| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ecology |
Botanical Institute, Göteborg University, P.O. Box 461, SE 405 30 Göteborg, Sweden
Received for publication April 16, 2004. Accepted for publication November 17, 2004.
ABSTRACT
Long-term studies on phenology are rarely reported from arctic and alpine areas, but are essential for understanding biotic and abiotic controls on flowering. We monitored first flowering day (FFD) for 144 species in a subarctic–alpine area in Swedish Lapland over a period of 10 yr (1992–2001). Temperature and global radiation were monitored continuously, and snowcover duration was observed. Thawing degree-days and snowcover duration (exposure) were the dominant environmental controls on phenology. We introduce a lability index (LI) to describe the interannual variability in FFD among species. The temporal sequence of species is very constant among years, although a few species are more labile. The species were also classified into the catagories "Functional type," "Raunkiær's life form," and "Sørensen's wintering floral type." The last two reflected the environmental data best, and together with "Exposure" they were combined into a phenology index (PI). The index was subsequently used in a triangular ordination together with FFD. The ordination illustrates whether species flower earlier or later than expected from their preconditions. We hypothesize that species having a delayed flowering respond more readily to global warming than species having an already optimized flowering.
Key Words: global change • lability index • life forms • long-term monitoring • phenology index
Phenology is the study of the timing of recurring biological events, the causes of their timing with regard to biotic and abiotic forces, and the interrelation among phases of the same or different species (Leith, 1974
). In arctic and alpine tundra, the growing season is extremely short, and its duration varies strongly among years (Molau, 1993
; Thórhallsdóttir, 1998
). In this biome, the timing of the onset of flowering is crucial to the reproductive success of flowering plants. In late-flowering species, the entire seed production is often lost in summers colder or shorter than the average (Molau, 1993
, and references therein; Henry and Molau, 1997
; Arft et al., 1999
). The study was initiated because it was noted that the order of species coming into flowering in the area was constant between years (1992 and 1993), although the actual timing of the individual species could be either retarded or enhanced (Molau, 1997
). We wanted to know if the trend could be confirmed over a longer period of time, and if so, what the environmental controls are.
Long-term studies of flowering phenology of local tundra floras are sparse in the literature, perhaps due to the logistic difficulties maintaining such studies over several years. The only existing study from the tundra biome of which we are aware, and which is based on 10 or more years, is the 11-yr monitoring of flowering phenology in central Iceland (subarctic–alpine) by Thórhallsdóttir (1998)
. However, the phenological stage of the flora at the site was assessed only once per year, in early July, and the proportion of species in flower at that time was correlated with climatic variables and snow conditions. Comparable long-term studies from neighboring biomes include the studies by Fitter et al. (1995)
, as well as Fitter and Fitter (2002)
, from England and Diekmann (1996)
from a deciduous woodland in southern Sweden. For a review on ecological responses to recent climate change, see Walther et al. (2002)
.
In the majority of phenological studies published, the timing variable reported is peak flowering day. Because it takes a great deal of effort to monitor peak flowering day for all species across a landscape, we have chosen to restrict our monitoring to a single date, the first flowering day (FFD). This is also the variable used by Fitter and Fitter (2002)
. In order to justify this choice, we investigated the correlation between FFD and peak flowering in three species representing different life forms typical for the biome.
To examine possible environmental controls on flowering phenology, we selected three abiotic factors: (1) temperature, (2) radiation, and (3) snowcover, or, seen from the opposite perspective, the degree of exposure, along with three biotic factors: (1) Raunkiær's life form (Raunkiær, 1934
), (2) Sørensen's wintering floral types (Sørensen, 1941
), and (3) functional types (Chapin et al., 1996
) for analysis.
Snow distribution and the temporal dynamics of the snowmelt process, governed by temperature and global radiation, are the major controls on flowering phenology in tundra plants (Sørensen, 1941
; Billings and Bliss, 1959
; Molau, 1993
, 1997
, 2001
; Thórhallsdóttir, 1998
). The distribution of the winter snowpack in an alpine landscape is far from uniform because of the interaction with topography and prevailing winds. Because snow distribution is fairly constant among years, snow duration is a main factor differentiating tundra plant communities (Billings and Bliss, 1959
; Molau et al., 2003
).
