We established Silver birch phenology in the Pyrenees. We highlight real evapotranspiration as the main driver and the altitudinal effect within a north-south phenological pattern.
Research Article
Phenology and Stem Growth Dynamics of Betula Pendula Roth. In the Spanish Pyrenees.
https://doi.org/10.21203/rs.3.rs-2210613/v1
This work is licensed under a CC BY 4.0 License
posted
You are reading this latest preprint version
We established Silver birch phenology in the Pyrenees. We highlight real evapotranspiration as the main driver and the altitudinal effect within a north-south phenological pattern.
Silver birch
band dendrometer
dbh increments
water stress
temperature
actual evapotranspiration
We have established the buds, leaves, and cambium activity phenological calendar of Betula pendula Roth. based on detailed ground weekly records for two years (2000-2001) in the Pyrenees. Using band dendrometers, we determined the growing period of stem radial growth, and the variation of the growth rates. The cumulative radial growth was characterized with the Gompertz model. We determined the effect of climatic variables on stem growth rates in diameter. Using data from different authors, we established a latitudinal phenological gradient along Europe.
The most important cues for bud burst are heat accumulation (base temperature 5ºC) and long days that reduced the thermal time to budburst. The leaves’ longevity, the period of stem cambial activity, and the variation of the stem radial growth rates throughout the year are mainly regulated by the actual evapotranspiration. We verified that leaves yellowing and shedding to advance if the hydric stress is pronounced. The latitudinal gradient of phenological events is not lineal. Spring phenology in the Pyrenees was delayed in relation to birches in central and Northern Europe due to the altitudinal effect on the birches of the Pyrenees. The south-north gradient of the maximum growth rate in diameter is not pronounced. In the Pyrenees, the cessation of stem growth in diameter takes place on similar dates to Northern Europe and earlier than in Central Europe due to water stress. The main effect of climate change might be a shorter growing period, earlier budburst but earlier growth cessation due to water stress.
Tree species phenology research serves to understand the variations in the annual rhythms of trees growth and the relationship that might have with climate change since the timing of the phenological events can be considerably modified by warming and water stress jeopardizing growth, reproduction and the survival of trees (Chuine 2010, Xu et al. 2022). In temperate and cold zones and in high mountain areas in Southern Europe, trees go through an annual cycle in which a period of growth and a dormant state in winter alternate. This phenological cycle is synchronized with seasonal changes in photoperiod, temperature, and humidity (Zeitgebers) by means of internal biological clocks, genetic timers of phenological rhythms. Daylength is the most reliable environmental cue, but temperature and humidity also constitute environmental timing cues (Welling et al. 2004; Dodd et al. 2005; McClung 2006; Harmer 2009; Mwimba et al. 2018, Xu et al. 2022). Temperature and water availability variability exert a significant influence on plant phenology and therefore phenological changes reflect climate change with more marked effects in spring than in autumn (Estrella and Menzel 2006; Pudas et al. 2008; Juknys et al. 2012; Piao et al 2019). Climate effects also drive cambium phenology, which is highly sensitive to variations in the annual climate regime as well (Fritts 1976, Campelo et al. 2021).
Birch (Betula pendula Roth.) phenological studies are abundant in northern and Central Europe, but not in Spain, where its geographical distribution is restricted to mountainous areas and does not form extensive forests (Blanco et al. 1998). Our paper research focuses on the phenology of birch (buds, leaves, and stem growth) based on detailed ground phenological records of a population located in the Spanish central Pyrenees. As stated by Liu et al. (2020) ground reliable phenological data are of paramount importance to quantify the impact of climate change.
Birch has a free-growing pattern, growing indefinitely on long-day photoperiod. The short-day photoperiod in early autumn induces growth cessation, bud set and leaves colouring (Junttila et al. 2003, Viherä-Aarnio et al. 2005; Cooke et al. 2012, Junttila and Hänninen 2012). It has also been described that leaf senescence can be advanced by high temperatures and reduced water availability in the summer (Li et al. 2002; Juknys et al. 2012; Mariën et al. 2021; Fernandez-Martinez et al. 2016). In autumn, lower temperatures favour acclimatization to adverse winter conditions (Sakai and Larcher 1987, Welling et al 2004, Li et al. 2005). During dormancy, internal physiological factors prevent bud development even when environmental conditions are favourable (Welling et al. 2004; Singh et al. 2017; Gil et al. 2019), birch trees become insensitive to the photoperiod and environmental temperature regulates physiological and molecular processes of dormancy (Singh et al. 2017).
