Introduction

In recent times, shifting phenology is considered to be one of the pronounced responses of vegetation to global climate change (Zohner et al. 2017; Liu and Zhang 2020). The ongoing climate warming is causing either a significant advancement in spring flowering phenology in temperate ecosystems or a significant delay in summer flowering phenology of many plant species, thereby altering their key ecological processes and biotic interactions (Chuine and Beaubien 2001; Elzinga et al. 2007; Cleland et al. 2012; Fridley 2012; Iler et al. 2021; Rosbakh et al. 2021). A better understanding of the relationship between contemporary climate warming and the associated phenological shifts of species is needed in order to unravel the mechanisms underlying these shifts and to better predict future changes in phenological events (Hassan et al. 2021). A large number of studies, particularly from the developed world countries, have reported these phenological shifts and investigated their relationship with warming climate (Wang et al. 2018; Love and Mazer 2021; Rosbakh et al. 2021). Most of these phenological studies have employed the observational approach to explore the effect of climate change on plant phenology (Lindh et al. 2018). Other studies adopted the herbarium-based reconstruction (Munson and Long 2017) or dated-photographic data approaches (Miller-Rushing et al. 2006; Sparks et al. 2006; Prevéy et al. 2020). The limitation of these approaches is that there are very few such long-term phenological observation datasets available, that too only in some countries from the developed world (Aono and Saito 2010). Although the use of herbarium and old photographic approaches is advantageous in providing the temporal depth, these datasets are beset with large temporal data gaps (Willis et al. 2017). In addition, these approaches limit our ability to distinguish the independent influence of climatic variables on phenology from other confounding factors like photoperiod, light intensity and relative humidity (Chuine et al. 2010; Ko¨rner and Basler 2010).

To overcome the limitations of observational approaches, an experimental approach by employing controlled conditions can help to tease apart the confounding effects of different climatic and biotic variables (Primack et al. 2015). Although a number of experimental phenology studies are available from the developed world (Ettinger et al. 2020), such studies are still insufficient from the developing world (Stuble et al. 2021). Very few studies reporting phenological shifts by climate warming using observational and historical data from the developing world have recently become available (Hart et al. 2014; Gaira et al. 2014; Ahmad et al. 2021; Hassan et al. 2021). As the developing world harbours a much higher proportion of the global biodiversity, conducting experimental phenology studies assumes a research priority.

Amongst the various experimental approaches employed in phenology research, winter dormant twigs of woody plants placed in the temperature-regulated growth chambers have been successfully used to study the influence of various temperature treatments on plant phenological events (Primack et al. 2015). The procedure is relatively simple and easily manageable, requires limited space and can be accomplished in growth chambers where factors like temperature, photoperiod and humidity can be easily controlled. Thus, this method has been employed to investigate the impact of climate warming on several different phenological responses (Primack et al. 2015; Jung et al. 2021). Several studies suggest twigs as effective model systems for depicting the response of whole plants (Polgar et al. 2014; Sønsteby and Heide 2014; Vitasse and Basler 2014), and they have been used in many recently published phenological studies (e.g. Basler and Körner 2012; Dantec et al. 2014; Menzel et al. 2020; Primack et al. 2015; Laube et al. 2014a, b).

Generally, the phenological response in a particular region depends upon the regional climate and is expected to be species-specific (Gillooly et al. 2001), as the sensitivity to seasonal warming differs amongst species (Fitter and Fitter 2002; Sherry et al. 2007; Zohner et al. 2017, 2016; Laube et al. 2014a, b). The closely related species show different phenological responses to climate warming, with certain species responding much faster than the others (Miller-Rushing and Inouye 2009). These differential phenological response patterns that exists amongst the related species can be used to make predictions about the response of plant communities to projected climate change scenarios. Therefore, specific case studies from the data-deficit regions are required to better predict the phenological response of congeners to future climate warming. Furthermore, the differential phenological responses shown by different species to climate warming could result in changed patterns of coexistence amongst species during reproduction (Sherry et al. 2007). As a result, the species interaction within communities could get disturbed, thereby changing the associated community composition (Waser and Real 1979; Chuine and Beaubien 2001; Post et al. 2001). Therefore, there is a need for research studies to be conducted at the regional scale taking into account the species-specific phenological responses (Fox and Jönsson 2019).