The relatively constant among-year pattern of snowcover within a tundra landscape makes the stature of the plants an important factor controlling flowering phenology. The classical concept of life forms launched by Danish botanist Raunkiær (1934)
is highly applicable in the tundra (Sørensen, 1941
). The position of leaf and flower buds in relation to the soil surface and snowcover has implications for how readily plants respond to summer microclimate. Besides Sørensen's (1941)
study in northeast Greenland, few efforts have been made to utilize Raunkiær's life form concept in tundra plant communities. Nevertheless, it seems to be highly relevant in studies on phenological differentiation in arctic and alpine environments.
While studying the phenology of flowering plants in northeast Greenland, Sørensen (1941)
classified the wintering stages of the flower buds or primordia of 181 species. He defined seven classes of wintering floral types, ranging from (I) totally undifferentiated (as in annuals) to the completely differentiated stage (VII), where mature pollen grains are in place before the onset of winter. In addition, he lumped a number of species in a class called "aperiodic," in which species observed to enter dormancy at any stage between III and VII and which seemed to be highly opportunistic in their strategies, were accommodated.
A third model for classification of tundra plants was proposed by Chapin et al. (1996)
. Based on growth form, species are grouped into "Functional types," such as shrubs, herbs, and graminoids. The potential drawback of this classification is that a functional type, such as the shrubs, may be heterogeneous in its composition regarding responses to climate change. In studies of plant responses to observed and experimental warming in subarctic ecosystems in northern Swedish Lapland, Graglia et al. (1997)
and Molau (2001)
have shown that species of a mainly boreal origin, living at their latitudinal and/or altitudinal upper limits in the Subarctic, respond much more vigorously to increased temperature than do species that can be regarded as "arctic specialists," regardless of functional type.
Fitter and Fitter (2002)
, with their 47-yr data set from southern England, have shown a common trend that the currently ongoing climatic warming has induced earlier first flowering date (FFD) in angiosperms. Climate change will, according to the forecasts by IPCC (2001)
, be strongest in polar regions. Therefore, it is particularly important to focus on the phenological responses of arctic and alpine plant species, in order to provide a forecast for the impacts of an accelerating climate change in northern regions.
In the present study, we analyzed the among-year difference in "onset of flowering" sequence of 144 angiosperm species during a 10-yr period (1992–2001) in the catchment area of Lake Latnjajaure in the mountains of northern Swedish Lapland. The Latnjajaure Field Station was continuously manned throughout the growing season all years, which allowed us to carry out the project with day-by-day monitoring.
In this paper, we address the following questions: (1) How constant is the species' temporal sequence of entering the stage of flowering among years over a longer period? (2) What are the environmental controls of the onset of flowering of individual species? (3) What ecological grouping of species explains most of that variation? and (4) What are the implications for impacts of Global Change?
MATERIALS AND METHODS
Field site and plant communities
The field work was conducted in northernmost Sweden at the Latnjajaure Field Station (LFS) in the valley of Latnjavagge, 68°21' N, 18°29' E, at around 1000 m elevation. The valley is covered by snow for most of the year, and the climate is classified as subarctic–alpine (Polunin, 1951
; Alatalo and Molau, 1995
) with cool summers and relatively mild, snowrich winters. The mean annual temperature varies between –1.5 and –2.9°C, and the annual minimum ranges from –27.3 to –21.7°C (data from 1992–2001). July is the warmest month, with a mean temperature ranging from +5.4°C (1992) to +9.9°C (1997). The total annual precipitation ranges from 605 mm (1996) to 990 mm (1993).
The vegetation in the Latnjajaure valley is representative of the Low Arctic, with Cassiope tetragona, Dryas octopetala, and Carex bigelowii among the dominant species (Polunin, 1951
; Molau and Alatalo, 1998
; Molau, 2001
). The bedrock varies from acidic to base-rich, and the soils of the plant communities vary from dry to wet and from siliceous and acidic (pH = 4) to calcareous and circumneutral (pH = 6; Björk, 2000
). The valley exhibits patchy permafrost, and there are extensive areas of patterned ground as well as other frost-induced phenomena (Kling, 1997
). This environmental heterogeneity within the landscape results in a wide range of plant communities, and the study area is among the most species-rich valleys in the northern Scandes.
Data collecting
The LFS was continually manned, with at least one field assistant, throughout the growing seasons (approximately 25 May to approximately 3 September) of 1992–2001, and at least one of the authors was always in place during the onset of flowering. Using a standardized preprinted protocol, we documented the date (as Julian day number) when flowers of all angiosperm species in the valley first opened. This day number is denoted "first flowering day," acronymized FFD. Most of the records were taken in the absolute vicinity of LFS (i.e., within a 500-m radius from the cabins), although a few species were monitored at the lower outskirts of the watershed area (i.e., Carex rariflora, Hierochloë alpina, Scirpus cespitosus). For most of the species, the same individuals or the same stand of plants was monitored throughout the 10-yr period.
