters: photon flux density (PFD) of photosynthetically active radiation, measured with a gallium arsenide sensor (JYP-1000, SDEC, Reignac sur Indre, France); Tveg, measured with non-contact semiconductor sensors (‘infracouple’, type OS39-MVC-6, Omega Engineering, Stamford, CT, USA); air temperature (Tair) at 5 cm height; and soil temperature (Tsoil) at 2·5, 7·5, 15 and 30 cm depth, all measured with precision centigrade temperature sensors (LM35A, National Semiconductor, Arlington, TX, USA). These readings were taken with separate sensors in each plot, except for PFD which was measured at a single location close to the plots. During the same period, soil moisture was measured 10 times with time-domain reflectometry (Trime-FM, Eijkelkamp Agrisearch Equipment, Giesbeek, the Netherlands) and depth of the active layer four times with a fibreglass rod (both recorded at four locations per plot, one per quadrant).
Every 2–3 days the leaf area of the youngest leaf was calculated from its length and width, in five plants per target species per plot. The formula of an ellipse was used in S. arctica, P. viviparum and P. grandiflora, and length × width (measured at half-length of the leaf) in C. bigelowii. Consecutively, the relative leaf extension rate (RLE) of the newly appearing leaf area on the plant (P. viviparum and P. grandiflora), twig (S. arctica) or shoot (C. bigelowii) was determined for each time interval, taking into account the initiation of new leaves. The basis for this RLE calculation was, in each case, the leaf area of the growing leaf, so the area of previously formed leaves that had already stopped growing at the start of the measurements was excluded. Stomatal conductance (gs) was measured with the same frequency as RLE on three randomly chosen leaves per target species per plot using an automatic porometer (Mk II, Delta-T Devices, Cambridge, UK). There were three measurement cycles for gs per day (rotating between plots), from 09 : 00 to 11 : 00, 14 : 00–16 : 00 and 19 : 00–21 : 00 LDT. Simultaneously, modulated chlorophyll fluorescence was measured with a Plant Efficiency Analyzer (Hansatech Ltd, King's Lynn, Norfolk, UK), on five randomly chosen leaves per target species per plot, which were dark-adapted (leaf clips) for 30 min. Leaves were exposed to a saturating pulse (>3000 µmol m−2 s−1) for 1 s, yielding maximum fluorescence (Fm). Minimum fluorescence (F0) was determined from the initial fluorescence on illumination. The PSII maximum efficiency (Fv/Fm = (Fm – F0)/Fm), a sensitive measure of potential non-cyclic photosynthetic electron transport, was calculated from these recordings (Schreiber, Bilger & Neubauer 1994). Optimal Fv/Fm = 0·832 ± 0·004 and is almost constant for many different plant species and ecotypes (Björkman & Demmig 1987). Lower values arise on stress exposure, indicating photoinhibition (Maxwell & Johnson 2000). As temperature may influence Fv/Fm values, instantaneous leaf temperature was measured with an infrared thermometer (C-1600, Calex, Bedfordshire, UK) after every Fv/Fm measurement in all target species, except in C. bigelowii, which has leaves too narrow for accurate measurement.
Repeated-measures (RM) anova was used to analyse the effect of the heatwave (treatment as fixed factor) on variables at plot level (e.g. the environmental parameters, cover or mortality) or on variables at species level if data were collected on the same plant (e.g. RLE). DOY was the repeated factor. When a significant interaction between treatment and DOY was found, a manova was used to analyse the treatment effect on the different days separately. Effects on Fv/Fm and gs were analysed with univariate anova, or with univariate analysis of covariance (ancova) in case a covariate could influence the dependent variable (e.g. PFD, soil moisture). When interaction between species and treatment was found, separate anovas or ancovas were carried out for each species. For post hoc multiple comparisons between observed means, the conservative Bonferonni (if RM-anova) or Tukey HSD (if anova or ancova) tests were used. In all tests, plot was nested within treatment to avoid pseudoreplication. All variables were examined for normality and heterogeneity of variance and all statistical tests were considered significant at P < 0·05. Analyses were performed using spss ver. 10·0 (SPSS Inc., Chicago, IL, USA).