The Himalaya, a global biodiversity hotspot region, has experienced rapid climate warming during recent times (Romshoo et al. 2015; Murtaza and Romshoo 2016). Recently, a few observation-based studies on the plant phenology in different parts of the Himalaya have been conducted (Hart et al. 2014; Basnett et al. 2019; Gaira et al. 2014; Ahmad et al. 2021; Hassan et al. 2021). The Kashmir Himalaya, a region located in the Himalaya global biodiversity hotspot, is recently experiencing rapid climate changes (Hassan et al. 2021), and therefore experimental phenological study from this region can help in filling the global geographical knowledge gap in phenology research. It is with this backdrop that our study attempts to determine the effects of different temperature treatments and species identity on the patterns of phenological shifts in Kashmir Himalaya.

Until now, most of the experimental studies employing the twig method have studied the influence of single variable like effect of temperature on the phenological events of single plant species (Primack et al. 2015). Experimental studies investigating the simultaneous effect of temperature and species identity using the multi-twig species approach can allow to evaluate the independent effects of climate variables and also to analyse the variable phenological response of multiple species. In this study, using four pairs of congeneric woody plant species from Kashmir Himalaya, we investigated the effect of experimental warming and species identity on the three phenological variables: date of first flower out, date of senescence and duration of flowering period. More specifically, we aimed to address the following research questions: (1) Do the experimental warming treatments (different temperature conditions) determine the initiation, end and duration of flowering phenology in the selected woody species? (2) Is the magnitude of studied phenological variables (dates of first flower out, senescence and duration of flowering period) species-specific? (3) Do the congeneric pairs of species show similar or different phenological responses to different temperature treatments?

This Himalayan region has experienced pronounced changes in temperature due to climate warming over the last decades (Hassan et al. 2021) with expected impacts on plant phenology. The phenological shifts induced due to climate warming not only affect the structure of plant community but also affect the biotic interactions, plant breeding systems and reproductive success (Inouye 2008; Theobald et al. 2017; McDonough MacKenzie et al. 2018). Therefore, the findings from the present study are expected to help in better understanding the species’ responses to climate warming in the region and predicting its impacts on key ecological processes.

Materials and methods

Study area

The present study was carried out between February and April 2021 at the Kashmir University Botanical Garden (KUBG), located in the main campus of the University of Kashmir situated in Srinagar, Kashmir Himalaya—a part of the Himalaya biodiversity hotspot (32°20′ to 34°54′ N latitudes and 73°55′ to 75°35′ E longitudes, altitude 1600 m asl) (Husain 2002; Dar and Khuroo 2020). A rich repository of plant species, either cultivated or growing in the wild, occur in the KUBG which are typical representatives of the flora of this Himalayan region. Generally, the climate in the region is of continental temperate type with four distinct seasons: winter (December–February), spring (March–May), summer (June–August) and autumn (September–November). The average temperature ranges from 31 °C maximum to 15 °C minimum in summer and from 4 °C maximum to − 4 °C minimum in winters. In recent times, however, the region has experienced increasing trends of climate warming with an increase in annual temperature of about 0.65℃/year, and decrease in annual precipitation of − 1.24 mm/year (Hassan et al. 2021; Romshoo et al. 2015; Murtaza and Romshoo 2016; Dad et al. 2021; Zaz et al. 2019). In particular, over the last decade, the region is experiencing relatively warmer months of February and early onset of spring flowering (Hassan et al. 2021). In recent times, such early flowering in the horticultural crops (e.g. apple, pear, peach) has caused huge economic losses due to freakish snowfall events post-flowering (Rashid et al. 2020; Wani et al. 2021).

