Subjects

Abstract

Advancing phenology is one of the most visible effects of climate change on plant communities, and has been especially pronounced in temperature-limited tundra ecosystems. However, phenological responses have been shown to differ greatly between species, with some species shifting phenology more than others. We analysed a database of 42,689 tundra plant phenological observations to show that warmer temperatures are leading to a contraction of community-level flowering seasons in tundra ecosystems due to a greater advancement in the flowering times of late-flowering species than early-flowering species. Shorter flowering seasons with a changing climate have the potential to alter trophic interactions in tundra ecosystems. Interestingly, these findings differ from those of warmer ecosystems, where early-flowering species have been found to be more sensitive to temperature change, suggesting that community-level phenological responses to warming can vary greatly between biomes.

Main

Warmer temperatures associated with climate change have advanced the phenology of organisms around the world1,2,3, and both temperature increases and phenological changes have been especially pronounced in temperature-limited tundra ecosystems4,5,6,7. Tundra ecosystems encompass cold regions above the latitudinal tree line (Arctic tundra) or altitudinal tree line (alpine tundra). Remote sensing studies indicate broad patterns of changing seasonality of vegetation productivity at high latitudes over time in relation to climate warming8,9,10; however, phenological responses to warmer temperatures have been shown to differ greatly among species and locations, with some species shifting dates of flowering and flower senescence more than others11,12,13,14,15. Studies from temperate ecosystems have found that early-flowering species often advance phenological events more in response to warmer temperatures than later-flowering species1,16,17,18,19; however, to date, the relationship between flowering time and phenological sensitivity has not been tested across tundra ecosystems.

Evidence suggests that across northern tundra ecosystems, the phenology of plants from colder sites at higher latitudes changes more with warmer temperatures than the phenology of plants from warmer, more southern latitudes7,15,20. However, within tundra plant communities, phenological responses to warming are often species specific, with no clear responses of specific functional groups3,21,22,23,24,25,26,27 or phylogenetic relationships28. A better understanding of the drivers of variation in phenological sensitivity will help determine how species and plant communities will respond to climate change in the future3,23,29, as well as contribute to our understanding of the adaptive nature of species-specific phenological responses to climate change.

The timing of life-history events, such as flowering, is of critical importance in harsh tundra ecosystems, and the fitness consequences of different phenological responses to climatic drivers can be substantial30,31. Plants that track snowmelt dates and not temperature (or thermal sums) may risk exposure to freezing events that can damage flowers and reduce seed production during early snowmelt years32,33,34,35, whereas plants that flower too late risk not being able to fully develop seeds before the end of the growing season, and may be at a competitive disadvantage compared with plants that do respond22,36.

There are diverse life-history strategies among species in tundra plant communities, even within the short growing seasons experienced at high latitudes and altitudes21,22,37. These various strategies could influence the species-specific responses of plants to warmer temperatures12,37,38. The relative flowering time of a species compared with other species in the plant community (hereafter, its ‘phenological niche’) could help explain the variation in phenological responses among species in tundra ecosystems. The existence of different phenological niches could promote species coexistence in many ecosystems39,40,41, as phenological niches can strongly influence competitive and trophic interactions42. Differential shifts in the phenological niche could lead to trophic mismatches in tundra ecosystems, altering food webs and influencing the abundance of pollinators or herbivores12,43,44,45. Classifying organisms using phenological niches could thus be a useful way to predict how species will respond to changes in environmental conditions in the future38.

Measuring the relative importance of different environmental cues for Arctic and alpine species, such as temperature and snowmelt date, will help determine how species will respond as the climate warms23,29. Although temperature influences the date of snowmelt, snowmelt can be decoupled from temperature because it is also influenced by the amount and quality of precipitation over winter and spring13. The phenology of early-flowering plant species may be influenced more by photoperiod or the timing of snowmelt, whereas the phenology of late-flowering species is probably more dependent on thermal heat sums accumulated over the growing season22,46. If early-flowering tundra species are less responsive to changes in summer temperature than late-flowering species, increases in summer temperature will probably accelerate the flowering phenology of late-flowering species more than early-flowering species. Additionally, if temperatures towards the end of the growing season are rising more rapidly than temperatures at the beginning of the year, the flowering phenology of late-flowering species will advance more than that of early-flowering species14,15. In both cases, a more rapid advance of late- than early-flowering species would result in a contraction of the community-level flowering season (Fig. 1)12, which could substantially change competitive and trophic interactions12,31,44,47. In particular, shorter flowering seasons could also strongly limit resource availability for pollinators, especially if the phenologies of pollinator species are responding to different drivers than those of plant communities12,48.

Fig. 1: Conceptual diagram showing how warmer summer temperatures may shorten the length of the flowering season in tundra ecosystems.
Fig. 1

If the phenology of early-flowering plant species is influenced primarily by photoperiod or the timing of snowmelt, and does not respond appreciably to warmer summer temperatures, but the phenology of late-flowering species is mostly dependent on accumulated heat sums over the growing season, and does shift earlier with warmer summers, there may be a contraction of the overall flowering season during warmer years.

In this data synthesis, we test how the temperature sensitivity of flowering relates to the phenological niches of tundra species using flowering observations of a total of 253 species, 23 sites and up to 20 years from Arctic and alpine ecosystems around the world, both from long-term monitoring plots and warming experiments (Fig. 2). With this global dataset, we tested three main hypotheses. (1) The flowering phenology of late-flowering tundra species is more sensitive to warmer summer temperatures than the flowering phenology of early-flowering species. We tested this hypothesis with both observational and experimental data, and hypothesized that the results would be similar for both observational and experimental data (that is, late-flowering species would be more sensitive to natural and experimental warming). (2) If late-flowering species are flowering earlier, but early-flowering species are not, the community-level flowering seasons will be shorter in warmer years. (3) As average summer temperatures at tundra sites have warmed in the recent past, the duration of community-level flowering seasons has decreased over this time period. We examined how the phenological niche of a species influenced the sensitivity of first flowering dates (FFDs) and flower senescence (that is, last flowering dates (LFDs)) to summer temperature indices, snowmelt date and experimental warming. To test for a contraction of community-level flowering seasons with warmer summers and over time, we investigated the relationship between community flowering season length and both mean June–July temperatures and year for six sites with observations of four or more species over ten or more years.