The total of vascular plant species recorded in the valley amounts to 176. Here, we concentrate on the 144 species for which we have data for 6 yr or more, i.e., generally the species we have known to be present from the start of the survey (Appendix 1; see Supplemental Data accompanying the online version of this article). Plant nomenclature follows Nilsson (1986)
with two exceptions: we acknowledge Cerastium glabratum at the species level and use the name Silene wahlbergella instead of S. uralensis (Rupr.) Bocq.
Climate data were obtained from the continuously operating weather station at LFS, the core of which is a Delta T logger (Delta-T Devices Ltd., Cambridge, UK) collecting data around the year. Data were stored as hourly means, maxima, and minima. During the growing seasons, the automatic weather station was supplemented with a manual weather station using a Stevenson screen with standard equipment and a precipitation gauge. For the design of ITEX climate stations, see Molau (1996)
. In the present study, we focus on two climatic variables: temperature and radiation. Temperature is used in the form of thawing degree days (TDD), the cumulative integrated temperature sum above 0°C (see Molau, 1996
). The daily contribution to TDD is the mean over 24 h of all hourly means above freezing. Radiation is measured as the accumulated global radiation (AGR) influx, based on daily totals.
The dates when 10, 50, and 90% of the valley is snowfree, used in Fig. 5, are based on snow depth measurements at the nodes of a grid net applied to the central valley. Snow depth was monitored every fifth day from approximately Julian date 145 during the years 1999–2001. The data was entered into a GIS model and the average snow depth was estimated by interpolation. The calculations were kindly provided by K. Lindblad, Göteborg University.
|
Classification of the species
We have used several keys to subdivide the material. Plant protection by snowcover was divided into three classes: exposed (ridges and cliffs), protected (heaths and meadows), and snowbeds (depressions in the landscape). The species were separated with regard to life forms according to Raunkiær (1934)
into phanerophytes, chamaephytes, hemicryptophytes, cryptophytes, and therophytes (annuals and biennials). Phanerophytes bear their leaf and flower buds well above the soil surface; the chamaephytes (dwarfshrubs and cushion plants) have their meristems located slightly above ground. The hemicryptophytes, many of them rosette plants, have their primordia right at the soil surface. The cryptophytes have meristems below ground, and the therophytes overwinter as seeds or seedlings. The classification of plants into hemicryptophytes and cryptophytes was ascertained by studies of herbarium specimens in the collections of herbarium (GB). Raunkiær (1934)
pointed out that species with wide distributions may attain several life form classes within their range. The classification of the species in the present study refers to their life forms at Latnjajaure and nearby areas.
Species were likewise divided into functional types (FTs), roughly as defined in the Toolik Lake project in northern Alaska (Chapin et al., 1996
). At Latnjajaure, we identified the following FTs: evergreen shrubs, deciduous shrubs, cushion plants, herbs, and graminoids. This classification has been used previously at Latnjajaure in studies of seed flux (Molau and Larsson, 2000
; Larsson and Molau, 2001
).
With regard to the plants' wintering floral stages, we follow the classification of Sørensen (1941)
. Eighty of the 144 vascular plant species in our study in northern Sweden were investigated and classified by Sørensen in his material from northeast Greenland collected in 1931–1937.
In addition, the flowering plants of Latnjajaure were grouped according to a number of criteria: their general geographic distribution (boreal, arctic, and alpine species), based on the distribution maps of Hultén (1958
, 1962
, 1970
), their main biome (forest or tundra), and their vegetative reproductive strategy (phalanx and guerilla species) (Lovett Doust, 1981
). In this case, classification was refined by using herbarium material at herbarium GB. Other grouping factors include lignification (woody vs. herbaceous), upper distributional limit in the Latnjajaure catchment, and pollination mode (entomophilous [insect pollinated] vs. anemophilous [wind-pollinated]).
Data analysis
Because we intended to focus on the variability of species' phenologies over the years, we used not only means of the variables (FFD, TDD, and AGR), but also their variability. As the most adequate parameter for variability, we have chosen to use the standard deviation (SD). Whence exponential, the variance tends to increasingly exaggerate deviations as the summer progresses, and hence it was disregarded. The coefficient of variance (CV) was avoided because we are dealing with data belonging to one sole population (Sokal and Rohlf, 1981
). The statistical data analyses were performed using the StatView, version 5.0.1 (SAS Institute Inc., Cary, North Carolina, USA) and SuperANOVA software, version 1.11 for Macintosh (Abacus Concepts Inc., Berkeley, CA).