Selected plant species

For the present study, we selected four pairs of congeneric woody species: Ulmus wallichiana, U. villosa, Prunus tomentosa, P. persica, Populus alba, P. deltoides, Viburnum cotinifolium and V. opulus growing in KUBG (Fig. 1). These species were selected because they are typical representatives of first flowering plants during the start of spring season in Kashmir Himalaya after a long winter (Malik et al. 2020).

Fig. 1
figure 1

A pictorial overview of the selected woody species in Kashmir Himalaya with the experimental setup showing twigs in the growth chambers maintained at different temperature regimes of 23℃, 20℃, 17℃ and 9℃ (control). A Prunus persica. B Prunus tomentosa. C Ulmus villosa. D Ulmus wallichiana. E Viburnum cotinifolium. F Viburnum opulus. G Populus deltoides. H Populus alba

Full size image

Experimental design

We adopted the twig experimental setup under different temperature regimes in growth chambers (Primack et al. 2015). Three plant growth chambers maintained at different temperature regimes of T2 (17℃), T3 (20℃) and T4 (23℃) were used (Fig. 1), with a photoperiod of 14.5 h and dark period of 9.5 h (Miller-Rushing and Primack 2008a) average of February to April months in the study region, with constant relative humidity of 65% (Ruiz et al. 2007; Campoy et al. 2012). The growth chambers were maintained at alternative day/night temperatures in order to provide the twigs natural-like conditions. The light intensity was kept around 55 mol m−2 s−1 (Ruiz et al. 2007; Campoy et al. 2012). These temperature regimes were selected as the range of average temperature in spring season during which these plant species flower lies between minimum of 17℃ to maximum of 23℃. On 12 February 2021, the twigs of the selected plant species with length ranging from 10 to 30 cm were cut obliquely with a sterilised hand pruner (Primack et al. 2015; Basler and Korner 2012), These were obtained from the single individual of particular plant species. Only the twigs bearing 2 to 5 intact reproductive buds were selected to be used for the experiment. The twigs were washed with sodium hypochlorite solution (200 ppm active chlorine) (Zohner and Renner 2015), followed again by a wash in distilled water. The washed twigs were placed in labelled 250-mL conical glass flasks with each flask containing single twig with 2–5 intact buds filled with 5% sucrose solution (Ruiz et al. 2007; Campoy et al. 2012; Sønsteby and Heide 2014). To avoid microbial infection, gentamicin sulphate (40 μg L−1) was poured into the flasks (Zohner and Renner 2015). The flasks were then transferred to the three growth chambers. The twigs of the selected species, each with three replicates, with each replicate consisting of single twig of particular plant species, were placed in each growth chamber by following completely randomised block design. In addition to the growth chambers, another set of twigs of the selected species were also kept in the pot house of KUBG to act as control. The pot house represented the natural conditions with temperature T1, around 9 ℃ (average temperature of February, March and April months during which experiment was active) (Fig. 1). After every week, the sucrose solution in the flasks was changed and the twigs were recut and placed in the fresh sucrose solution. The growth chambers and the pot house were monitored on a daily basis to record the data on three phenological variables. For the twigs of the selected species, we recorded dates of first flower out (FFO) and senescence (DOS) in terms of the day of year (DOY) and duration of flowering (DOF) period based on number of days. The FFO was taken as the day of the year when the corolla of first flowers (around 10% in each twig) fully unfolds; the DOS was taken when the corolla of flowers (around 10% in each twig) dries out; and the DOF was recorded as the number of days from FFO to DOS. The data obtained on these three phenological variables was used to analyse the role of different temperature regimes and species identity on the patterns and magnitude of phenological shifts in the selected plant species.