Fig. 2: Map of long-term observational and experimental warming studies.
Fig. 2

Site names are listed in order from the site with the coldest (2.8 °C) to the site with the warmest (11.9 °C) summer temperatures (June–August for Northern Hemisphere sites, and December to February for the Southern Hemisphere site; Supplementary Fig. 1). Site symbols shown on the map correspond to the symbols and colours in Figs. 3 and 4. Asterisks indicate sites used in community flowering season analyses.

Results

FFDs of late-flowering species were more temperature sensitive than those of early-flowering species (that is, FFDs of late-flowering species advanced more per °C increase in summer temperature, and in response to experimental warming, than those of early-flowering species; Figs. 3a and 4a, Supplementary Fig. 1 and Supplementary Table 4). The results of analyses using June temperature for all species, or the average daily temperature from snowmelt through the average flowering date, also indicated a significant influence of phenological niche on temperature sensitivity of flowering (Fig. 3b,c and Supplementary Table 4). However, the phenological niche of a species did not influence the sensitivity of FFDs to snowmelt timing (Fig. 3d and Supplementary Table 4). Overall, species from sites with colder summer temperatures had greater temperature sensitivity of FFDs (Supplementary Table 4). Analyses from warming experiments yielded similar results, with greater differences in FFDs between experimentally warmed and control plots for late-flowering species than for early-flowering species (Fig. 4a). There was no influence of phenological niche on the temperature sensitivity of LFDs in either long-term monitoring plots or warming experiments (Supplementary Table 5 and Fig. 4b).

Fig. 3: Temperature sensitivity of FFDs was greater for late- versus early-flowering species.
Fig. 3

ad, Relationships between the phenological niches of species and sensitivities of FFDs to mean monthly temperature until flowering (a), mean June temperature (b), mean daily temperature between snowmelt and flowering (c) and the date of snowmelt (d). Points represent the estimated temperature sensitivities for each species at each site, and vertical grey lines span the 95% credible intervals (CIs) for each species-by-site-level estimate. Colours and symbols correspond to site names in Fig. 2. The ‘phenological niche’ is the average flowering date of a species compared with the site-level mean flowering date of all species at a site. Solid black lines denote significant hierarchical model slopes, dashed black lines indicate non-significant model slopes and the horizontal grey line denotes the zero line. Hierarchical model slopes and 95% CIs are listed in the bottom left of each graph. The phenological niches significantly predict phenological responses (at the 5% level) if the 95% CIs do not overlap zero.

Fig. 4: The change in FFDs in response to experimental warming was greater for late- versus early-flowering species.
Fig. 4

a,b, Relationships are shown between phenological niches of species and the timing of FFDs (a) and LFDs (b) in experimentally warmed plots compared with control plots. Points represent the estimated temperature sensitivities for each species at each site, and vertical grey lines span the 95% CIs for each species-by-site-level estimate. Colours and symbols correspond to site names in Fig. 2. The ‘phenological niche’ is the average flowering date of a species compared with the site-level mean flowering date of all species at a site. Solid black lines denote significant hierarchical model slopes, dashed black lines indicate non-significant model slopes and the horizontal grey line denotes the zero line. Hierarchical model slopes and 95% CIs are listed in the bottom left of each graph. The phenological niches significantly predict phenological responses (at the 5% level) if the 95% CIs do not overlap zero.

The community-level flowering seasons across the 6 sites with 10 or more years of data were 3.96 d shorter per 1 °C warmer June–July temperature (95% CI = −7.31 to −0.79; Fig. 5a and Supplementary Table 5). The length of the flowering season was estimated as the duration between the average FFD of the earliest-flowering species and the average LFD of the latest-flowering species per site in each year. Community-level flowering seasons became shorter over time at all six sites, but the change was significant only at Alexandra Fiord, Daring Lake and Zackenberg. Across all sites, the flowering season length shortened by 0.43 d yr−1, but the credible interval (CI) on this parameter overlapped 0 (95% CI = −0.87 to 0.06; Fig. 5b). Annual June–July temperatures increased by 0.07 °C yr−1 (95% CI = 0.02 to 0.12; Fig. 5c).

Fig. 5: Warming was related to the change in the duration of the flowering season over time at sites across the tundra biome.
Fig. 5

a, Difference in the duration of the community-level flowering season compared with the difference in mean June–July temperatures from site averages. b, Change in the duration of the community-level flowering season over time. c, Yearly June–July temperature over time. Flowering season length and average June–July temperatures were mean-centred for each site so that they could be compared across sites. Points represent the change in the community-level flowering season per site and year. Solid black lines denote significant hierarchical model slopes and dashed black lines indicate non-significant model slopes. Coloured bands show the 95% CIs for site-level slopes. Hierarchical model slopes and 95% CIs are listed in the bottom left of each graph.

Discussion

Our results reveal an overall shortening of community-level flowering seasons with summer warming across the tundra biome. We additionally found evidence of a contraction of the community-level flowering season over time at a subset of sites. In both cases, the shortening of the flowering season was due to greater temperature sensitivity of flowering of late-flowering than early-flowering species. On average, the temperature sensitivity of FFDs was greater for tundra species that flowered later in the growing season compared with those that flowered earlier. This pattern was evident both in long-term monitoring plots over time and in warming experiments. Additionally, observations from long-term monitoring plots indicated that, on average, plants at colder sites were more phenologically sensitive, consistent with results from ref. 20 using a largely overlapping dataset, and that late-flowering plant species at the coldest tundra sites exhibited the highest phenological sensitivities in the dataset. Our analyses of long-term monitoring and experimental warming data indicate that late-flowering tundra species may alter their flowering phenology more than early-flowering species in a warmer world, resulting in a shortening of community-level flowering seasons at sites across the tundra biome.

The finding of greater temperature sensitivity of late-flowering species differs from the results of many studies conducted at lower latitudes and altitudes6,18,19,49. Studies from warmer biomes found that early-flowering species often advance phenological events more in response to warmer temperatures than late-flowering species1,16,17,18,19,50,51. Mid- and late-season moisture limitation probably plays a greater role in structuring the phenology of plants in warmer ecosystems52. However, in cold tundra ecosystems with relatively short summers, moisture limitation may not be as important a phenological driver as in warmer, drier ecosystems53. Additionally, selection might be stronger at the start of the growing season under the harsher climate conditions experienced by early-flowering plants in tundra sites relative to more temperate biomes46.

