Effect of climate change on bud phenology of young aspen plants (Populus tremula. L)

Abstract Boreal tree species are excellent tools for studying tolerance to climate change. Bud phenology is a trait, which is highly sensitive to environmental fluctuations and thus useful for climate change investigations. However, experimental studies of bud phenology under simulated climate change outdoors are deficient. We conducted a multifactorial field experiment with single (T, UVA, UVB) and combined treatments (UVA+T, UVB+T) of elevated temperature (T, +2°C) and ultraviolet‐B radiation (+30% UVB) in order to examine their impact on both male and female genotypes of aspen (Populus tremula L.). This study focuses on the effect of the treatments in years 2 and 3 after planting (2013, 2014) and follows how bud phenology is adapting in year 4 (2015), when the treatments were discontinued. Moreover, the effect of bud removal was recorded. We found that elevated temperature played a key role in delaying bud set and forcing bud break in intact individuals, as well as slightly delaying bud break in bud‐removed individuals. UVB delayed the bud break in bud‐removed males. In addition, both UVA and UVB interacted with temperature in year 3 and even in year 4, when the treatments were off, but only in male individuals. Axillary bud removal forced both bud break and bud set under combined treatments (UVA+T, UVB+T) and delayed both under individual treatments (T, UVB). In conclusion, male aspens were more responsive to the treatments than females and that effect of elevated temperature and UV radiation on bud set and bud break of aspen is not disappearing over 4‐year study period.

generally finely tuned to the seasonality of their environment. The autumn phenology of trees was traditionally thought to be mainly controlled by day length (Wareing, 1956), and spring phenology by temperature (Zohner, Benito, Svenning, & Renner, 2016). Recently conducted studies, however, have shown that rising temperatures influences the flushing dates of northern tree species (e.g., Chung et al., 2013;Ibáñez et al., 2010;Menzel, 2000). There are also signs of autumn phenology being affected by increase in temperature (Westergaard & Eriksen 1997, Kalcsits, Salim, & Tanino, 2009Tanino, Kalcsits, Silim, Kendall, & Gray, 2010;Rohde, Bastien, & Boerjan, 2011;Hänninen & Tanino, 2011;Way, 2011). The commencement of autumn leaf senescence and dormancy is based on a combination of numerous developmental and environmental signals (Cooke, Eriksson, & Junttila, 2012). Climate warming can accelerate the growth cessation in many tree species, cultivars, and ecotypes (Tanino et al., 2010). In some cases, the growth cessation occurs as a combination of low night temperatures and photoperiod. However, in some northern ecotypes of Picea abies, Salix pentandra, and Betula pubescens whose dormancy induction is insensitive to photoperiod, the autumn growth cessation is delayed due to the temperature increase resulting in longer growing seasons (Hänninen & Tanino, 2011).
The effects of climate factors on phenology have mostly been studied in growth chambers or in common gardens (e.g., Beuker, 1994;Burley, 1966;Hannerz, 1994;Hänninen, 1990;Hurme, Repo, Savolainen, & Pääkkönen, 1997;Leinonen, 1996;Myking & Heide, 1995;Oleksyn, Tjoelker, & Reich, 1998) while there are few outdoor studies with a specific experimental setup. The growth conditions used in almost all experiments indoors are temporarily and spatially less variable than those in natural environments (Frenkel, Jankanpaa, Moen, & Jansson, 2008). The temperatures are usually constant or at the best varying between diurnal values, while the light intensity is generally lower and with a spectral composition differing quite significantly from that of the sun (Leonidopoulos, 2000;Young, McMahon, Rajapakse, & Decoteau, 1994). On the other hand, while in common garden experiments, phenology can be studied in natural light and temperature conditions, the local versus foreign effect cannot be tested rigorously (Allendorf & Lundquist, 2003). Thus, field experiments with experimentally modulated light and temperature supplementation are needed in order to gain a better understanding of the effects of climate factors on phenology.