Data analysis

All the analyses were carried out using R statistical software version 4.0.2 (R Core Team 2020). We used a simple linear regression model to investigate the studied phenological variables, comparing date of first flower out, date of senescence and duration of flowering period with temperature. We analysed the data in two different ways, comparing (i) the sensitivities (slope) of each species individually, and (ii) each genus separately. We also subjected the data on the phenological variables to Levene’s test and the Shapiro–Wilk test to check the assumptions of homogeneity of variance and normality of distribution in the data respectively. Since the assumptions in the data were not met, we log-transformed the data of the phenological variables to meet the required assumptions. To test whether there is a significant difference in the phenological variables, we performed the two-way analysis of variance (ANOVA) test using temperature (T), species identity (SI) and their interactions as factors. In the case that the effect was significant (P ≤ 0.05), we used Tukey’s HSD (Honest Significant Differences) for performing multiple pairwise-comparison between the means of groups.

Results

The results of the two-way ANOVA showed that both the temperature and species identity had a significant effect on the studied phenological variables (Table 1), thereby showing that the species differ in their phenological response to different temperature treatments. The results of the linear regression revealed a significant negative effect of temperature on the date of first flower out, indicating that the flowering in the selected woody species starts much earlier under warmer temperature conditions (Fig. 2; Supplementary Table 1). This pattern was consistent across all the species (Fig. 2) and genera studied, though the magnitude of the effect varied greatly (Fig. 5; Supplementary Table 1). Likewise, a significant decrease in the date of senescence with varying magnitudes was observed with increasing temperature for all the species (Fig. 3; Supplementary Table 1) and genera studied (Fig. 6; Supplementary Table 1). Similarly, we found a significant negative effect of temperature on duration of flowering, indicating that the length of flowering period in the selected species decreases with increasing temperature (Fig. 4; Supplementary Table 1). This effect was seen within all individual species, where advancement in the dates of senescence was greater than the advancement in dates of first flower out, resulting in a shorter duration of flowering (Supplementary Table 1). Once again, the observed effect was consistent at the genus level as well, though the magnitude of the effect varied considerably (Fig. 7; Supplementary Table 1). For all the studied phenological variables, our results revealed that the species’ sensitivities are more similar within congeneric pairs than the distantly related species (Fig. 2, 3 and 4).

Table 1 Summarized results of the two-way ANOVA test. Shown are the associated significance levels along with mean (and standard deviation). ***P ≤ 0.001
Full size table
Fig. 2
figure 2

Relationships between the temperature and first flower out (DOY) of the studied species. The solid lines represent the best fitting regression lines while the shaded areas represent the 95% confidence interval of the regression lines

Full size image
Fig. 3
figure 3

Relationships between the temperature and date of senescence (DOY) of the studied species. The solid lines represent the best fitting regression lines while the shaded areas represent the 95% confidence interval of the regression lines

Full size image
Fig. 4
figure 4

Relationships between the temperature and duration of flowering (no. of days) of the studied species. The solid lines represent the best fitting regression line while the shaded areas represent the 95% confidence interval of the regression line

Full size image
Fig. 5
figure 5

Relationships between the temperature and date of first flower out (DOY) in four pairs of selected congeneric woody species. The black solid line represents the best fitting regression line at the genus level, the coloured dashed lines the best fitting regression lines for the two species in the particular genus and the grey shaded areas the 95% confidence interval of the regression lines

Full size image
Fig. 6
figure 6

Relationships between the temperature and date of senescence (DOY) in four pairs of selected congeneric woody species. The black solid line represents the best fitting regression line at the genus level, the coloured dashed lines the best fitting regression lines for the two species in the particular genus and the grey shaded areas the 95% confidence interval of the regression lines

Full size image
Fig. 7
figure 7

Relationships between the temperature and duration of flowering (no. of days) in four pairs of selected congeneric woody species. The black solid line represents the best fitting regression line at the genus level, the coloured dashed lines the best fitting regression lines for the two species in the particular genus and the grey shaded areas the 95% confidence interval of the regression lines