Our finding of a contraction of the flowering season with warmer temperatures also differs from studies in other ecosystems. Some studies have found a divergence of flowering dates of early- versus late-flowering species with warming in temperate grasslands49, montane and subalpine meadows54,55 and deserts53, with less overlap of the flowering times of species49, and a mid-season depression in flower abundance54,55. Individual studies conducted in temperate ecosystems, and global meta-analyses of phenology experiments and long-term monitoring projects, have concluded that early-flowering species are more responsive to climate warming6,18,51. However, our results show that Arctic and alpine plants exhibit the opposite pattern, suggesting that community-level phenological responses to warming can vary greatly among biomes19,56.

For the six Arctic sites with over ten years of observations, we documented a contraction of the flowering season with warmer temperatures and a trend towards shorter flowering seasons over time, although this pattern was not significant at all sites. A contraction of the flowering season is in agreement with previous single-site studies in Arctic ecosystems5,12,48. Shorter flowering seasons could lead to possible phenological mismatches if late-season pollinators or herbivores are not following the same cues as late-season plant species48,57. Additionally, less dispersion among the flowering times of species in a community may increase competition for pollinators58 or, alternatively, increase exposure to more pollinators because plant species are all flowering at similar times59. However, it is important to note that we did not directly measure how the abundance of plant species, or the abundance of open flowers, changed with temperature or over time. The timing of peak flowering may shift less than the timing of FFDs55; thus, changes in the coverage and abundance of flowers over the season may exhibit different patterns than changes in the overall length of the flowering season60.

Increased temperature sensitivity of flowering may be advantageous if it allows plants to track ideal temperature conditions for growth and reproduction30,61. Our results suggest that late-flowering species that track temperature more than snowmelt date or photoperiod may be more able to optimize the timing of flowering, and this could be an advantage as the temperature increases or becomes more variable29,62. Phenological plasticity may also be indicative of plasticity of other plant traits, so plant species that can shift phenology to changing conditions may be better able to adjust to climate change over time. To date, there have been few studies of the relationship of phenological traits versus other plant traits and changes in plant abundance (but see refs 30,61). However, as phenological data for tundra plant species accumulate, the next logical step will be to link phenological measurements to performance measurements to aid predictions of vegetation change in tundra ecosystems in the future63.

Phenological responses are one of the most easily observable effects of climate change on plant communities2, but identifying the underlying mechanisms driving phenological responses to warming is crucial in accurately estimating food-web dynamics and plant–pollinator interactions. Our data synthesis demonstrates an agreement between long-term and experimental data to identify how plants respond to warmer temperatures64,65. In temperature-limited tundra ecosystems, late-flowering species advance flowering more in warmer years, and this can lead to a contraction of the flowering season of the entire plant community. Additionally, these changes are most pronounced at the coldest tundra sites where temperature increases have been greatest20. Thus, our study demonstrates that the phenological niches of plant species can be useful predictors of how the flowering of tundra species will respond to warmer temperatures, and can aid predictions of plant and ecosystem responses to climate change in the future.

Methods

Compilation of the flowering phenology database

We compiled a database of flowering phenology observations from a total of 253 species at 23 sites in Arctic and alpine ecosystems from both long-term monitoring plots and warming experiments (Supplementary Table 1 and Fig. 2). Portions of the dataset were analysed and reported in Oberbauer et al.7 and Prevéy et al.20; however, two additional monitoring sites and ten additional warming experiments are included in this analysis (Supplementary Table 1). Phenological observations were made at each site following a standardized protocol that was originally developed for the International Tundra Experiment network66,67. Following the International Tundra Experiment protocol, observers recorded the phenological status of plants one to three times per week over the snow-free season, and specifically recorded the FFD and LFD of each species per individual or plot. The FFD was defined as the date when the first flower was open, the first pollen was visible or the first anthers were exposed. The LFD was defined as the date when the withering of anthers, first petal drop or last petal drop was observed. However, both FFD and LFD were recorded consistently at each site over time. We include data only from long-term monitoring plots that had three or more years of flowering phenology observations per species per plot.

Effects of species phenological niches on the sensitivity of flowering

We calculated the phenological niche of a species at each site as the average FFD of the species at each site across all years of measurements (Supplementary Table 2). We examined the relationship between phenological niche and temperature (expressed in several ways) and snowmelt dates at long-term monitoring plots. Temperature was expressed as the mean monthly temperature until flowering, the mean June temperature or the mean daily temperature between snowmelt and flowering. Flowering dates for the Southern Hemisphere alpine site were adjusted by 210 d to match those of the Northern Hemisphere growing season, and to assist with model convergence in analyses. We specified mean monthly temperature until flowering separately for each species and site as the average monthly air temperature from June through the average month of flowering, except for 29 site-by-species combinations where species flowered in May, for which we used the average May temperature (Supplementary Table 2). For example, if the phenological niche of a species was 30 June, the mean June temperature was used as the summer temperature variable for that species. However, if the phenological niche was 15 July, the average June–July temperature was used (Supplementary Table 2). To test the influence of the temperature windows on the results, we also performed the analyses with June temperature as the predictor variable for all sites and species, because preliminary analysis showed that June temperature was the strongest predictor of flowering across all species and sites (Supplementary Table 2). We used average monthly temperatures because they were available for all sites in the analyses; thus allowing us to incorporate the largest set of phenological data available. We recognize that using monthly mean temperatures may bias the results, as the sensitivity of flowering time for species flowering in the early parts of months is obviously not affected by temperatures experienced after they flower. Thus, for the subset of 12 sites with both daily temperature data and snowmelt dates available we calculated the mean daily temperature as the average daily air temperature from the date of snowmelt through the average date of flowering for each species and year. Finally, we examined the association between the timing of snowmelt and flowering in long-term monitoring plots by assessing the phenological niches of species in relation to snowmelt timing for the subset of 13 sites that had recorded snowmelt dates over time.