Growth and dormancy cycles in plants are also controlled by light quality (Olsen & Lee, 2011). Blue, red, and far-red light influences the growth, dormancy, and bud formation of many plant species (e.g., Campbell et al., 2008;Mølmann, Junttila, Johnsen, & Olsen, 2006;Olsen, 2010). The effects of the UVA and UVB regions of the solar spectrum, however, have not been widely investigated. Most of the research on UVB and plants has concentrated on the effect of UVB as a stress factor. However, there is increasing evidence that UVB may also function as an environmental signal (Jenkins, 2009;Zlatev, Lidon, & Kaimakanova, 2012). Strømme et al. (2015) found that UVB accelerates bud set and bud break in European aspen (Populus tremula. L) plants after one growing season, while Sivadasan, Randriamanana, Julkunen-Tiitto, and Nybakken (2015) found variations in secondary metabolites and bud size in male and female Salix myrsinifolia buds under UV radiation.
Populus tremula is a diecious deciduous tree that is widely distributed throughout Europe and Asia. Populus species are of great ecological importance, as a large number of organisms, including several endangered species, are found in association with these trees (e.g., Lindroth, 2008). They are, therefore, monitored in several national and international phenology networks. Populus species are widely used as model organisms among woody plants in experimental botany (e.g., Bradshaw, Ceulemans, Davis, & Stettler, 2000;DiFazio, Slavov, & Joshi, 2011). Numerous studies on climate change are conducted with different Populus species as its circumboreal range largely overlaps with areas where drastic climate change is predicted to occur (IPCC 2014). Populus species had been found to show sex-specific responses under different climatic stress factors (Li et al., 2014;Randriamanana et al., 2014). For example, Strømme et al. (2015) have shown that in young P. tremula plantlets, elevated temperature delayed bud set and forced bud break in one growing season old seedlings. However, in their studies, under combined UVB+T treatment, bud set was forced in both males and females while bud break was delayed only in males.
Moreover, at low temperatures, females of P. cathayana showed earlier growth cessation and more chilling injuries in the chloroplast ultrastructure, cellular membranes, and leaf morphology compared to males (Zhang, Jiang, Peng, Korpelainen, & Li, 2011). Also, variations between sexes as to the magnitude of morphological, physiological, and biochemical traits have been documented under drought and elevated temperature in P. cathayana (Xu, Peng, Wu, Korpelainen, & Li, 2008). Males of P. cathayana having higher basal diameter, leaf nitrogen, and lower concentration of abscisic acid and UV-absorbing compounds and exhibited greater resistance under enhanced UVB than did females (Xu et al., 2010).
This study used the same experimental setup as Strømme et al. (2015), except that we investigated the subsequent 2-year effects of climate change on P. tremula bud phenology, and the potential carryover effects for 1 year after the treatments were discontinued. In addition, the effect of axillary bud removal on the timing of bud set and bud break was tested. We hypothesized that (i) enhanced temperature will delay bud set and force bud break, while UVB will force bud set, and the responses will be mitigated by the experimental years due to acclimation, (ii) the effects of enhanced temperature on bud break will be sustained over the following season, even when the treatments are discontinued, (ii) bud removal will change the growing period due to resource restrictions, and (iv) males and females vary in their responses to the treatments.