During the past 15 years the average July air temperature recorded by the Arctic Station's meteorological station equalled 7·1 ± 1·1(SD) °C (data from 1990–2004; R. Erjnaes, personal communication). As the average air temperature for July 2003 was 7·4 °C, background temperatures during the experiment were close to average. To determine a realistic duration for the simulated extreme event, we compared the daily mean air temperatures between 1961 and 2004 (data between 1961 and 1989 from the Danish Meteorological Institute at Ilulissat, 60 km from the study site; data between 1990 and 2004 from the Arctic Station; R. Erjnaes, personal communication) with the temperatures in the heated plots. A direct comparison, however, would be biased because (i) Tair measured in the plots at 5 cm height is different from the air temperature measured at 2 m height by the meteorological station of the Arctic station (Tair,station), and (ii) because Tair under free air temperature increase is less representative than Tveg (the latter is controlled directly by the heaters, while the former is distorted by the outside air crossing the plots; Nijs et al. 1996). For these reasons, we first carried out a regression between Tveg measured in the plots under ambient conditions and Tair,station (linear regression, P < 0·05, F1,984 = 268·4, r2 = 0·21). Using this regression, we calculated the Tair,station that corresponded with the on average 7·6 ± 2·2 °C higher Tveg recorded in the heated plots during the period of heating. This reconstructed Tair,station equalled 9·2 °C during the applied heatwave. The longest time interval between 1961 and 2004 with average Tair,station (including day and night values) exceeding 9·2 °C lasted for 12 days (summer 1978). Therefore we considered a period of 13 days’ heating as a realistic extreme temperature event.
During the heatwave, the separate FATI units increased the instantaneous Tveg by increments that approximated the target value of 8 °C (Table 1), and the instantaneous Tsoil close to the surface by increments of about 5 °C (soil warming decreased towards deeper layers, Fig. 1a). During the recovery period, Tsoil values in both treatments became equal again and the steep Tsoil profile that characterized the heatwave largely disappeared. In both treatments, soil moisture was depleted during the heatwave period, and restored afterwards through substantial precipitation (Fig. 1b). Mean soil moisture was always lower in the warmed plots (RM-anova, P < 0·05, F1,18 = 35·3), except for the first and the last two measurement days (manova of soil moisture with treatment as fixed factor, P < 0·05, F1,18 4·76–24·7). Thawing depth was increased significantly during the experimental period, with 4·0 ± 1·3 cm above an initial value of 49·6 ± 3·5 cm in the control plots and with 7·5 ± 2·7 cm above an initial value of 49·4 ± 4·7 cm in the heated plots (RM-anova, P < 0·05, F3,48 = 7·92, Bonferroni). However, the 3·5-cm difference between treatments was not significant (RM-anova, P > 0·05, F1,16 = 1·39).
|Parameter||Tveg (°C)||ΔTveg (°C ± SD)||Tair (°C)||ΔTair (°C ± SD)|
|FATI 1||14·0||8·9 ± 2·5||9·3||3·6 ± 1·9|
|FATI 2||14·4||6·3 ± 2·2||9·7||3·9 ± 1·8|
|FATI 3||14·5||7·5 ± 1·8||9·2||3·6 ± 2·1|
|FATI 1||9·3||−0·2 ± 0·7||8·8||0·4 ± 0·8|
|FATI 2||8·9||0·2 ± 1·1||9·1||0·1 ± 0·9|
|FATI 3||10·3||0·3 ± 0·7||9·0||−0·1 ± 0·7|
The average cover of living vascular plants declined significantly in the heated plots (Fig. 2a, RM-anova, P < 0·05, F2,4 = 10·3, Bonferroni), but not in the unheated plots (P > 0·05, F2,4 = 4·60, Bonferroni). This was reflected in the substantial death caused by the exposure, whereas the percentage of dead material in the control plots remained constant and negligible (Fig. 2b, RM-anova, P < 0·05, F1,4 = 44·1). Mortality could be ascribed mainly to S. arctica and V. uliginosum (Fig. 2c, RM-anova of percentage dead cover by species, P < 0·05, F1,4 = 190·1 and 8·27, respectively), which were also the most abundant species (Fig. 2d). Next to leaf mortality, leaf growth could be a determining factor in the observed cover differences. However, RLE was not significantly altered by the heatwave in S. arctica and C. bigelowii (RM-anova, P > 0·05, F1,7 = 0·53, F1,10 = 0·61, respectively), where it amounted to on average 0·81 and 0·30 mm2 mm−2 day−1, respectively. In P. viviparum and P. grandiflora, on the other hand, RLE was lower in the heated plots (0·01 relative to 0·08 mm2 mm−2 day−1 and 0·01 relative to 0·03 mm2 mm−2 day−1; RM-anova, P < 0·05, F1,17 = 9·04 and F1,8 = 6·39, respectively), but this was probably caused by the observed curling of leaves.