Models also included the effect of mean site-level summer temperatures (June–August) from 1981–2000 as an additional predictor variable of species phenological responses, since a previous synthesis found that flowering dates of species from colder tundra sites were more sensitive to changes in temperature than those from warmer sites20. Mean monthly temperatures for sites were obtained from local weather stations when available. If no long-term (1981–2010) weather data were available near sites, mean monthly temperatures were estimated using 0.5° gridded temperature data from the Climate Research Unit68 (Supplementary Table 1). Temperatures and phenological niches were mean-centred by site for all species for long-term monitoring plot data. Plot within site and year within site were included as random variables. We also tested for the interaction between phenological niche and temperature.

In total, the analyses of FFDs with summer temperature windows or mean June temperatures as predictor variables included 14,324 observations from 318 unique site-by-species combinations at 19 sites. The analyses of FFDs with snowmelt date included 9,918 observations from 141 unique site-by-species combinations at 13 sites, and the analyses of FFDs using average daily temperatures included 9,713 observations from 143 unique site-by-species combinations at 11 sites. The analyses of LFDs with summer temperature windows or mean June temperatures as predictor variables included 9,226 observations from 88 unique site-by-species combinations at 11 sites. The analyses of LFDs with snowmelt date included 7,661 observations from 80 unique site-by-species combinations at 11 sites, and the analyses of LFDs using average daily temperatures included 7,341 observations from 74 unique site-by-species combinations at 9 sites.

Effects of phenological niches on the temperature sensitivity of flowering in warming experiments

We examined observations from warming experiments that utilized open-top chambers (OTCs) to investigate how experimental warming influenced the flowering dates of species with different phenological niches. In the warming experiments, plots were warmed with around 1 m2 fibreglass or polycarbonate OTCs, in either cone or hexagonal shapes, that increased the air temperature by 0.5–3.0 °C66,69,70,71 (Supplementary Table 3). The OTCs were either placed on plots only over the summer or left on plots throughout the year, depending on the site (Supplementary Table 3).

To examine how the phenological niche of a species influenced its phenological sensitivity to experimental warming, we first calculated the average difference in the timing of phenological events (either FFD or LFD) between control and experimentally warmed plots at each site and year for every species that occurred in both treatments. Then, we assessed the relationship between the phenological niches of each species and the difference in the number of days between the FFD or LFD in experimentally warmed and control plots for each species, site and year combination. Mean site-level summer temperature was not included as a predictor variable in the warming experiment analyses because the amount of experimental warming differed between experiments at different sites (Supplementary Table 3). We also examined how differences in the amount of warming in different warming experiments may have altered the results by calculating the difference in the number of days between the FFDs or the LFDs in experimentally warmed and control plots divided by the mean number of degrees of warming reported for chambers at each site or subsite within site (Supplementary Table 3) to obtain an estimate of the change in flowering date per °C of warming.

In total, the analyses of FFDs in warming experiments included 1,219 flowering observations from 164 unique site-by-species combinations at 16 sites. Analyses of LFDs in warming experiments included 743 observations from 96 unique site-by-species combinations at 11 sites.

Statistical analyses of effects of phenological niches on sensitivity of flowering

To statistically analyse phenological observations over the different numbers of sites, years of observations and species, we used Bayesian hierarchical modelling. This approach allowed for estimation of the uncertainties of phenological responses among sites, plots, years and species, and the incorporation of these uncertainties in the final correlation of phenological niche and phenological responses per species per site72.

For data from long-term monitoring plots, we used two-level regression models. At the lower level, we estimated phenological sensitivities by relating the date of phenological events (FFD or LFD) to the temperature or snowmelt date. At the higher (species) level, we related species’ phenological sensitivities to their phenological niches. For data from warming experiments, the difference (in days) of FFD or LFD between warmed and control plots was directly included as a response variable in the species-level regression.

We fit Bayesian models using the programme Stan73, which was accessed using the package Rstan74 in the statistical programme R 3.2.2 (ref. 75). Each model was run with 2 chains of 20,000 iterations, using Hamiltonian Monte Carlo sampling. We used flat priors for all parameter estimates. Full model details and codes are included in Supplementary Information Section 7. We checked for convergence of chains for all parameters both visually with trace plots and with the Gelman–Rubin convergence statistic76. Trace plots showed that chains mixed well and converged to stationary distributions for all parameter estimates. Gelman–Rubin convergence statistics for parameter estimates of all models were <1.02.

Duration of flowering season

To test for a contraction of community-level flowering seasons in association with warmer summers, we conducted analyses that only included sites with FFDs and LFDs for four or more species over ten or more years. This limited analyses to the six Arctic sites with long-term monitoring data: Alexandra Fiord, Atqasuk, Utqiaġvik, Daring Lake, Toolik Lake and Zackenberg. Flower count or peak flowering data were not available for all sites, so we used a proxy for the community flowering season calculated as the number of days between the average FFD of the earliest-flowering species at a site per year and the average LFD of the latest-flowering species at a site per year. We used the earliest- and latest-flowering species in each year to avoid any bias caused by uneven shifts in flowering times among species. Although changes in FFDs and LFDs are not always representative of changes over the entire flowering season55,77, we believe our proxy can provide an estimate of how the length of the flowering season may change with future warming. Additionally, a previous synthesis found that reproductive phenological events within the same species are highly correlated7.

We related this proxy for the duration of the community-level flowering season to the average June–July temperature at a site per year using a Bayesian hierarchical modelling approach. We mean-centred both flowering season length and average June–July temperatures for each site so we could assess the relationship between the change in community-level flowering seasons and the change in June–July temperatures across sites. Because all sites chosen for these analyses had relatively long records of phenological measurements (>10 years), we also examined whether flowering season length or June–July temperatures have changed significantly over time. We analysed associations between community flowering season length and summer temperature and time with a Bayesian hierarchical model using mean-centred June–July temperature as the predictor variable for the temperature sensitivity models and year as the predictor variable for the temporal change models, and an intercept and slope that varied by site. We also examined whether mean June–July temperatures changed over time using the same models with year as the predictor variable. Full model details and codes are included in Supplementary Information Section 7.

Reporting Summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

The data that support the findings of this study have been archived in the Polar Data Catalogue: https://doi.org/10.21963/12961.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    Fitter, A. H. & Fitter, R. S. R. Rapid changes in flowering time in British plants. Science 296, 1689–1691 (2002).