Full size image

The post hoc comparison revealed that the observed differences in the phenological variables were due to the significant differences between all the possible pairwise temperature comparisons; the one exception was for the T3 vs T4 comparison for duration of flowering phase; in this case, the observed difference in duration of flowering was non-significant (Supplementary Table 2). The post hoc test of species identity showed that observed difference in date of first flower out was significant for all the cases except for Ulmus villosa vs Populus alba, Prunus persica vs Populus deltoides and Viburnum opulus vs Ulmus wallichiana pairs (Supplementary Table 3). Similarly, the post hoc test of species identity showed that the observed difference in date of senescence was significant for all the cases except for Ulmus villosa vs Populus alba pair (Supplementary Table 4). In contrast, the post hoc test of species identity showed that observed difference in duration of flowering phase was significant for a few pairwise cases only (Supplementary Table 5).

Discussion

The present study, to the best of our knowledge, is the first attempt from the developing world to experimentally investigate the effect of warming temperature and species identity on the phenological events of woody species in Kashmir Himalaya.

Role of temperature in flowering phenology

The timing of various spring phenological events is governed by changes in environmental conditions particularly changes in temperature (Post and Stenseth 1999; Elmendorf et al. 2016; Luna 2016; Tao et al. 2017). Our results clearly show the decreasing effect of temperature on the date of first flower out, date of senescence and duration of flowering for all the selected woody species. We observed that the first flower out of all the species occurs earlier under warmer conditions. A likely reason for this effect is that growing degree days (GDD) accumulate faster with warmer temperatures, thereby allowing plants to flower earlier once the threshold has been met (Piao et al. 2015; Suonan et al. 2017). Our results showing advances in flowering phenology are in agreement with a large number of studies conducted in temperate ecosystems around the world (Neil and Wu 2006; Cleland et al. 2007; Stuble et al. 2021). Increase in temperature has reported to significantly advance first flowering dates in 28 woody species in North America (Flynn and Wolkovich 2018), Salix × smithiana (Caffarra and Donnelly 2011) and Sternbergia vernalis in Kashmir Himalaya (Hassan et al. 2021). Our results are further supported by several other studies (Sherry et al. 2007; Yu et al. 2010; Lee 2011; Prevéy et al. 2019), which have reported the advancement of phenological events under warming conditions. Warming not only advances first flower out dates, but also advances date of senescence by accelerating protein degradation and affecting cell division and cell expansion (Furguson et al. 1990; Scheurwater et al. 2000). Furthermore, higher temperature promotes sugar accumulation in tissues, which also contributes to early senescence (Haba et al. 2014).

Interestingly, in the present study, we found that the duration of flowering phase in all the species shows a corresponding decrease along the gradient of increasing temperature. This shortening of flowering phase under warmer temperature is largely because of early flower senescence due to increased temperature (Kehrberger and Holzschuh 2019). Under warmer conditions, the advancement in the dates of senescence was greater than the advancement in dates of first flower out for all the species, thereby resulting in shorter duration of flowering (Supplementary Table 1). These findings reveal that the flower senescence phenophase in the studied species is more sensitive to climate warming, which is in agreement with findings of some recent studies (Nagahama et al. 2018; Kehrberger and Holzschuh 2019). This climate warming–induced decrease in flower duration can potentially have a negative effect on plant reproductive success by decreasing the probability of pollinators visiting the plant species (Kehrberger and Holzschuh 2019; CaraDonna and Waser 2020). Such phenological constraints on reproductive potential in plant communities can alter the structure and stability of ecological communities, which could have long-term consequences for provision of ecosystem goods and services (Liu et al. 2011).