Overall, the treatment affected Fv/Fm and gs during both the heatwave and the recovery period (Fig. 3). We first examined these effects with an anova and an ancova of Fv/Fm and gs, respectively, with treatment, plot (nested within treatment), DOY, species and time of day as between-subject factors and PFD as covariate with gs, for both the heatwave and the recovery period (Table 2). During the heatwave, PSII maximum efficiency usually decreased later in the day in all four target species (data not shown, Tukey HSD). Also, gs was generally lower later in the day, but only significantly during the recovery period and not in all species (data not shown, Tukey HSD). Because the responses to the warming treatment also varied with species (Table 2), we examined these effects using separate anovas and ancovas per species, with treatment, plot (nested within treatment) and DOY as between-subject factors and PFD as covariate in the case of gs (Table 3). This revealed that PSII maximum efficiency and stomatal conductance of P. viviparum leaves were not affected by the heatwave at any time (Fig. 3a,b), whereas C. bigelowii, P. grandiflora and S. arctica were affected (Fig. 3c–h). In C. bigelowii and P. grandiflora, both parameters did not decline until the recovery period, with the exception of gs in P. grandiflora, the decline of which already occurred during the heatwave (Table 3). By DOY 205, gs in C. bigelowii and P. grandiflora had recovered (manova per DOY and per species, P > 0·05, F1,18 = 0·42 and F1,47 = 4·05 for C. bigelowii and P. grandiflora, respectively, on DOY 205), whereas PSII maximum efficiency had not (manova, P < 0·05, F1,66 = 5·19 and 8·57, respectively; Fig. 3c–f). These two species exhibited Fv/Fm values already well below 0·832 at the start of the heatwave, indicating suboptimal functioning of the electron transport chain in ambient conditions. In S. arctica, on the other hand, both Fv/Fm and gs were reduced during the heatwave (Table 3), and were not re-established afterwards (manova, P < 0·05, F1,66 = 13·87 and F1,38 = 13·69 for Fv/Fm and gs, respectively, Fig. 3g,h).
|Parameter||Fv/Fm||gs (mol m−2 s−1)|
|Heatwave period||Recovery period||Heatwave period||Recovery period|
|PFD||F1,1207 = 4·241*||F1,633 = 0·028|
|DOY||F6,1708 = 9·261***||F6,1066 = 9·878***||F6,1207 = 30·939***||F6,633 = 4·708***|
|Species||F3,1708 = 125·708***||F3,1066 = 207·537***||F3,1207 = 128·536***||F3,633 = 88·108***|
|Treatment||F1,1708 = 28·064***||F1,1066 = 71·631***||F1,1207 = 23·225***||F1,633 = 31·195***|
|Plot(treatment)||F4,1708 = 24·962***||F4,1066 = 12·920***||F4,1207 = 6·340***||F4,633 = 1·669|
|Time of day||F3,1708 = 16·541***||F3,1066 = 2·595||F2,1207 = 1·594||F2,633 = 9·673***|
|Treatment × DOY||F6,1708 = 11·183***||F6,1066 = 1·507||F6,1207 = 3·568**||F6,633 = 0·370|
|Treatment × species||F3,1708 = 7·536***||F3,1066 = 8·023***||F3,1207 = 7·087***||F3,633 = 5·845**|
|Treatment × time of day||F3,1708 = 5·343**||F3,1066 = 0·344||F2,1207 = 1·125||F2,633 = 1·557|
|Species × time of day||F9,1708 = 0·870||F9,1066 = 1·769||F6,1207 = 3·171**||F6,633 = 7·648***|
|Species × DOY||F18,1708 = 1·531||F18,1066 = 2·081*||F18,1207 = 5·583***||F18,633 = 4·763***|
|DOY × time of day||F16,1708 = 3·470***||F16,1066 = 0·752||F12,1207 = 9·970***||F12,633 = 4·479***|
|Parameter||Fv/Fm||gs (mol m−2 s−1)|
|F||Polygonum viviparum||Carex bigelowii||Pyrola grandiflora||Salix arctica||F||Polygonum viviparum||Carex bigelowii||Pyrola grandiflora||Salix arctica|
|Treatment × DOY||F6,435||3·687**||1·697||5·700***||4·832**||F6,354||1·456||0·498||1·456||4·975***|
|Treatment × DOY||F6,273||2·492||0·049||3·364*||1·140||F6,182||0·926||0·686||0·169||0·382|
In the previous paragraph, the influence of treatment at any give time potentially consisted of four different effects: (i) instantaneous effect of imposed heat; (ii) instantaneous effect of the indirect change in soil moisture; and (iii, iv) the two cumulative effects of the heatwave (historical effects of exposure to heat or drought on the days preceding the measurement). Although the experiment was not designed to disentangle these components, we tried to separate them by carrying out ancovas of Fv/Fm or gs with both soil moisture and leaf temperature as covariates; species and treatment as fixed factors; plot (nested within treatment) and PFD as third covariate in the case of gs. Inclusion of multiple covariates was allowed because they were not highly correlated (Pearson's correlation, r between −0·19 and −0·34, Field 2000). Because interactions between treatment and species were significant (P < 0·05, F3,2096 = 18·42 and F3,2198 = 19·44 for gs and Fv/Fm, respectively), we repeated the ancova tests per species. The covariates soil moisture and leaf temperature reflect the aforementioned instantaneous components (i) and (ii) of the heatwave, as they were measured at the same time as gs and Fv/Fm. The covariate soil moisture interacted with treatment in gs (P < 0·05, F1,580 10·38–15·79 except for P. viviparum, P > 0·05, F1,556 = 1·53), whereas in Fv/Fm leaf temperature and treatment interacted (P < 0·05, F1,724−729 4·97–6·11). As a consequence, the regression slopes were not homogeneous, and separate analysis by treatment was therefore warranted in order to establish which covariates affect gs and Fv/Fm in each treatment (Glantz & Slinker 2001). On the other hand, any fixed factor (treatment) effects additional to these covariates could be determined. These include the cumulative effects of exposure to both heat and drought prior to the measurements (components iii and iv), although these cumulative effects cannot be separated in our analysis. In P. viviparum, no such treatment effect was found on gs or Fv/Fm (P > 0·05, F1,556 = 1·17 and F1,724 = 0·02 for gs and Fv/Fm, respectively), in accordance with the absence of a heatwave influence in Fig. 3. Surprisingly, in S. arctica and P. grandiflora, gs and Fv/Fm were no longer significantly affected by the treatment when the influence of the covariates was removed (P > 0·05, F1,556−729 0·00–3·36). This indicates that these species suffered only from the immediate effects of the heatwave (heat and/or drought). In contrast, the additional treatment effect did modify gs in C. bigelowii (P < 0·05, F1,379 = 28·58), indicating a long-term influence of the exposure.
Stepwise multiple linear regressions were then conducted to determine the effects of the covariates per treatment. Regressions per treatment were carried out for each species, with soil moisture, leaf temperature and PFD (for gs only) as predictors for Fv/Fm and gs. Combining these predictors in one analysis was allowed, as multicollinearity was absent, correlations were low, and the variance inflation factor <2 and tolerance >0·5 in all cases (Field et al. 2000). In P. viviparum and C. bigelowii (all plots) and in P. grandiflora and S. arctica (unheated plots), the best model predicting gs included only soil moisture (highest F values) and excluded leaf temperature and PFD (Fig. 4a–d). When these latter two predictors were included, the model was also significant but the coefficient of determination (r2) did not increase substantially. In P. grandiflora and S. arctica the heatwave caused stomatal closure, especially at higher soil-moisture levels, which dissolved the linear relationship found in the unheated treatment (Fig. 4c,d). Note that, in P. viviparum, the treatments were combined as the interaction of treatment with soil moisture was absent (see above). In contrast to gs, the stepwise regression of Fv/Fm always excluded soil moisture. Leaf temperature, on the other hand, was included for P. grandiflora, P. viviparum and S. arctica in the heated plots (no leaf temperature data for C. bigelowii, see Materials and methods), whereas in the unheated plots the linear relationship was never significant (Fig. 5).
These results confirm that, in all species, gs was influenced mainly by soil moisture, hence the observed decrease in gs resulted mainly from desiccation, the indirect effect of the heatwave. PSII maximum efficiency, on the other hand, was influenced by the instantaneous heat. The sensitivity to the long-term components (iii) and (iv) of the heatwave appears to be species-specific, as a treatment effect was found only in C. bigelowii.
An experimentally imposed summer heatwave significantly enhanced mortality of several tundra species, which indicates that extreme temperatures potentially induce significant change in arctic communities. Coinciding physiological signals were decreased stomatal conductance and PSII maximum efficiency, although not in all four target species: P. viviparum was never stressed; S. arctica exhibited stress during the exposure; and the other two species showed a delayed response. This confirms that plant species in tundra communities respond individually to temperature extremes, similarly to responses to moderate warming (Henry & Molau 1997; Dormann & Woodin 2002; Callaghan et al. 2004). Drought, as an indirect effect of the heat treatment, played an important role in the observed impact. Rising temperatures can modify soil moisture content through different mechanisms (Oechel & Billings 1992), but the primary effect is increased evapotranspiration, drying the top soil. A thicker active layer, due to melting of the permafrost, ameliorates drainage and further desiccates the upper soil layers. It is therefore possible that temperature extremes will ultimately modify community composition by altering the competitive interactions between drought-sensitive and drought-tolerant species (Bassow et al. 1994).