  2. 2.

    Parmesan, C. & Yohe, G. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421, 37–42 (2003).

  3. 3.

    Thackeray, S. J. et al. Phenological sensitivity to climate across taxa and trophic levels. Nature 535, 241–245 (2016).

  4. 4.

    Arft, A. M. et al. Responses of tundra plants to experimental warming: meta-analysis of the International Tundra Experiment. Ecol. Monogr. 69, 491–511 (1999).

  5. 5.

    Høye, T. T., Post, E., Meltofte, H., Schmidt, N. M. & Forchhammer, M. C. Rapid advancement of spring in the High Arctic. Curr. Biol. 17, R449–R451 (2007).

  6. 6.

    Parmesan, C. Influences of species, latitudes and methodologies on estimates of phenological response to global warming. Glob. Change Biol. 13, 1860–1872 (2007).

  7. 7.

    Oberbauer, S. F. et al. Phenological response of tundra plants to background climate variation tested using the International Tundra Experiment. Phil. Trans. R. Soc. B 368, 20120481 (2013).

  8. 8.

    Xu, L. et al. Temperature and vegetation seasonality diminishment over northern lands. Nat. Clim. Change 3, 581–586 (2013).

  9. 9.

    Park, T. et al. Changes in growing season duration and productivity of northern vegetation inferred from long-term remote sensing data. Environ. Res. Lett. 11, 084001 (2016).

  10. 10.

    Xu, C., Liu, H., Williams, A. P., Yin, Y. & Wu, X. Trends toward an earlier peak of the growing season in Northern Hemisphere mid‐latitudes. Glob. Change Biol. 22, 2852–2860 (2016).

  11. 11.

    Bradley, N. L., Leopold, A. C., Ross, J. & Huffaker, W. Phenological changes reflect climate change in Wisconsin. Proc. Natl Acad. Sci. USA 96, 9701–9704 (1999).

  12. 12.

    Høye, T. T., Post, E., Schmidt, N. M., Trøjelsgaard, K. & Forchhammer, M. C. Shorter flowering seasons and declining abundance of flower visitors in a warmer Arctic. Nat. Clim. Change 3, 759–763 (2013).

  13. 13.

    Bjorkman, A. D., Elmendorf, S. C., Beamish, A. L., Vellend, M. & Henry, G. H. R. Contrasting effects of warming and increased snowfall on Arctic tundra plant phenology over the past two decades. Glob. Change Biol. 21, 4651–4661 (2015).

  14. 14.

    Panchen, Z. A. & Gorelick, R. Flowering and fruiting responses to climate change of two Arctic plant species, purple saxifrage (Saxifraga oppositifolia) and mountain avens (Dryas integrifolia). Arct. Sci. 1, 45–58 (2015).

  15. 15.

    Panchen, Z. A. & Gorelick, R. Prediction of Arctic plant phenological sensitivity to climate change from historical records. Ecol. Evol. 7, 1325–1338 (2017).

  16. 16.

    Price, M. V. & Waser, N. M. Effects of experimental warming on plant reproductive phenology in a subalpine meadow. Ecology 79, 1261–1271 (1998).

  17. 17.

    Dunne, J. A., Harte, J. & Taylor, K. J. Subalpine meadow flowering phenology responses to climate change: integrating experimental and gradient methods. Ecol. Monogr. 73, 69–86 (2003).

  18. 18.

    Menzel, A. et al. European phenological response to climate change matches the warming pattern. Glob. Change Biol. 12, 1969–1976 (2006).

  19. 19.

    Miller-Rushing, A. J. & Inouye, D. W. Variation in the impact of climate change on flowering phenology and abundance: an examination of two pairs of closely related wildflower species. Am. J. Bot. 96, 1821–1829 (2009).

  20. 20.

    Prevéy, J. S. et al. Greater temperature sensitivity of plant phenology at colder sites: implications for convergence across northern latitudes. Glob. Change Biol. 23, 2660–2671 (2017).

  21. 21.

    Shaver, G. R. & Kummerow, J. in Arctic Ecosystems in a Changing Climate: An Ecophysiological Perspective (eds Chapin, F. S., Jefferies, R. L., Reynolds, J. F., Shaver, G. R. & Svoboda, J.) 193–211 (Academic Press, San Diego, 1992).

  22. 22.

    Molau, U. Relationships between flowering phenology and life history strategies in tundra plants. Arct. Alp. Res. 25, 391–402 (1993).

  23. 23.

    Keller, F. & Körner, C. The role of photoperiodism in alpine plant development. Arct. Antarct. Alp. Res. 35, 361–368 (2003).

  24. 24.

    Hollister, R. D., Webber, P. J. & Tweedie, C. E. The response of Alaskan arctic tundra to experimental warming: differences between short- and long-term responses. Glob. Change Biol. 11, 525–536 (2005).

  25. 25.

    Semenchuk, P. R., Elberling, B. & Cooper, E. J. Snow cover and extreme winter warming events control flower abundance of some, but not all species in High Arctic Svalbard. Ecol. Evol. 3, 2586–2599 (2013).

  26. 26.

    Iler, A. M. et al. Maintenance of temporal synchrony between syrphid flies and floral resources despite differential phenological responses to climate. Glob. Change Biol. 19, 2348–2359 (2013).

  27. 27.

    Post, E., Kerby, J., Pedersen, C. & Steltzer, H. Highly individualistic rates of plant phenological advance associated with arctic sea ice dynamics. Biol. Lett. 12, 20160332 (2016).

  28. 28.

    CaraDonna, P. J. & Inouye, D. W. Phenological responses to climate change do not exhibit phylogenetic signal in a subalpine plant community. Ecology 96, 355–361 (2015).

  29. 29.

    Körner, C. & Basler, D. Phenology under global warming. Science 327, 1461–1462 (2010).

  30. 30.

    Cleland, E. E. et al. Phenological tracking enables positive species responses to climate change. Ecology 93, 1765–1771 (2012).

  31. 31.

    Wheeler, H. C., Høye, T. T., Schmidt, N. M., Svenning, J.-C. & Forchhammer, M. C. Phenological mismatch with abiotic conditions—implications for flowering in Arctic plants. Ecology 96, 775–787 (2015).

  32. 32.