| Plant materials
The aspen plantlets used in this experiment originated from Eastern and Southern Finland, as presented in Randriamanana, Nissinen, Moilanen, Nybakken, and Julkunen-Tiitto (2015). In 2012, when the field experiment started, they were micropropagated from buds of six male and six female aspen trees, about 30-40 years old. Each genotype was collected from the following locations: Kaavi 62°43′N, and planted in soil on 11 June. We followed the same individuals for bud phenology as in Strømme et al. (2015). Some mortality occurred during the study period occurred due to Venturia shoot blight, and also as a result of some mechanical and herbivore damages. All the plants in the experimental field were scored for bud break and bud set. As a consequence of Venturia infections on the apical meristems, numerous buds remained dead or fell off during scoring, and only the plants that

| Experimental setup
The experimental setup included 36 plots in a 6 × 6 matrix, as explained in detail by Nybakken, Hörkkä, and Julkunen-Tiitto (2012). The plants within each plot received one of the following six treatments or treatment combinations: enhanced temperature (T), enhanced ultraviolet-B radiation (UVB), ultraviolet-A radiation (UVA), UVB+T, UVA+T, and control with ambient temperature and UV radiation (C). The enhanced levels of T and UVB were continuously regulated to increases of +2°C and 30%, respectively. A 10 cm layer of 0.8% limed mineral soil was added to each plot. A distance of 3 m was kept between the plots in all directions, and adjustable aluminum frames (1.5 × 2.0 m) holding the lamps and heaters above the plots were bolted to metallic posts.
A metal net fence of 1.5 m was constructed around the experimental F I G U R E 1 The bud break stages used for scoring the spring phenology were (a-e), a (0)-a closed brown bud, b (1)closed bud with protruding green leaf tips, c (2)-green leaf tips out of the bud with leaf bases hidden, d (3)-broken bud with at least one petiole and e (4)-unfolded leaf with visible leaf blade and stalk. The bud set stages used for autumn phenology scoring were f-h, f (1)-apices between full, active growth to apices with an open bud, g (0.5)-a closed green bud, and h (0)-a closed brown bud field in order to prevent the intrusion of large mammals, and a protective metal sheet was implanted 60 cm into the soil, reaching 60 cm above the soil level, to prevent vole intrusion.
To each aluminum frame, six 40 W UV fluorescent lamps (1.2 m long, UVB-313, Q-Panel Co., Cleveland, OH, USA) were appended, following a cosine distribution (Björn, 1990 plots, IR radiators were replaced with wooden boards in order to attain the same shading levels. The filters were changed every 3 weeks, and the frames were lifted every third week in order to maintain a 60 cm distance between the highest shoot tip and the radiators/UV lamps. Four Thies Clima sensors (Thies, Göttingen, Germany) were used for measuring UVB radiation. These sensors measured the radiation between 250 and 325 nm, with a peak of 300 nm. Two sensors were placed above the control frames for ambient UVB levels, and two under the frames of UVB enhancement plots for set-point values.

| Axillary bud removal
In summer 2014 (2nd July) (day 183), three axillary buds from four lateral shoots were removed from one individual of every clone in each plot (total 410 individuals). In autumn 2014 (20th October) (day 293), another three buds were removed from the same individual in order to see whether the bud removal had any effect on bud set and bud break for the coming growing season.

| Scoring the bud set and bud break stages
In 2013

| Statistical analysis
The effects on bud break (

| Bud set in 2013 and 2014
During autumn 2013, elevated temperature delayed bud set in both males and females (Figure 2a). In autumn 2014, UVA treatment delayed bud set in males, but the combination of UVA and temperature (UVA+T) forced the buds to set earlier in males than in females (Table 1, Figure 2b). In the same year, removal of axillary buds resulted in delayed bud set in males under elevated temperature, relative to females (Figure 2c). However, after bud removal, the combined treatment UVB+T forced the buds to set earlier in males compared to females. The combination treatment UVA+T also forced the bud break in males, to a smaller extent (Table 1, Figure 2c).

| Bud break in 2014 and 2015
Bud break was forced under temperature treatments in both males and females in 2014 by 2 days (Figure 3a) and to a smaller degree in spring 2015 (Figure 3b). The significant negative coefficient of males showed that the female bud break was earlier than males, independent of treatments (Table 2). In 2015, when buds had been removed, elevated T and UVB slightly delayed bud break in males when compared to female clones (Table 2). In contrast, the combination of temperature and UVA (UVA+T) enhanced bud break in males compared to female clones. (Table 2, Figure 3c).