The degree to which plant species can tolerate or take advantage of changing climate conditions depends on characteristics such as growth form, phenology, and allocation and storage patterns of carbon and nutrients (Shaver & Kummerow 1992). Based on previous warming experiments with smaller temperature increments during entire seasons or years, positive responses to temperature increase (taking advantage of it) appear to follow the general pattern of forbs > graminoids > deciduous shrubs (Henry & Molau 1997). This is in agreement with our finding that P. viviparum (forb) tolerated the heatwave better than C. bigelowii (graminoid, sedge), while S. arctica (deciduous shrub) was most sensitive. The hydraulic architecture of deciduous species such as S. arctica, which possesses large xylem vessels, allows rapid and efficient uptake of water and nutrients and high rates of photosynthesis and growth during favourable conditions (Semikhatova, Gerasimenko & Ivanova 1992; Gorsuch, Oberbauer & Fisher 2001), but the same hydraulic features can be a disadvantage during drought or frost. For example, Tyree, Davis & Cochard (1994) have argued that wide xylem vessels have a cost in terms of high vulnerability to frost-induced cavitation. The relationship with drought-induced cavitation is weaker, but still apparent (Tyree et al. 1994). As drought was aggravated by the heatwave, obstruction of water flow may have been the basis of the high mortality in S. arctica. Vulnerability to drought-induced cavitation also depends on pit-membrane properties (Sperry & Tyree 1990; Sperry & Saliendra 1994), so anatomical studies may help predict the drought tolerance of arctic species in future studies. The fact that plant responses in the current study were much more species-specific than in the Marchand et al. (2005) experiment, where drought was absent, points at the possible importance of this factor for interspecific differences to become expressed. The Marchand et al. (2006) study, on the other hand, which had two consecutive heatwaves but no appreciable change in soil water status, suggests that species responses may also diverge under prolonged exposure to heat.
Similarly to the interspecific differences in response, the sequence of responses can also give indications about the mechanisms underlying the impact of extreme events. It is well known that stomatal limitation is a primary event under mild stress, followed later by changes in the photosynthetic reactions (Cornic & Briantais 1991; Flexas et al. 2002). This was observed in P. grandiflora, in which gs decreased during the heatwave but Fv/Fm only in the recovery period, possibly reflecting the cumulative stress of the imposed heat and/or drought (Larcher, Wagner & Lutz 1997). Loss of cold acclimation during the heatwave would be an alternative or additional explanation for this delay, if plants that had acclimated to the imposed warmer conditions experienced the return to (low) ambient temperatures as stressful (Marchand et al. 2005). In the leaves of P. grandiflora and C. bigelowii, the difference in gs between treatments disappeared by the end of the experiment, indicating probable recovery in these species. This would be in accordance with the fact that mild heat stress inhibits net CO2 uptake, mainly by reversible conformational changes in the thylakoids (Weis 1981). On the other hand, PSII maximum efficiency in these species was still smaller in the heated plots by the end of the recovery period, so the improvement was either partial or not yet completed. In any case, our findings suggest that the physiological impact of temperature extremes is significantly less ephemeral than the microclimatic disruption itself.
Which components of the heatwave were responsible for its impact? Similarly to species of other biomes, stomatal conductance in tundra plants is strongly correlated with soil moisture (Dawson & Bliss 1989). In our experiment, soil moisture content decreased from the start and remained low during the heatwave until precipitation refilled the soil profile towards the end of the recovery period. This corresponds well with the course of gs. Because the multiple regressions also indicated instantaneous soil moisture as the most important driver of gs in all four species, we conclude that the effects of the heatwave on stomatal behaviour were mainly indirect through increased desiccation. However, in P. grandiflora and S. arctica, gs in the heated plots did not respond to the restored soil moisture at the end of the experiment. This tilted the regression lines in Fig. 4(c,d), suggesting that other factors controlled gs. Possibly a change in senescence in these species was responsible for this, as senescent leaves can display significant decreases in stomatal conductance associated with loss of CO2 assimilatory capacity (Lou & Zhang 1998). Because leaf temperature was picked up as a second (but weak) predictor in the multiple regressions, the instantaneous temperature component of the heatwave apparently also contributed to controlling stomatal conductance. In arctic sedges (Eriophorum vaginatum and C. bigelowii) and S. arctica, Dawson & Bliss (1989) and Starr, Neuman & Oberbauer (2004), respectively, observed that the sensitivity of gs can even extend to temperature fluctuations in the soil. We conclude that the disturbance exerted by the heatwave on the leaf-level CO2 and H2O exchange of our arctic tundra species (by affecting the stomata) was driven mainly by soil moisture, but can still involve other factors depending on species. In the heatwave experiment of Marchand et al. (2005) in July 2001, which had three of the four species in common with the current study but was not characterized by drought, negative responses arose much later. In fact, in 2001 the plants even grew faster during the heatwave, and exhibited stress symptoms only in the aftermath. Although other differences might also be responsible for this – the 2001 experiment being shorter (8 days instead of 13) and at a different location (Zackenberg, High Arctic) – we put forward the absence of drought as the most likely explanation, given the strong effects of water shortage identified in the current study. Positive responses were likewise found in 2004 during the first of two consecutive heatwaves in Marchand et al.'s (2006) study, where the heating did not induce appreciable additional drought either. This would confirm the pattern of water shortage precluding an initial positive response to extreme heat.