    Inouye, D. W. Effects of climate change on phenology, frost damage, and floral abundance of montane wildflowers. Ecology 89, 353–362 (2008).

  33. 33.

    Wipf, S., Stoeckli, V. & Bebi, P. Winter climate change in alpine tundra: plant responses to changes in snow depth and snowmelt timing. Clim. Change 94, 105–121 (2009).

  34. 34.

    Wheeler, J. A. et al. Increased spring freezing vulnerability for alpine shrubs under early snowmelt. Oecologia 175, 219–229 (2014).

  35. 35.

    Wheeler, J. A. et al. The snow and the willows: earlier spring snowmelt reduces performance in the low-lying alpine shrub Salix herbacea. J. Ecol. 104, 1041–1050 (2016).

  36. 36.

    Cooper, E. J., Dullinger, S. & Semenchuk, P. Late snowmelt delays plant development and results in lower reproductive success in the High Arctic. Plant Sci. 180, 157–167 (2011).

  37. 37.

    Parsons, A. N. et al. Growth responses of four sub-Arctic dwarf shrubs to simulated environmental change. J. Ecol. 82, 307–318 (1994).

  38. 38.

    Molau, U., Nordenhäll, U. & Eriksen, B. Onset of flowering and climate variability in an alpine landscape: a 10-year study from Swedish Lapland. Am. J. Bot. 92, 422–431 (2005).

  39. 39.

    Grubb, P. J. The maintenance of species-richness in plant communities: the importance of the regeneration niche. Biol. Rev. 52, 107–145 (1977).

  40. 40.

    Higgins, S. I., Delgado-Cartay, M. D., February, E. C. & Combrink, H. J. Is there a temporal niche separation in the leaf phenology of savanna trees and grasses? J. Biogeogr. 38, 2165–2175 (2011).

  41. 41.

    Sanz-Aguilar, A., Carrete, M., Edelaar, P., Potti, J. & Tella, J. L. The empty temporal niche: breeding phenology differs between coexisting native and invasive birds. Biol. Invasions 17, 3275–3288 (2015).

  42. 42.

    Wolkovich, E. M. & Cleland, E. E. The phenology of plant invasions: a community ecology perspective. Front. Ecol. Environ. 9, 287–294 (2011).

  43. 43.

    Hegland, S. J., Nielsen, A., Lázaro, A. & Bjerknes, A.-L. How does climate warming affect plant–pollinator interactions?. Ecol. Lett. 12, 184–195 (2009).

  44. 44.

    McKinnon, L., Picotin, M., Bolduc, E., Juillet, C. & Bêty, J. Timing of breeding, peak food availability, and effects of mismatch on chick growth in birds nesting in the High Arctic. Can. J. Zool. 90, 961–971 (2012).

  45. 45.

    Kerby, J. T. & Post, E. Advancing plant phenology and reduced herbivore production in a terrestrial system associated with sea ice decline. Nat. Commun. 4, 2514 (2013).

  46. 46.

    Wipf, S. Phenology, growth, and fecundity of eight subarctic tundra species in response to snowmelt manipulations. Plant Ecol. 207, 53–66 (2010).

  47. 47.

    Post, E., Pedersen, C., Wilmers, C. C. & Forchhammer, M. C. Warming, plant phenology and the spatial dimension of trophic mismatch for large herbivores. Proc. R. Soc. B 275, 2005–2013 (2008).

  48. 48.

    Schmidt, N. M. et al. An ecological function in crisis? The temporal overlap between plant flowering and pollinator function shrinks as the Arctic warms. Ecography 39, 1250–1252 (2016).

  49. 49.

    Sherry, R. A. et al. Divergence of reproductive phenology under climate warming. Proc. Natl Acad. Sci. USA 104, 198–202 (2007).

  50. 50.

    Steltzer, H. & Post, E. Seasons and life cycles. Science 324, 886–887 (2009).

  51. 51.

    Wolkovich, E. M. et al. Warming experiments underpredict plant phenological responses to climate change. Nature 485, 494–497 (2012).

  52. 52.

    Prevéy, J. S. & Seastedt, T. R. Seasonality of precipitation interacts with exotic species to alter composition and phenology of a semi-arid grassland. J. Ecol. 102, 1549–1561 (2014).

  53. 53.

    Diez, J. M. et al. Forecasting phenology: from species variability to community patterns. Ecol. Lett. 15, 545–553 (2012).

  54. 54.

    Aldridge, G., Inouye, D. W., Forrest, J. R. K., Barr, W. A. & Miller-Rushing, A. J. Emergence of a mid-season period of low floral resources in a montane meadow ecosystem associated with climate change. J. Ecol. 99, 905–913 (2011).

  55. 55.

    CaraDonna, P. J., Iler, A. M. & Inouye, D. W. Shifts in flowering phenology reshape a subalpine plant community. Proc. Natl Acad. Sci. USA 111, 4916–4921 (2014).

  56. 56.

    Cook, B. I. et al. Sensitivity of spring phenology to warming across temporal and spatial climate gradients in two independent databases. Ecosystems 15, 1283–1294 (2012).

  57. 57.

    Høye, T. T. et al. Phenology of High-Arctic butterflies and their floral resources: species-specific responses to climate change. Curr. Zool. 60, 243–251 (2014).

  58. 58.

    Hocking, B. Insect–flower associations in the High Arctic with special reference to nectar. Oikos 19, 359–387 (1968).

  59. 59.

    Janzen, D. H. Synchronization of sexual reproduction of trees within the dry season in Central America. Evolution 21, 620–637 (1967).

  60. 60.

    Meng, F. D. et al. Changes in flowering functional group affect responses of community phenological sequences to temperature change. Ecology 98, 734–740 (2017).

  61. 61.

    Hulme, P. E. Contrasting impacts of climate-driven flowering phenology on changes in alien and native plant species distributions. New Phytol. 189, 272–281 (2011).

  62. 62.

    Craine, J. M., Wolkovich, E. M., Gene Towne, E. & Kembel, S. W. Flowering phenology as a functional trait in a tallgrass prairie. New Phytol. 193, 673–682 (2012).

  63. 63.

    Miller-Rushing, A. J., Høye, T. T., Inouye, D. W. & Post, E. The effects of phenological mismatches on demography. Philos. Trans. R. Soc. Lond. B 365, 3177–3186 (2010).