All species were given rank numbers for their position in the flowering sequence for each year, and a mean rank for each species was calculated. Based on this data set, a "lability index" (LI) for each species was calculated according to the formula
In order to visualize the impact of what we consider to be the three main factors shaping phenology, a phenology index was calculated and plotted against FFD in a triangular ordination. The phenology index (PI) was calculated from the formula PI = a + 2b + 3c where a is Sørensen's wintering floral type (0, type VII, anthers fully developed before onset of winter; 1, type VI; 2, type VIII, aperiodic; 3, type V; 4, type IV; 5, type III; 6, type II; 7, type I, anthers undeveloped before onset of winter), b is Raunkiær's life form (0, phanerophyte; 1, chamaephyte; 2, hemicryptophyte; 3, cryptophyte; 4, therophyte), and c is snowcover (0, exposed; 1, protected; 2, snowbed). In the formula, a takes a total of seven values, b five values, and c only three. To avoid bias towards one of the classes, the values are weighted accordingly. The ordination gives an overview of how well species use their ability to flower early under given snow conditions.
RESULTS
The three species monitored for peak flowering in permanent plots over 10 yr all showed highly significant positive correlations between mean FFD and mean date of peak flowering (Diapensia lapponica, r2 = 0.662, P = 0.0026; Dryas octopetala, r2 = 0.920, P < 0.0001; Ranunculus nivalis, r2 = 0.713, P = 0.0013). Thus, we feel confident that FFD is a useful phenological tool.
During our 10-yr study in the mountains of northern Swedish Lapland, we could verify that the sequence in which species came into flower remained relatively constant year after year, even though summer climates differed widely (Table 1; Appendix 1, see Supplemental Data accompanying the online version of this article). The lability index (LI) was introduced in order to describe the disorder in the sequence of onset of flowering among species and years. A lability index value of 0 indicates that the species remains at the same rank in the flowering sequence year after year. In contrast, an index value over 10 indicates that a species changes its phenology from early to late or vice versa each year. All species at Latnjajaure have LIs in the range 0–1 indicating that they are rather constant and appear in a certain order each year. A bivariate scattergram with sequence lability plotted as a function of mean FFD is presented in Fig. 1, and it shows that lability is independent of the time of onset of flowering. A few species deviate by being more labile. Those having a LI above 0.5 are marked in the figure. There are two labile early-flowering species: Ranunculus nivalis and Eriophorum vaginatum. The remaining are late flowering and some are confined to snowbeds: Poa glauca, Veronica fruticans, Saxifraga foliolosa, Deschampsia flexuosa, Deschampsia alpina, Phippsia algida, and Gentianella tenella. Snow protection has a significant impact on the LI of a species (one-way ANOVA, square-root transformed data, F2,141 = 5.386, P = 0.0056), and it is the species growing in snowbeds that have significantly higher LIs than those that are protected or exposed (Fisher's protected LSD post-hoc test, P < 0.05 and P < 0.01, respectively). Saxifraga oppositifolia is outstanding in having a LI of zero, a result of constantly and without competition being the first species to come into flower in the study area.
|
|
There are large differences in FFD, TDD, and AGR means between snowbed species, snowprotected species, and those that are growing in exposed, wind-swept situations on cliffs and ridges (one-way ANOVAs; FFD nontransformed means, F2,141 = 15.24, P < 0.0001; TDD square root-transformed means, F2,141 = 18.79, P < 0.0001; AGR square root-transformed means, F 2,141 = 15.77, P < 0.0001; all pairwise comparisons were significant at the 5% level in Fisher's protected LSD post-hoc tests). The inhabitants of exposed places are also always less variable among years regarding FFD (see under LI above), as well as TDD and AGR requirements.
The plant species first appearing above the snow are generally also the first to flower (Fig. 2). Phanerophytes are on average the first to open their flowers, hemicryptophytes and cryptophytes follow about 3 wk later. The therophytes flower in late season. Annuals develop from seed every year and a photosynthesizing shoot must precede flowering. The early phanerophytes have less variable FFDs than the late-flowering therophytes. The later the flowering, the larger the variation.
|
Phanerophytes and chamaephytes start flowering after a certain amount of TDD has accumulated (Fig. 2). The year-to-year variability (in terms of SD) is not very large. Hemicryptophytes and cryptophytes show a larger variation in their TDD requirement. Therophytes have the largest variation in TDD requirements among years. There seems to be no relationship between life form and variation in AGR (Fig. 2).