| DISCUSSION
In this study, we followed, to our knowledge, for the first time, the effects of temperature and UV on bud set and bud break in woody plant seedlings for three subsequent years. In line with our first hypothesis, and similar to the first year (Strømme et al., 2015), bud set was delayed by elevated temperature during the second year (autumn 2013).
This confirms that temperature modifies sensitivity to day length signals at growth cessation and can influence the duration of bud formation in P. tremula, as seen earlier both in another outdoor study carried out at different field sites for two seasons with hybrids from P. nigra, P. trichocarpa, and P. deltoides (Rohde et al., 2011), and in a study conducted in a controlled environment with hybrids from P. nigra, P. petrowskyana, and P. deltoides (Kalcsits et al., 2009). Likewise, bud removal delayed bud set in males when compared to females. Bud removal can result in an overall reduction in sugar levels; sugars having a cross-talk with the pathways of phytohormones also causing imbalances between them (Eyles et al., 2013;Gibson, 2005;Little & Wareing, 1981). It is also found that sugars do play an important role in controlling bud dormancy by influencing the phytohormones (Anderson, Chao, & Horvath, 2001;Horvath, Anderson, Chao, & Foley, 2003). High abscisic acid concentration is also associated with bud dormancy (Rinne, Tuominen, & Junttila, 1994), and the removal of buds could lower the concentration of this hormone, leading to a delayed bud set in males under elevated temperature. The bud-removed individuals in our experimental field showed increased height growth (Sobuj et al. unpublished data), and it has also been found that the increased height growth is genetically associated with delayed bud set in P. balsamifera (Riemenschneider, McMahon, & Ostry, 1992). Earlier studies show that spring events, such as leaf unfolding or needle flush, are particularly sensitive to temperature (Lechowicz, 1995;Sarvas, 1972Sarvas, , 1974. In accordance with our results on the after effect of the treatments in 2015, Fu et al. (2012) found that Betula  (Pagter, Andersen, & Andersen, 2015), and bud break is associated with low levels of soluble sugars (Lipavská, Svobodová, & Albrechtová, 2001). In this case, the enhanced temperature would have reduced the sugar concentrations, as was also seen in B. pendula seedlings (Riikonen et al., 2013), which could be the mechanism behind the temperature-forced bud break in 2014 and 2015.
Contrary to our hypothesis, UVB did not affect bud break and bud set during the treatment period. In the first growing season (2012) (Jansen et al., 1996). Responsiveness to UV dose gradually decreases in leaves as plants age (Kakani, Reddy, Zhao, & Gao, 2004;Klem et al., 2012;Urban, Tuma, Holub, & Marek, 2006), which can also be one reason for the disappearance of the UVB effect after the first year in our experiment.
After removal of axillary buds, males previously exposed to elevated temperature set their buds later than females. However, within plants previously exposed to UVB+T, only females had delayed bud set (autumn 2014) and also had earlier bud break (spring 2015) than males. This gender differences in bud phenology in response to UVB and temperature treatments are difficult to explain. In budremoved individuals, UVB might have also caused some fluctuations in the carbohydrate levels during the bud break causing a delay as explained by Lindroth, Hofman, Campbell, McNabb, and Hunt (2000) and Quaggiotti, Trentin, Dalla, and Ghisi (2004 According to our hypothesis, the bud phenology of male and female aspens differed in their responses to environmental changes. The males of some species of Populus are more growth-oriented than females (Lloyd & Webb, 1977). In a greenhouse experiment, the males of P. tremula were taller had higher shoot biomass and greater leaf area when compared to females (Randriamanana et al., 2014). This may mean that male bud development and entry into the vegetative phase may be faster than in females. In P. tomentosa, An et al. (2011) found that during the time of floral budding, the male buds normally progress and senesce earlier than female floral buds. This is due to the female requirement for more resources to prepare for reproduction than that of males (Hultine, Bush, West, & Ehleringer, 2007;McDowell, McDowell, Marshall, & Hultine, 2000;Pickering & Arthur, 2003). In a study conducted with one, four-and ten-year-old Populus × canadensis aimed at checking the expression of miRNA's during vegetative phase change, it was found that the change is evident in minor changes in leaf shape and internode length . Our experimental plants changed leaf shape from triangular form to round form over the study years, demonstrating their transition from juvenile to vegetative phase. It may be that during the phase change, females show more responsive growth patterns compared to males, in order to compensate for their reproductive requirements.
To sum up, elevated temperature was influential in delaying and forcing the bud formation and development. The effect of UV-B diminished during the second growing season, but was again seen in bud-removed individuals for bud break 2015. Males were more responsive compared to females, and the removal of invested resources (axillary buds) from the plants resulted in delay and forcing of autumn and spring bud phenology in males when compared to females. As the timing of bud break and bud set represents events in survival and growth, discernment of these mechanisms and their interactions with climatic variables is a key to understand the consequences of the projected climate change for Populus forests.

ACKNOWLEDGEMENTS
We sincerely thank Mr. Norul Alam Sobuj, Mr. Shahed Saifullah, and Mr. Apu Sarvar for their help during the data collection from the experimental field. This project would have been impossible without the support of the Academy of Finland (Project number: 14918) and a spearhead project in the University of Eastern Finland.

CONFLICT OF INTEREST
None declared.

AUTHORS CONTRIBUTION
As the corresponding author Unnikrishnan Sivadasan was responsible for the study conception and design, data acquisition, analysis and interpretation the study, and writing the manuscript. As my supervisors Riitta Julkunen Tiitto and Line Nybakken designed, directed, and coordinated the study providing conceptual and technical guidance for all aspects of the study. Chenhao Cao contributed to the data acquisition. Tendry Randriamanana contributed to the statistical analysis.
Virpi Virjamo provided valuable comments in the research study. All authors contributed to the drafting of the manuscript to its final version of submission.