When the influence of the heatwave on PSII maximum efficiency is considered, an instantaneous effect of desiccation (component ii) is unlikely, as all the multiple regressions rejected soil moisture as a predictor of Fv/Fm. This is in agreement with observations that drastic water deficits are needed to reduce Fv/Fm (Havaux 1992; Yordanov et al. 2003). Reductions in Fv/Fm can be attributed to injury to the PSII complexes, specifically a decrease in the number of functional PSII centres (Rohacek & Bartak 1999; Hendrickson et al. 2003), and our observation of decreased senescence after the heatwave (10·8% mortality in the heated plots compared with 1·8% in the unheated plots) clearly indicates that the plants were damaged. Generally, reductions in PSII efficiency are found to be species-specific and depend on interactions between temperature and incident PFD (Gamon & Pearcy 1990; Huxman et al. 1998). In accordance with this, the second predictor – instantaneous leaf temperature – was significant, but was selected only in the heated plots of P. viviparum, P. grandiflora and S. arctica, where it was not of major importance. Instantaneous effects of heat on PSII maximum efficiency were therefore limited, in agreement with findings on Arctic and Antarctic vascular plants (Xiong, Ruhland & Day 1999, Starr et al. 2004). Nevertheless, as the photosynthetic pathway is very sensitive to heat (Havaux et al. 1988), direct effects of the heatwave may have affected other parts of the photosynthetic apparatus (e.g. inactivation of Calvin-cycle enzymes such as Rubisco: Law & Crafts-Brandner 1999; Salvucci & Crafts-Brandner 2004), contributing to the observed enhanced senescence in the heated plots.
The occurrence in our experiment of substantial changes in soil moisture induced by heating was unplanned, and probably arose from the warmer conditions compared with Marchand et al. (2005, 2006). To what extent would a fully factorial design with controlled drought (present/absent) crossed with the heatwave (present/absent) be more suitable to disentangle both factors? At first sight, such a setup would make it possible to separate statistically the impacts of heat and drought and their possible interaction. However, in practice it is not straighforward to simulate water shortage in the field in a controlled way, for example, the use of rain shelters would constitute a partial enclosure of the plots, whereas the free-air technique was designed to avoid precisely this. Second, which drought level should be applied in the ambient treatment? As the additional drought was induced by the heating, its level cannot be calculated beforehand. A more feasible option for future work (but one that would require twice the number of heaters) would be to eliminate the additional drought through irrigation in half the plots of the heated treatment. However, this would also modify the temperature increment, because the reduced evapotranspiration of drier heated plots reinforces the warming (not shown). A well watered, heated treatment (or better, a heated treatment tracking the soil moisture in the ambient plots) would therefore experience a less intense heatwave. Controlled separation of the influences of heat and drought in a natural system in the absence of enclosure therefore seems difficult to realize. On the other hand, the technique used here (separation of factors heat and drought as multiple covariates in ancova), was justified as their correlation was weak. It is therefore improbable that shared variance that cannot not be attributed to either factor has corrupted the analysis.
We conclude that a simulated extreme event (+7·6 °C for 13 days) induced changes in Arctic tundra plants, which can probably be ascribed to a combination of heat and drought causing dysfunctions inter alia in the photosynthesis apparatus, ultimately leading to enhanced senescence and leaf mortality. Soil moisture apparently dominated the observed changes over the effects of heat. As extreme temperature events often coincide with desiccation (Chaves 1991), and as the responses we observed were highly species-specific, more frequent temperature extremes in a future climate might alter tundra communities. Studies based on increases in mean temperature alone may overlook these impacts.
We thank the University of Copenhagen for providing access to, and logistics at, the Arctic Station located on Disko Island, Greenland. This study was supported by the Fund for Scientific Research–Flanders (FWO, Belgium) under contract G.0357·02.