The flowering sequence among FTs is shown in Fig. 3. The first three FTs to come into flower (evergreen shrubs, deciduous shrubs, and cushion plants) all include partially or fully lignified species, the latter two (graminoids and herbs) include nonlignified plants. The difference in onset of flowering between lignified and herbaceous species is highly significant (one-way ANOVA, nontransformed data, F1,142 = 25.459, P < 0.0001). The cushions and shrubs are less variable among years in FFD than are graminoids and herbs. Evergreen shrubs, deciduous shrubs, and cushions (which are often semi-evergreen) require the least amount of accumulated heat (TDD) and the variation among years is lowest for evergreen shrubs and cushion plants (Fig. 3). Deciduous shrubs have a distinctly higher variability in TDD required before onset of flowering than the other two groups, and yet, they almost always flower on the same date each year. The graminoids and herbs start flowering at very different accumulated temperatures (TDDs) but this is tied to a high variability in FFD.
|
. Our data show that aperiodic species start flowering in the same time intervals as species of types III– VII, which corroborates this observation. Species having the least developed flowers by onset of winter are labile in terms of FFD and TDD, which may vary a lot from year to year, but rather constant in terms of AGR requirement. In contrast, species of type VI, having quite developed flowers by onset of winter, depend on temperature rather than radiation.
|
noted that especially many grasses project their inflorescence quite early but without opening the flowers. They linger on in this stage for some time, but the reason for this phenomenon has not been elucidated. FFD mean and TDD mean are at a similar level across distribution groups. Species with a predominantly arctic distribution do not flower earlier, or quicker, than species with a boreal distribution. However, if only entomophilous plants are analyzed, there is a strong tendency toward a break-up of the boreal plants from the arctic and arctic/alpine species toward later flowering and increased TDD and AGR requirements (one-way ANOVAs; square-root transformed mean TDD, F2,88 = 2.91, P = 0.0598; nontransformed mean AGR, F2,88 = 2.95, P = 0.0578).
Plants having their main distribution area in the Arctic show the least variation in FFD and TDD requirements. The arctic-alpine species have a slightly higher mean variability in these two variables compared with the previous group. Boreal plants vary the most among years. The opposite is true for the variable AGR.
When classified according to their main biomes, tundra plant species did not differ significantly from forest plants with regard to FFD, TDD, and AGR means. The same is true for the FFD, TDD, and AGR SD variables. However, if only the entomophilous species are considered, forest species flower later than tundra species. The among-year variability is still comparable between the groups.
As to their altitudinal distribution, early-flowering species, which do not require a large temperature sum before they start to flower, have a higher upper limit than species flowering late (Fig. 6).
|
|
DISCUSSION
The variability in climate among summers in the North is well documented (e.g., Shaver et al., 1986
; Thórhallsdóttir, 1998
; Molau, 2001
), and the data from 1992–2001 at Latnjajaure is no exception (Table 1). The pilot study by Molau (1997)
on the phenological sequence of 15 selected species during two consecutive summers with contrasting climate, indicated that the flowering sequence would be constant between years, even though the actual timing of individual species could differ largely. In our 10-year material, summer climates have differed much more than in the pilot study, with an unusually cold summer in 1995 and exceptionally warm ones in 1996 and 1997. The greater variability did not affect the phenological sequence of species among years and, thus, corroborates the previous results. The constructed lability index (LI), where phenological rank is corrected for temporal crowding of flowering dates, shows that most of the 144 species in our study are stable rather than labile with regard to phenological sequence. It is also obvious that the most labile species are either early-flowering opportunists or late snowbed plants. The LI applies at the landscape level. It captures the individual species' phenology in relation to all species in the landscape. Because none of the terms are exclusively alpine or arctic, it should be applicable for all biomes, and it could be a useful tool for sensitivity analysis in relation to climate change impacts.
In the subarctic–alpine landscape at Latnjajaure, snow protection of the plants proved to be very important. The plants growing on cliffs and ridges are usually not snowcovered from mid-May, and the same situation applies for meristems of phanerophytes. Snowfall during winter varies from year to year, but the snow distribution pattern, with snowbeds and snowfree ridges, is relatively constant over the years at the landscape level. Hence, species growing in exposed environments flower at approximately the same time every year. In contrast, the detailed pattern of snow distribution varies considerably in sites with medium and high snow protection. More TDDs and/or increased rainfall are required to melt the snow in snowrich years, and this process takes longer the more snow there is. Life processes in the low herbaceous and graminoid plants do not begin until the snow has melted and, hence, both FFD and TDD are correlated with snow conditions. Onset of flowering seems to be unrelated to AGR for all life forms.