  64. 64.

    Elmendorf, S. C. et al. Experiment, monitoring, and gradient methods used to infer climate change effects on plant communities yield consistent patterns. Proc. Natl Acad. Sci. USA 112, 448–452 (2015).

  65. 65.

    Hollister, R. D. et al. Warming experiments elucidate the drivers of observed directional changes in tundra vegetation. Ecol. Evol. 5, 1881–1895 (2015).

  66. 66.

    Molau, U. & Mølgaard, P. International Tundra Experiment (ITEX) Manual (Danish Polar Center, 1996).

  67. 67.

    Henry, G. H. R. & Molau, U. Tundra plants and climate change: the International Tundra Experiment (ITEX). Glob. Change Biol. 3, 1–9 (1997).

  68. 68.

    Harris, I., Jones, P, Osborn, T. J. & Lister, D. H. Updated high-resolution grids of monthly climatic observations—the CRU TS3.10 dataset. Int. J. Climatol. 34, 623–642 (2014).

  69. 69.

    Marion, G. M. et al. Open-top designs for manipulating field temperature in high-latitude ecosystems. Glob. Change Biol. 3, 20–32 (1997).

  70. 70.

    Hollister, R. D., Webber, P. J., Nelson, F. E. & Tweedie, C. E. Soil thaw and temperature response to air warming varies by plant community: results from an open-top chamber experiment in Northern Alaska. Arct. Antarct. Alp. Res. 38, 206–215 (2006).

  71. 71.

    Walker, M. D. et al. Plant community responses to experimental warming across the tundra biome. Proc. Natl Acad. Sci. USA 103, 1342–1346 (2006).

  72. 72.

    Latimer, A. M. Geography and resource limitation complicate metabolism-based predictions of species richness. Ecology 88, 1895–1898 (2007).

  73. 73.

    Stan Modeling Language User’s Guide and Reference Manual Version 2.17.0 (Stan Development Team, 2017).

  74. 74.

    RStan: the R Interface to Stan R Package Version 2.17.3 (Stan Development Team, 2018).

  75. 75.

    R Development Core Team R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2017).

  76. 76.

    Gelman, A. & Rubin, D. B. Inference from iterative simulation using multiple sequences. Stat. Sci. 7, 457–472 (1992).

  77. 77.

    Miller-Rushing, A. J., Inouye, D. W. & Primack, R. B. How well do first flowering dates measure plant responses to climate change? The effects of population size and sampling frequency. J. Ecol. 96, 1289–1296 (2008).

Download references

Acknowledgements

We are grateful to the many individuals who established experiments and collected detailed phenological observations. There are too many to name them all; however, we especially thank: M. Dalle Fratte, D. Cooley, O. Durey, C. Eckert, J. F. Johnstone, C. Kennedy, V. Lamarre, G. Levasseur, C. Spiech, J. Svoboda and R. Wising; the Herschel Island Qikiqtaruk Territorial Park staff, including E. McLeod, S. McLeod, R. Joe, P. Lennie, D. Arey, L. Meyook, J. McLeod, P. Foisy, C. Gordon, J. Hansen, A. Rufus and R. Gordon; Quttinirpaaq National Park staff; the Greenland Ecosystem Monitoring team; and Warming and species Removal in Mountains (WaRM) coordinators N. Sanders, A. Classen and M. Sundqvist. These observations were made possible with the support of many funding agencies and grants, including: ArcticNet; the Natural Sciences and Engineering Research Council of Canada; the Canadian International Polar Year Program; the Polar Continental Shelf Program of Natural Resources Canada; the Danish Environmental Protection Agency; the Swiss Federal Institute for Forest, Snow and Landscape Research; the National Geographic Society; the US National Science Foundation (grant numbers PLR1525636, PLR1504141, PLR1433063, PLR1107381, PLR0119279, PLR0902125, PLR0856728, PLR1312402, PLR1019324, LTER 1026415, OPP1525636, OPP9907185, DEB1637686, 0856710, 9714103, 0632263, 0856516, 1432277, 1432982, 1504381, 1504224, 1433063, 0856728, 0612534, 0119279 and 9421755; the Danish National Research Foundation (grant CENPERM DNRF100); the Danish Council for Independent Research (Natural Sciences grant DFF 4181-00565); the Deutsche Forschungsgemeinschaft (grant: RU 1536/3-1); the Natural Environment Research Council (grant NE/M016323/1); the Department of Energy (grant SC006982); a Semper Ardens grant from the Carlsberg Foundation to N. J. Sanders; and an INTERACT Transnational Access grant.

Author information

Affiliations

  1. Pacific Northwest Research Station, US Forest Service, US Department of Agriculture, Olympia, WA, USA

    • Janet S. Prevéy
  2. WSL Institute for Snow and Avalanche Research , Davos, Switzerland

    • Janet S. Prevéy
    • , Christian Rixen
    • , Chelsea L. Chisholm
    •  & Sonja Wipf
  3. German Centre for Integrative Biodiversity Research Halle-Jena-Leipzig, Leipzig, Germany

    • Nadja Rüger
  4. Smithsonian Tropical Research Institute, Panama City, Panama

    • Nadja Rüger
  5. Department of Bioscience and Arctic Research Centre, Aarhus University, Aarhus, Denmark

    • Toke T. Høye
    •  & Niels Martin Schmidt
  6. Ecoinformatics and Biodiversity, Department of Bioscience, Aarhus University, Aarhus, Denmark

    • Anne D. Bjorkman
  7. Senckenberg Gesellschaft für Naturforschung, Biodiversity and Climate Research Centre, Frankfurt, Germany

    • Anne D. Bjorkman
  8. University of Edinburgh, Edinburgh, Scotland

    • Isla H. Myers-Smith
  9. Institute for Arctic and Alpine Research, University of Colorado, Boulder, CO, USA

    • Sarah C. Elmendorf
    • , Jane G. Smith
    •  & Katharine N. Suding
  10. Northern Great Plains Inventory and Monitoring Network, National Park Service, Rapid City, SD, USA

    • Isabel W. Ashton
  11. Department of Science and High Technology, Università degli Studi dell’Insubria, Como, Italy