For many Nordic species, including birch, buds break need to be exposed to cold temperatures for a period of time (Myking and Heide 1995; Rinne et al. 1997; Cooke et al. 2012). After this chilling requirement, buds break date is considered to depend on degree-day accumulation which drives ontogenetic development and buds burst (Heide 1993; Rousi et al. 2007; Linkosalo et al. 2006; Junttila et al. 2012). Some authors have suggested that breaking winter dormancy requires a specific signal of light irradiance, length of night or possibly changes in the quality of the light spectrum to initiate safe ontogenetic shoot development so that winter rest is not broken too early (Häkkinen et al. 1998; Linkosalo et al. 2000). Thus, once the critical photoperiod has passed, the main external factor influencing bud opening appears to be the accumulation of hours of heat, and is therefore controlled by air temperature (Junttila et al. 2012) with the consequent loss progressive tolerance to frost.
There are many studies of birch phenology, but few have considered the joint phenology of the shoots and the cambium (Schmitt et al. 2004; Vitas 2011; Dox et al. 2020; Marchand et al. 2021). As has been described for deciduous pioneer species with wood diffuse porosity, among them the birch (Suzuki et al. 1996), cambium activity begins after leaves formation ends before the leaves turn yellow. On the other hand, the growth period and the growth rate in trunk increments are not constant (Schmitt et al. 2004; Vitas 2011), the accumulation of wood increments throughout the growth period follows a sigmoidal pattern (Schmitt et al. 2004; Rossi et al. 2005). The span of the growth period and the growth rate can vary greatly between individuals and from one year to another due to water stress (Breda et al. 2006; Cooke et al. 2012; Mwimba et al. 2018; Dox et al. 2020), the availability of nutrients, particularly nitrogen (Cooke et al. 2012; Gil et al. 2019) or by a shortening of growth period due to an earlier leaf senescence (Chen et al. 2020; Dox et al. 2020). These authors state that the same factors trigger leaf senescence and wood formation. The differences between individuals are due to genetic differences (Viherä-Aarnio et al. 2005; Rousi and Pusenius 2005) and the size/age of the individuals (Vitas 2011). All these factors determine the physiological characteristics and the response that the tree can give to climate variability, if the phenological calendar of B. pendula is regulated mainly by the annual climate regime, the phenological dates will vary between years, but there will be a synchronization in the phenological calendar and in the variability of trunk growth rates between individuals and between years.
Considering what has been briefly exposed and taking into account the latitudinal gradient, it would be expected that the birches in the Pyrenees, low latitude, have a more extended growth period and therefore wider annual rings than those of the birches in Northern Europe. But other factors come to play, e.g. altitude and annual precipitation distribution, that exert a significant effect on the birch phenology. For Betula spp. it is well documented that bud burst of high altitude trees are delayed in relation to low altitude trees (Myking and Heide 1995; Ovaska et al. 2005; Mir et al. 2016). While in the Nordic areas birch trees grow at low altitudes, the birch trees studied in the Pyrenees are located at 1,536 m asl, where mountain climatic conditions could minimize the latitudinal effect to the point that the growth period and/or the growth rate could be significantly reduced. In relation to precipitations its distribution and amount could also affect the period length and/or the rate of growth; in Southern Pyrenees, summer droughts and high temperatures in autumn can impair physiological processes (Fernández-Martínez et al. 2016) and growth arrest due to water stress (Cooke et al. 2012; Mwimba et al. 2018, Xu et al. 2022). Other differential factors with respect to Northern Europe, such as the length of the night and daily temperature fluctuations during the growth period, could also have a significant impact on the growth rate.
The main objective of our research was to determine the annual growth pattern in thickness of the trunk of Betula pendula and its relationship with the phenology of the shoots (buds burst and leaves longevity) in a location of the southern Pyrenees. To achieve this main objective, we have considered the following partial objectives: (1) to establish the phenological calendar: dates of bud burst and leaf formation, dates of start and cessation of cambium activity; (2) to estimate the growing degree-days (thermal time) for the initiation of bud and leaves development, and cambium activity; (3) to determine growth rates in diameter of the trees stem and to characterize growth dynamics using the Gompertz model; and (4) to assess the effect of climate on tree trunk thickness growth rates. To achieve our goals, we weekly monitored birch phenology and recorded stem growth increments with dendrometers on three trees over two years (2000–2001). We have also used additional phenological data from studies carried out in central and Northern Europe that have allowed us to compare the phenological calendars and the growth pattern in trunk diameter, and to establish and complete the latitudinal phenological gradient from the Pyrenees to northern Finland.