The two different life form classification systems tested, viz., the traditional life form spectrum by Raunkiær (1934)
and the system of FTs of plants developed for the tundra by Chapin et al. (1996)
, turned out to be almost equally useful. This result is in agreement with the findings of Thórhallsdóttir (1998)
from central Iceland, another subarctic–alpine landscape, where snowmelt and TDD were found to be the proxy controls. Our results concerning hemicryptophytes and cryptophytes seem to reflect the among-year variation in snowcover thickness. With buds at the ground surface, the snow has to melt almost completely before growth and flowering are initiated. A variable amount of snow among years will lead to variation in the date when a site becomes snowfree and hence also the TDD requirements of the plants at the site.
The phenological sequence found among species in northern Swedish Lapland correlates strongly with the classification of wintering developmental stages of the species in Sørensen's (1941)
study from northeast Greenland. Eighty species were shared between his and our data sets. The less developed the buds are before winter sets in, the later the onset of flowering occurs the next year. Thus, wintering developmental stage of the reproductive organs has direct bearing on the phenology of flowering in the subsequent summer. Sørensen's detailed study of plant wintering stages is unique, and his classification of the species obviously holds true throughout their distribution ranges.
The phenology index applies at the species level and is in our case specially designed for arctic and alpine tundra. It extracts the potential for early flowering by combining bud development and position with snow protection. As such, it indicates how well adapted a particular species is to its environment. This index could work well in other high latitude systems. In lower latitude biomes, other strategy classification systems may be more important, and a modification of the PI presented here can probably be constructed for these situations. Plotting PI against FFD in a triangular graph (Fig. 5) makes it possible to visualize our conclusion that we are dealing with a phenology that is space-delimited by three main strategies: early, late, and delayed flowering. We anticipate that the species plotted along the E–L line and beyond are not as responsive in their flowering phenology to global warming than are the more delayed species. That this is actually true can be demonstrated by extracting data from the literature reporting on phenological responses to artificial temperature increment experiments conducted under ITEX. The strongest response is obtained for Cassiope tetragona (Molau, 2001
), a species found quite far down the delayed axis in the triangle. Other species that significantly antedate their onset of flowering under experimental warming are Silene acaulis (Alatalo and Totland, 1997
), Salix herbacea (Jones et al., 1997
), Carex bigelowii (Stenström and Jónsdóttir, 1997
), Dryas octopetala, and Polygonum viviparum (Molau, 2001
). These are all found somewhat to the right of the E–L line. In contrast, two species are shown not to respond to increased temperatures. These are Saxifraga oppositifolia sitting on the E–L line (Stenström et al., 1997
) and Ranunculus nivalis, which is already overcompensating (Molau, 2001
). In conclusion, we regard the presently delayed species as the ones that will respond most positively to global warming. The species on the E–L line, or to the left of it, will not be able to respond to global warming in terms of flowering phenology and may suffer a competitive disadvantage. There is a general trend that increased temperature induces earlier flowering and, hence, a better seed set with respect to number and/or mass (Henry and Molau, 1997
; Arft et al., 1999
). With the plasticity that we have shown to exist in the flora of the Latnjajaure landscape, we anticipate that an increased climatic warming will bring about, on average, earlier flowering in most species of the local flora, thereby ensuring a better seed set (Molau, 1993
). Another effect may be an increasingly better seed production from boreal invaders, causing a feedback on the rate of community change.
The finding that anemophilous woody species flower significantly earlier than their entomophilous counterparts may be phylogenetically constrained since the anemophilous woody species almost exclusively belong to the plant family Salicaceae and the woody, entomophilous ones to Ericaceae. However, the later onset of flowering in insect-pollinated species may also be interpreted as an adaptation to the pollination syndrome rather than physiological constraints on growth. The abundance of insects is low very early in the season, and they do not move about much. Hence seed set in species depending on insect visits would be low. Thus, the mean FFD is expected to be pushed forward towards later dates.