- 1994) The response of temperate tree seedlings grown in elevated CO2 to extreme temperature events. Ecological Applications 4, 593–603. , & (
- 1987) Photon yield of O2 evolution and chlorophyll fluorescence characteristics at 77K among vascular plants of diverse origins. Planta 170, 489–504. & (
- 2005) Soil moisture and predicted spells of extreme temperatures in Britain. Journal of Geophysical Research – Atmospheres 110, D05104. , , & (
- 1997) Impacts of climate change on demographic processes and population dynamics in Arctic plants. Global Change and Arctic Terrestrial Ecosystems (eds W.C.Oechel, T.V.Callaghan, T.Gilmanov et al.), pp. 129–152. Springer Verlag, New York. & (
- 2004) Responses to projected changes in climate and UV-B at the species level. Ambio 7, 418–435. , , et al. (
- 1991) Effects of water deficit on carbon assimilation. Journal of Experimental Botany 42, 1–16. (
- 1991) Partitioning of photosynthetic electron flow between CO2 and O2 reduction in a C3 leaf (Phaseolus vulgaris L.) at different CO2 concentrations and during drought stress. Planta 183, 178–184. & (
- 1989) Intraspecific variation in the water relations of Salix arctica, an arctic–alpine dwarf willow. Oecologia 79, 322–331. & (
- 2002) Climate change in the arctic: using plant functional types in a meta-analysis of field experiments. Functional Ecology 16, 4–17. & (
- 2000a) Observed variability and trends in extreme climate events: a brief review. Bulletin of the American Meteorological Society 81, 417–425. , , , , & (
- 2000b) Climate extremes: observations, modelling, and impacts. Science 289, 2068–2074. , , , , & (
- 2000) Discovering Statistics using SPSS for Windows: Advanced Techniques for the Beginner. Sage, London. (
- 2002) Steady-state chlorophyll fluorescence (Fs) measurements as a tool to follow variations of net CO2 assimilation and stomatal conductance during water-stress in C3 plants. Physiologia Plantarum 114, 231–240. , , et al. (
- 1990) Photo-inhibition in Vitis california– interactive effects of sunlight, temperature and water status. Plant, Cell & Environment 13, 267–275. & (
- 1984) Heat resistance and energy budget in different Scandinavian plants. Holarctic Ecology 7, 1–78. (
- 2001) Primer of Applied Regression and Analysis of Variance. McGraw-Hill, New York. & (
- 2001) Comparative vessel anatomy of arctic deciduous and evergreen dicots. American Journal of Botany 88, 1643–1649. , & (
- 1999) Indicators of climate change for the Russian Federation. Climatic Change 42, 219–242. , , & (
- 1992) Stress tolerance of photosystem II in vivo. Plant Physiology 100, 424–432. (
- 1988) Correlation between heat tolerance and drought tolerance in cereals demonstrated by rapid chlorophyll fluorescence tests. Journal of Plant Physiology 133, 555–560. , & (
- 2003) Assessment of photoprotection mechanisms of grapevines at low temperature. Functional Plant Biology 30, 631–642. , , , & (
- 1997) Tundra plants and climate change: The International Tundra Experiment (ITEX). Global Change Biology 3, 1–9. & (
- 2001) Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge, UK. , , , , & (
- 1998) Gas exchange and chlorophyll fluorescence responses of three south-western Yucca species to elevated CO2 and high temperature. Plant, Cell & Environment 21, 1275–1283. , , & (
- 1997) The effect of heat on photosynthesis, dark respiration and cellular ultrastructure of the arctic–alpine psychrophyte Ranunculus glacialis. Photosynthetica 34, 219–232. , & (
- 1999) Inhibition and acclimation of photosynthesis to heat stress is closely correlated with activation of ribulose-1,5-bisphosphate carboxylase/oxygenase. Plant Physiology 120, 173–181. & (
- 1998) Modifications in photosystem II photochemistry in senescent leaves of maize plants. Journal of Experimental Botany 49, 1671–1679. & (
- 1995) Testing predictions of the resistance and resilience of vegetation subjected to extreme events. Functional Ecology 9, 640–649. , , et al. (
- Nijs, I. (2005) Performance of High Arctic tundra plants improved during but deteriorated after exposure to a simulated extreme temperature event. Global Change Biology 11, 2078–2089. , , , &
- 2006) Are heat and cold resistance of arctic species affected by successive extreme temperature events? New Phytologist 170, 291–300. , , , & (