Species identity, congeners and flowering phenology response

The phenological response to a particular temperature varies between different species and thus is species-specific in magnitude (Rice et al. 2021). This interspecific variation in phenological response to different temperature treatments is evident, as temperature is one of the prime determinants of the plant phenology (Hassan et al. 2021; Rice et al. 2021). In the present study, the flowering phenology response of all the woody species to a particular temperature regime showed shifts independent of one another, which can be ascribed to the species-specific adaptive response to a particular temperature cue. Therefore, a relatively warmer temperature to a plant species than the normal conditions brings about the specific phenological shifts, as supported by majority of studies on species’ phenological shifts due to climate warming (Fitter and Fitter 2002; Lehikoinen et al. 2004; Sherry et al. 2007; Miller-Rushing and Primack 2008a, b). Previously, the experimental phenological studies have shown the response of a single species to the climate warming (Matsumoto et al. 2003; Doi 2008). Our study makes advance by investigating the variation in the phenological response amongst multiple sets of species to experimental warming, which is consistent with the findings of other studies (Rice et al. 2021; CaraDonna et al. 2014; Primack et al. 2009). The response of individual species to climate warming is crucial for understanding the influence of global warming on its interactions within population, communities and ultimately on ecosystems and the biodiversity they support (Primack et al. 2009). Increased temperature has the potential to cause major changes in biological communities by altering the biotic relationships amongst the species (Kharouba et al. 2018; Forrest and Miller-Rushing 2010), affecting species synchrony, co-flowering patterns and flowering duration, which can ultimately shift the diversity and distribution of biodiversity. Furthermore, the observed species-specific response to climatic conditions particularly temperature can be used to predict the effects of future climate warming on flowering phenology, especially in temperate communities where temperature during the early spring is showing warming trends (Gherardi and Sala 2015; Chen et al. 2016).

Furthermore, our study showed a significant difference in the phenological response between the species of all the genera studied. More specifically, a significant difference in the first flower out, date of senescence and duration of flowering was observed for majority of the congeneric pairwise species. We found that species of the same genus showed variable phenological shifts to same temperature treatment. The probable reason for the variable phenological response can be that each plant species requires an optimum and specific temperature cue for a particular phenological event to occur (Rice et al. 2021). Our findings reveal that the species sensitivities are more similar within congeneric pairs than the distantly related species. This can be attributed to the phylogenetically conserved traits amongst congeneric species that can account for a significant amount of variation in phenological shifts amongst the studied genera (Buckley and Kingsolver 2012). Our results are also in agreement with the recent studies that reported the congeners differ in their phenological response to particular temperature (Primack et al. 2009; Gerst et al. 2017). Due to rapidly changing climate in the Himalayan region, the shifts in phenological patterns may lead to mismatch in the plant-pollinator interactions, thus affecting the biological communities in general and synchrony amongst insect and plant species in particular (CaraDonna et al. 2014).

Conclusion

A shift to earlier spring phenological phases is a well-known response of plant species to the climate warming. The present study investigated the effect of warmer temperatures on the flowering phenology of selected woody species of Kashmir Himalaya. We experimentally showed that, with an increase in temperature, there is an advancement of dates of first flower out and senescence for all the selected species, and a decrease in duration of flowering phase. These results provide important insights about the likely impacts of phenological shifts on the pollination behaviour and other biological interactions of plant species due to ongoing climate warming, which merit detailed investigation in near future. Since the studied plant species are reported to be mostly bee-pollinated (Ara et al. 2019), an advancement in flowering time can affect the plant-pollinator interaction due to early spring warming, thereby influencing their reproductive success. In addition, we found that the different species (including congeneric pairs) significantly differ in their phenological response to particular temperature treatments, indicating that these relationships are highly species-specific. We conclude that the warmer temperature plays a significant role in advancement of first flower out and flower senescence in the studied spring-flowering woody species of Kashmir Himalaya, concomitantly influencing the duration of flowering phase as well. Overall, this study furthers our understanding of the key drivers underlying phenological shifts in this data-deficient Himalayan biodiversity hotspot region, with wider implications for similar regions elsewhere in the world.

Lastly, our results should be taken with the caveat that we have specifically attempted to determine the effect of increasing temperature on the flower senescence by comparing the twigs placed in similar conditions, not by comparing the twigs with the whole plants in nature. We would like to caution in directly extrapolating the results of our study to the field conditions, which therefore leaves much scope for field validation, and variation if any, in future phenological studies.