When analyzing the residuals among years for the date of onset of flowering, some species turned out to be most constant in terms of TDD, other in terms of AGR, and some just went by the clock, showing little variability in FFD. Species with a mainly boreal distribution seemed to be more triggered by temperature than did pure tundra species, a finding that correlates well with the studies by Graglia et al. (1997)
and Molau (2001)
. This could be explained by the fact that global radiation is at a maximum early in the season and decreases as the summer proceeds. Early in the season, the daily global radiation, when it is high, varies greatly between sunny and cloudy days. Because the AGR is smaller early in the season, the relative effect on the accumulated sum is larger early in the season than late in the season. Thus, plants flowering early experience relatively more variation in AGR than later-flowering species. In contrast, Fitter and Fitter (2002)
found that species far away from their distribution center were less likely to respond to a rise in temperature. Their results, however, are based on species from the temperate biome in the British Isles and may not apply in the Arctic.
FOOTNOTES
This study was carried out thanks to grants from the Swedish Natural Sciences Research Council (NFR, B-AA/BU 08424), the Kempe Foundation, the Swedish Agency for Nature Protection, and a starting grant from the Swedish Society for Nature Conservation. We thank the following field assistants and contributors for their work during the 1992–2001 field seasons: Juha Alatalo, Sünne Burmeister, Elizabeth Cooper, Lars Gerre, Karsten Hagmann, Annika Jägerbrand, Anna Karall, Olga Khitun, Per Larsson, Antonia Liess, Karin Lindblad, Eva Nilsson, Henrik Pärn, Priitta Pöyhtäri, Raija Kaarina Ratilainen, Sarah Richardson, Satu Räsänen, Anna Stenström, Mikael Stenström, and Tomás Zicha. We also thank the Abisko Natural Sciences Research Station (ANS) and its staff for help and hospitality. 
LITERATURE CITED
Alatalo J. M U Molau 1995 Female frequencies and sex-associated morphological differences in gynodioecious Silene acaulis (Caryophyllaceae). Nordic Journal of Botany 15: 251-256[ISI]
Alatalo J. M Ø Totland 1997 Response to simulated climatic change in an alpine and subarctic pollen-risk strategist, Silene acaulis. Global Change Biology 3: (Supplement 1) 74-79
Arft A. M et al 1999 Response patterns of arctic plant species to experimental warming. Ecological Monographs 69: 491-511[CrossRef]
Billings W. D L. C Bliss 1959 An alpine snowbank environment and its effect on vegetation, plant development, and productivity. Ecology 40: 388-397[CrossRef][ISI]
Björk R 2000 Soil properties and plant community types at Latnjajaure. B.Sc. thesis, Botanical Institute, Göteborg University, Göteborg, Sweden
Chapin F. S. III M. S Bret-Harte S. E Hobbie H Zhong 1996 Plant functional types as predictors of transient responses of arctic vegetation to global change. Journal of Vegetation Science 7: 347-358[CrossRef][ISI]
Diekmann M 1996 Relationship between flowering phenology of perennial herbs and meteorological data in deciduous forests of Sweden. Canadian Journal of Botany 74: 528-537[ISI]
Fitter A. H R. S. R Fitter 2002 Rapid changes in flowering time in British plants. Science 296: 1689-1691
Fitter A. H R. S. R Fitter I. T. B Harris M. H Williamson 1995 Relationship between first flowering date and temperature in the flora of a locality in central England. Functional Ecology 9: 55-60
Graglia E S Jonasson A Michelsen I. K Schmidt 1997 Effects of shading, nutrient application and warming on leaf growth and shoot densities of dwarf shrubs in two arctic-alpine plant communities. Ecoscience 4: 191-198
Henry G. H. R U Molau 1997 Tundra plants and climate change: the International Tundra Experiment (ITEX). Global Change Biology 3: (Supplement 1) 1-9
Hultén E 1958 The Amphi-Atlantic plants. Almqvist & Wiksell, Stockholm, Sweden
Hultén E 1962 The Circumpolar plants. I. Vascular cryptogams, conifers, monocotyledons. Almqvist & Wiksell, Stockholm, Sweden
Hultén E 1970 The Circumpolar plants. II. Dicotyledons. Almqvist & Wiksell, Stockholm, Sweden
IPCC. 2001 Climate Change 2001: impacts, adaptations and vulnerability. Cambridge University Press, Cambridge, UK