- 1992) Arctic climate: potential for change under global warming. Arctic Ecosystems in a Changing Climate – An Ecophysiological Perspective (eds F.S.Chapin, R.L.Jefferies, J.F.Reynolds, G.R.Shaver & J.Svoboda), pp. 11–34. Academic Press, San Diego, CA, USA. (
- 2000) Chlorophyll fluorescence – a practical guide. Journal of Experimental Botany 51, 659–668. & (
- 2001) Meteorological observations in 2000 at the Arctic Station, Qeqertarsuaq (69°15′ N), Central West Greenland. Danish Journal of Geography 101, 155–158. , & (
- 1996) Free Air Temperature Increase (FATI): a new tool to study global warming effects on plants in the field. Plant, Cell & Environment 19, 495–502. , , , , & (
- 2000) Climate-warming simulation in tundra: enhanced precision and repeatability with an improved infrared-heating device. Arctic, Antarctic and Alpine Research 32, 242–253. , , , , & (
- 1996) The Physiology of Plants under Stress. John Wiley & Sons, New York. & (
- 1992) Effects of global change on the carbon balance of Arctic plants and ecosystems. Arctic Ecosystems in a Changing Climate – An Ecophysiological Perspective (eds F.S.Chapin, R.L.Jefferies, J.F.Reynolds, G.R.Shaver & J.Svoboda), pp. 139–168. Academic Press, San Diego, CA, USA. & (
- 1999) Changes in climate extremes over the Australian region and New Zealand during the twentieth century. Climatic Change 42, 183–202. , , et al. (
- 2002) Changes in seasonal and annual high-frequency air temperature variability in the arctic from 1951 to 1990. International Journal of Climatology 22, 1017–1032. (
- 1999) Technique of the modulated chlorophyll fluorescence: basic concepts, useful parameters, and some applications. Photosynthetica 37, 339–363. & (
- 2004) Inhibition of photosynthesis by heat stress: the activation state of Rubisco as a limiting factor in photosynthesis. Physiologia Plantarum 120, 179–186. & (
- 1994) Chlorophyll fluorescence as a non-destructive indicator for rapid assessment of in vivo photosynthesis. Ecological Studies 100, 49–70. , & (
- 2001) Photosynthetic response of Amphibolis antarctica and Posidonia australis to temperature and desiccation using chlorophyll fluorescence. Marine Ecology Progress Series 220, 119–130. & (
- 1992) Photosynthesis, respiration, and growth of plants in the Soviet Arctic. Arctic Ecosystems in a Changing Climate – An Ecophysiological Perspective (eds F.S.Chapin, R.L.Jefferies, J.F.Reynolds, G.R.Shaver & J.Svoboda), pp. 169–192. Academic Press, San Diego, CA, USA. , & (
- 1992) Phenology, resource allocation and growth of Arctic vascular plants. Arctic Ecosystems in a Changing Climate – An Ecophysiological Perspective (eds F.S.Chapin, R.L.Jefferies, J.F.Reynolds, G.R.Shaver & J.Svoboda), pp. 193–212. Academic Press, San Diego, CA, USA. & (
- 1994) Intra- and inter-plant variation in xylem cavitation in Betula occidentalis. Plant, Cell & Environment 17, 1233–1241. & (
- 1990) Water stress induced xylem cavitation in three species of conifers. Plant, Cell & Environment 13, 427–437. & (
- 2004) Ecophysiological analysis of two arctic sedges under reduced root temperatures. Physiologia Plantarum 120, 458–464. , & (
- 1994) Biophysical perspectives of xylem evolution – is there a trade-off of hydraulic efficiency for vulnerability to dysfunction? IAWA Journal 15, 335–360. , & (
- 2001) Survival, gap formation, and recovery dynamics in grassland ecosystems exposed to heat extremes: the role of species richness. Ecosystems 4, 797–806. , , , & (
- 1981) Reversible heat-activation of the Calvin-cycle: a possible mechanism of the temperature regulation of photosynthesis. Planta 151, 33–39. (
- 2005) Changing frequency of occurrence of extreme seasonal temperatures under global warming. Geophysical Research Letters 32, L20721. & (
- 2000) Sensitivity of three grassland communities to simulated extreme temperature and rainfall events. Global Change Biology 6, 671–684. , , & (
- 2001) Impacts of extreme climatic events on competition during grassland invasions. Global Change Biology 7, 1–13. , , & (
- 1985) Impact of extreme events. Nature 316, 106–107. (
- 1999) Photosynthetic temperature response of the Antarctic vascular plants Colobanthus quitensis and Deschampsia antarctica. Physiologia Plantarum 106, 276–286. , & (
- 2003) Plant responses to drought and stress tolerance. Bulgarian Journal of Plant Physiology 2003, S187–S206. , & (