4.1 Plasticity of bud phenology
The significant effect of site on bud phenology indicated a significant phenotypic plasticity of leaf development for sugar maple. As sessile organisms, plants benefit from a high phenological plasticity to respond quickly to environmental changes, matching the optimal moment of the year for flushing, and avoiding the negative consequences of early or late growth reactivations (Allevato et al. 2019). Individuals with high phenological plasticity in bud phenology are favored during the evolutionary process (Vitasse et al. 2010). This phenological plasticity of sugar maple may be an important factor that enables a broad tolerance of the species to various environmental conditions, and probably contributes to the wide latitudinal range of the species across eastern North America.
The interaction between site and seed origin at the beginning and ending of leaf development indicated differences in phenological plasticity among sugar maple populations. These different responses may be related to contrasting adaptation strategies among populations within the same species, and involving the trade-off between lengthening the growing season and avoiding frost damages. In six temperate deciduous species, the phenological plasticity was lower in populations originating from higher elevations (Vitasse et al. 2013). A lower plasticity limits the sensitivity to warmer spring temperatures, thus delaying bud burst and ensuring the avoidance of frost. A high plasticity is expensive for plants, and requires resources for the continuous monitoring of the environment through the synthesis or use of chemical substances (Van Buskirk and Steiner 2009). For Pinus contorta and Pinus monticola growing under stressing conditions, more resources were invested to resist drought and frost, thus, a reduced plasticity was observed for growth (Chuine et al. 2006). This different magnitude of plasticity among populations plays an important role in matching environmental changes and ensuring sustained growth in the long term.
4.2 Comparison of bud phenology in two common gardens
Buds of sugar maple reactivated earlier in the southern plantation (Ripon). In our study, the mean spring temperature (April-May) in 2020 in Ripon was 4.6 ℃, which was higher than that in Chicoutimi (2.4 ℃). Warmer spring conditions in Ripon speed up the achievement of forcing temperatures, thus advancing bud burst. Our results are consistent with previous studies conducted in temperate and boreal ecosystems. The timings of leaf unfolding of Fagus sylvatica L. and Quercus petraea (Matt.) Liebl. advanced at a rate of 5.7 days per additional degree Celsius along five common gardens (Vitasse et al. 2010). Similarly, buds of Populus fremontii in Arizona flush earlier in trees growing in the common gardens of warmer regions (Cooper et al. 2019).
According to the 19 bioclimatic variables, the climatic distance between the two common gardens could be considered small when compared to the broad geographical gradient of the seed origins used in this study. However, we detected a difference of 12 days for the beginning of leaf development between Ripon and Chicoutimi. The two common gardens are located near or at the border of the northern distribution of sugar maple. Individuals growing at the boundary of species distribution are more sensitive to the environment (Jump et al. 2006; Normand et al. 2009). For example, the effects of winter temperature on the radial growth of European beech (Fagus sylvatica L.) were significant only at the colder part of the distribution (Weigel et al. 2018). Similarly, European beech was more sensitive to drought at the drier boundary of the range (Jump et al. 2006; Roibu et al. 2017). In our study, the difference in spring temperature between the two common gardens reached 2.2 ℃, which explained the observed delay of bud phenology in Chicoutimi.
The time gap between Ripon and Chicoutimi decreased for the later leafing stages (7 and 8). The difference in photoperiod is an important factor for spring phenology. Day length during bud burst (phase 1) was 15.0 h in Chicoutimi, which was longer than that recorded in Ripon (14.1 h) due to the different latitudes and timings of growth reactivation between common gardens. A longer photoperiod could have speeded up bud development, thus resulting in similar timings of full leaf expansion. A previous study demonstrated that sugar maple leafing benefits from a longer day length, and photoperiod can outweigh the delaying effects of colder springs (Ren et al. 2020). However, the similar ending of leaf expansion in the two common gardens remains partially unexplained, and could result from the weather events occurring during the studied year. Thus, a better understanding of the impact of current weather on leaf development is needed and requires a long-term monitoring of bud phenology in the two common gardens.
4.3 Plasticity vs local adaptation
Our study revealed a higher contribution of plasticity to variance compared with local adaptation, mainly at the beginning of leaf development. Similar results were also observed in other species of boreal and temperate ecosystems (Baliuckas and Pliura 2003; Vitasse et al. 2010). The contribution of plasticity to bud burst phenology ranged from 55 to 86% in seven deciduous species in Europe, one order of magnitude higher than that of local adaptation (0.3-9%) (Vitasse et al. 2013). In another study combining experiments in situ and common garden, the genetic differentiation explained <28% of variance in the morphological and physiological traits of leaves of sessile oak and European beech, suggesting a minor effect of local adaptation on leaf functional traits (Bresson et al. 2011).
Local adaptation and phenotypic plasticity act as concurring processes in the response of plants to changes in the environment, playing a different role at spatial and temporal scales. Under stable local environments, well adapted populations could maintain competitive fitness to survive. However, in a context of rapidly changing conditions, which requires fast response mechanisms (Van Kleunen and Fischer 2005), phenotypic plasticity can be favoured. Species with wide distributions, especially under continental climates, can experience a wide inter-annual variability in environmental conditions, principally temperature, one of the main limiting factors for bud phenology in spring. Individuals with high plasticity can respond to weather events quickly and gain a comparably longer period for carbon fixation and high competitive abilities (Kramer 1995). In addition, because of the long lifespan and slow and intermittent regeneration periods (masting years) of trees, individuals could rely predominantly on phenotypic plasticity for survival and growth rather than adaptation (Fox et al. 2019).
It has been predicted that the mid-latitudes of North America will have experienced warming up to 7 °C at the end of the 21st century (Feng et al. 2014). On the one hand, plastic species may benefit from these changes by lengthening the growing season and increasing fitness. A previous study demonstrated that the recent advancement of 13 flowering days under global warming has helped plastic trees to improve their fitness by 40% (Anderson et al. 2012). On the other hand, an earlier bud burst seriously increases the risk of frost damage to the young developing tissues and leaves (Howe et al. 2003). Under the ongoing global changes, the climatic variability increases in magnitude, resulting in more frequent extreme weather events such heat waves (Hegerl et al. 2011; Min et al. 2011) or cooling (Wang et al. 2011). These extreme events have proved to be detrimental for tree growth and survival in the short term. Phenotypic plasticity is an important functional trait to deal with the increased uncertainty of climate in the future (Harmon et al. 2009; Donohue et al. 2013).
In this study, we observed a high contribution of residuals to the variance in bud phenology, which is in agreement with previous studies (Sole-Medina et al. 2020; Varsamis et al. 2018). This large heterogeneity in phenology unexplained by our factors may suggest a high variability among individuals (Perrin et al. 2017), in addition to the potential effect of microsite conditions and sampling errors during field observations. An important effect of microsite on plant phenology seems unlikely, because the two common gardens are located in cropland areas, which were submitted to crop production in the past, and are therefore expected to be more homogeneous than natural sites. The wide variation in bud phenology within the same population may represent a diverse gene reservoir for the long-term survival of the species (Rousi and Heinonen 2007). These various genotypes ensure a potential matching between some individuals of the populations and the environmental conditions, thus allowing local persistence of the species.