Jones M. H C Bay U Nordenhäll 1997 Effects of experimental warming on arctic willows (Salix spp.): a comparison of responses from the Canadian High Arctic, Alaskan Arctic, and Swedish Subarctic. Global Change Biology 3: (Supplement 1) 55-60
Kling J 1997 Observations on sorted circle development, Abisko, northern Sweden. Permafrost and Periglacial Processes 8: 447-453[CrossRef][ISI]
Larsson E.-L U Molau 2001 Snowbeds trapping seed rain—a comparison of methods. Nordic Journal of Botany 21: 385-392[ISI]
Leith H 1974 Phenology and seasonality modeling. Springer-Verlag, Berlin, Germany
Lovett-Doust L 1981 Population dynamics and local specialization in a clonal perennial (Ranunculus repens). I. The dynamics of ramets in contrasting habitats. Journal of Ecology 69: 743-755[CrossRef]
Molau U 1993 Relationships between flowering phenology and life history strategies in tundra plants. Arctic and Alpine Research 25: 391-402[CrossRef][ISI]
Molau U 1996 ITEX climate stations. In U. Molau and P. Mølgaard [eds.], ITEX manual, 2nd ed., 6–10. Danish Polar Center, Copenhagen, Denmark
Molau U 1997 Phenology and reproductive success in arctic plants: susceptibility to climate change. In W. Oechel, T. V. Callaghan, T. Gilmanov, J. I. Holten, B. Maxwell, U. Molau, and B. Sveinbjörnsson [eds.], Global change and arctic terrestrial ecosystems, 153–170. Springer-Verlag, New York, New York, USA
Molau U 2001 Tundra plant responses to experimental and natural temperature changes. Memoirs of National Institute of Polar Research, special issue 54: 445–466
Molau U J. M Alatalo 1998 Responses of subarctic-alpine plant communities to simulated environmental change: biodiversity of bryophytes, lichens, and vascular plants. Ambio 27: 322-329[ISI]
Molau U J Kling K Lindblad R Björk J Dänhardt A Liess 2003 A GIS assessment of alpine biodiversity at a range of scales. In L. Nagy, G. Grabherr, Ch. Körner, and D. B. A. Thompson [eds.], Alpine biodiversity in Europe. Ecological Studies 167: 221–229. Springer-Verlag, Berlin
Molau U E.-L Larsson 2000 Seed rain and seed bank along an alpine altitudinal gradient. Canadian Journal of Botany 78: 728-747[ISI]
Nilsson Ö 1986 Nordisk fjällflora. Bonniers, Stockholm, Sweden
Polunin N 1951 The real Arctic: suggestions for its delimitation, subdivision and characterization. Journal of Ecology 39: 308-315[CrossRef]
Raunkiær C 1934 The life forms of plants and statistical plant geography. Clarendon Press, Oxford, UK
Shaver G. R N Fetcher F. S Chapin III 1986 Growth and flowering in Eriophorum vaginatum: annual and latitudinal variation. Ecology 67: 1524-1535[CrossRef][ISI]
Sokal R. R F. J Rohlf 1981 Biometry, 2nd ed. W. H. Freeman, New York, New York, USA
Sørensen T 1941 Temperature relations and phenology of the northeast Greenland flowering plants. Meddelelser om Grønland 125: 1-305
Stenström M F Gugerli G. H. R Henry 1997 Response of Saxifraga oppositifolia L. to simulated climate change at three contrasting latitudes. Global Change Biology 3: (Supplement 1) 44-54
Stenström A I. S Jónsdóttir 1997 Responses of the clonal sedge, Carex bigelowii, to two seasons of simulated climate change. Global Change Biology 3: (Supplement 1) 89-96
Thórhallsdóttir T. E 1998 Flowering phenology in the central highland of Iceland and implications for climatic warming in the Arctic. Oecologia 114: 43-49[CrossRef][ISI]
Walther G.-R E Post P Convey A Menzel C Parmesan T. J. C Beebee J.-M Fromentin O Hoegh-Guldberg F Bairlain 2002 Ecological responses to recent climate change. Nature 416: 389-395[CrossRef][Medline]
This article has been cited by other articles:
|
|
 |
|
  A. J. Miller-Rushing, T. Katsuki, R. B. Primack, Y. Ishii, S. D. Lee, and H. Higuchi Impact of global warming on a group of related species and their hybrids: cherry tree (Rosaceae) flowering at Mt. Takao, Japan Am. J. Botany, September 1, 2007; 94(9): 1470 - 1478. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. Miller-Rushing, R. B. Primack, D. Primack, and S. Mukunda Photographs and herbarium specimens as tools to document phenological changes in response to global warming Am. J. Botany, November 1, 2006; 93(11): 1667 - 1674. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