    • Nicoletta Cannone
  12. Center for Macroecology, Evolution and Climate, Natural History Museum of Denmark, Copenhagen, Denmark

    • Chelsea L. Chisholm
  13. Environment and Natural Resources, Government of the Northwest Territories, Yellowknife, Northwest Territories, Canada

    • Karin Clark
  14. Institute for Arctic and Marine Biology, UiT - The Arctic University of Norway, Tromsø, Norway

    • Elisabeth J. Cooper
    •  & Philipp R. Semenchuk
  15. Center for Permafrost, Department of Geosciences and Natural Resource Management, University of Copenhagen, Copenhagen, Denmark

    • Bo Elberling
  16. Faroese Museum of Natural History, Hoyvík, Faroe Islands

    • Anna Maria Fosaa
  17. Department of Geography, University of British Columbia, Vancouver, British Columbia, Canada

    • Greg H. R. Henry
  18. Biology Department, Grand Valley State University, Allendale, MI, USA

    • Robert D. Hollister
  19. Institute of Life and Environmental Sciences, University of Iceland, Reykjavík, Iceland

    • Ingibjörg Svala Jónsdóttir
  20. University Centre in Svalbard, Longyearbyen, Norway

    • Ingibjörg Svala Jónsdóttir
  21. Faculty of Environmental Sciences and Natural Resource Management, Norwegian University of Life Sciences, Ås, Norway

    • Kari Klanderud
  22. Department of Botany, University of British Columbia, Vancouver, British Columbia, Canada

    • Christopher W. Kopp
  23. Université du Québec à Trois-Rivières, Trois-Rivieres, Québec, Canada

    • Esther Lévesque
  24. Center for Ecosystem Science and Society, Northern Arizona University, Flagstaff, AZ, USA

    • Marguerite Mauritz
    •  & Edward Schuur
  25. Department of Biology and Environmental Sciences, University of Gothenburg, Gothenburg, Sweden

    • Ulf Molau
  26. Woods Hole Research Center, Falmouth, MA, USA

    • Susan M. Natali
  27. Department of Biological Sciences, Florida International University, Miami, FL, USA

    • Steven. F. Oberbauer
    •  & Tiffany Troxler
  28. Department of Biology, Dalhousie University, Halifax, Nova Scotia, Canada

    • Zoe A. Panchen
  29. Department of Wildlife, Fish and Conservation Biology, University of California, Davis, Davis, CA, USA

    • Eric Post
  30. Department of Botany and Biodiversity Research, University of Vienna, Vienna, Austria

    • Sabine B. Rumpf
    •  & Philipp R. Semenchuk
  31. Department of Ecology and Evolutionary Biology, University of Colorado, Boulder, CO, USA

    • Katharine N. Suding
  32. Department of Biological Sciences, University of Bergen, Bergen, Norway

    • Ørjan Totland
  33. Centre for Integrative Ecology, School of Life and Environmental Sciences, Deakin University, Geelong, Victoria, Australia

    • Susanna Venn
  34. Research Centre for Applied Alpine Ecology, La Trobe University, Melbourne, Victoria, Australia

    • Carl-Henrik Wahren
  35. UArctic and University of Oulu, Oulu, Finland

    • Jeffrey M. Welker
  36. Department of Biological Sciences, University of Alaska, Anchorage, AK, USA

    • Jeffrey M. Welker

Authors

  1. Search for Janet S. Prevéy in:

  2. Search for Christian Rixen in:

  3. Search for Nadja Rüger in:

  4. Search for Toke T. Høye in:

  5. Search for Anne D. Bjorkman in:

  6. Search for Isla H. Myers-Smith in:

  7. Search for Sarah C. Elmendorf in:

  8. Search for Isabel W. Ashton in:

  9. Search for Nicoletta Cannone in:

  10. Search for Chelsea L. Chisholm in:

  11. Search for Karin Clark in:

  12. Search for Elisabeth J. Cooper in:

  13. Search for Bo Elberling in:

  14. Search for Anna Maria Fosaa in:

  15. Search for Greg H. R. Henry in:

  16. Search for Robert D. Hollister in:

  17. Search for Ingibjörg Svala Jónsdóttir in:

  18. Search for Kari Klanderud in:

  19. Search for Christopher W. Kopp in:

  20. Search for Esther Lévesque in:

  21. Search for Marguerite Mauritz in:

  22. Search for Ulf Molau in:

  23. Search for Susan M. Natali in:

  24. Search for Steven. F. Oberbauer in:

  25. Search for Zoe A. Panchen in:

  26. Search for Eric Post in:

  27. Search for Sabine B. Rumpf in:

  28. Search for Niels Martin Schmidt in:

  29. Search for Edward Schuur in:

  30. Search for Philipp R. Semenchuk in:

  31. Search for Jane G. Smith in:

  32. Search for Katharine N. Suding in:

  33. Search for Ørjan Totland in:

  34. Search for Tiffany Troxler in:

  35. Search for Susanna Venn in:

  36. Search for Carl-Henrik Wahren in:

  37. Search for Jeffrey M. Welker in:

  38. Search for Sonja Wipf in:

Contributions

J.S.P. and C.R. designed and led the study. J.S.P. and C.R. led the collection of data for the phenology database. J.S.P., N.R., A.D.B., I.H.M.-S. and S.C.E. performed the statistical analyses. J.S.P., C.R., N.R., T.T.H., A.D.B., I.H.M.-S. and S.C.E. drafted the paper. J.S.P., C.R., A.D.B., I.H.M.-S., I.W.A., N.C., K.C., C.C., E.J.C., B.E., A.M.F., G.H.R.H., R.D.H., I.S.J., K.K., C.W.K., E.L., M.M., U.M., S.N., S.O., Z.A.P., E.P., S.B.R., N.M.S., E.S., P.R.S., J.G.S., K.N.S., Ø.T., T.T., S.V., C.-H.W., J.M.W. and S.W. contributed data. All authors were involved in writing and editing the manuscript.

Competing interests

The authors declare no competing interests.

Corresponding author

Correspondence to Janet S. Prevéy.

Supplementary information

  1. Supplementary Information

    Supplementary Tables 1–7, Supplementary Figure 1, Supplementary Code and Supplementary References

  2. Reporting Summary

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/s41559-018-0745-6