The study site is located in the southern Central Pyrenees (Catalonia, NE Spain), specifically in the Forest of Arrose in Esterri d'Àneu (42º 36' 11''N, 1º 6' 18''E), altitude of 1,536 m, on a steep slope (32°) and north orientation (16°N). The study site is a forest of young Scots pine trees (tree age about 60 years), tree density is 825 trees/ha and without understory. The average height of the pines is 11.48 m, mean diameter at breast height (DBH) is 24.61 cm. Small groups of multi-stemmed Betula pendula trees grow in the forest gaps, birches age is about 30 years at DBH. To carry out our study, we selected three birch trees located in different clumps about 50 m apart. The climate of the study site is mid-mountain (see on line resource 3 Fig. S1a) with a strong Mediterranean influence that is manifested in frequent summer droughts and a high variability in the annual climate regime, as shown by the highly contrasted climate regime of the two years of study. In 2000 (see on line resource 3 Fig. S1b), there was a strong drought in summer, while in 2001 there were droughts in spring (June) and summer (August) but rains were abundant in July that year (see on line resource 3 Fig. S1c). The main soil type of the area is entisol developed on metamorphic argillaceous limestone.
We carried out the study during the years 2000 and 2001. We recorded phenological data of the bud opening and of the fully expanded leaves, the appearance of the first yellow leaves, and the date of the fall of all the leaves every 7 days except in winter time.
Changes of Betula pendula in stem perimeter at breast height (DBH) were monitored using stainless-steel manual band dendrometers (Agricultural Electronics Co., Tucson, USA); (Keeland and Sharitz 1993; Gutiérrez et al. 2011). In February 2000, we installed band dendrometers on the dominant trees of 3 multi-stemmed trees of B. pendula at a height of 1.30 m from the trunk base. The DBH of the three selected trees (B1, B2, and B3) was 21.04, 10.11 and 22.79 cm, respectively.
The sampling period spanned from February 2000 to September 2001. Mean sampling frequency from April to November was 7.7 ± 2.4 days (mean ± 1SD). Sampling frequency was greater, 25.6 ± 10.7 days (mean ± 1SD) in winter, from December to March. The data obtained with the dendrometers were converted into diametrical growth, assuming that the trunk was cylindrical. Cumulative growth was obtained by summing the increments in diameter between consecutive samplings. Daily diametrical growth rates were obtained by dividing the increments produced between consecutive samplings by the number of days between samplings.
Band dendrometers detect changes in the tree trunk perimeter not related to the formation of wood, such as expansion and contraction of the bark and xylem, after an episode of rain or a prolonged drought respectively. To minimize errors, we took six measurements per dendrometer with an electronic caliper with a precision of 0.01 mm. We considered abnormal measurements those that could correspond to sudden changes (increase or decrease in diameter) of more than 15% of the total annual increase, and anomalous changes in diameter observed in a tree but not in others (Gutiérrez et al. 2011). Some anomalous measurements could be caused by animals or blockages in the parts of the dendrometer due to insects. This happened twice, and the measurements were not taken into account, the dendrometer was adjusted again.
The meteorological data used are from the meteorological station of Esterri d'Àneu, at an altitude of 948 m asl and an approximate distance of 3 km from the study area. Since the study site is located at 1,536 m asl, we made the altitudinal correction of the temperatures, considering a decrease of 0,6ºC for every 100 m (adiabatic gradient):
\({T}_{a}^{^\circ }={T}_{E}^{^\circ }-\left(\frac{\left(h\text{a}-h\text{E}\right)}{100}\right)0.6\) (Eq. 1)
Where \({T}_{E}^{^\circ }\) and hE are the temperatures and altitude of Esterri d'Àneu, and \({T}_{a}^{^\circ }\) and ha are the temperatures and altitude of the study site.
We calculated the variables that characterize the monthly and daily water regime of the study site. Evapotranspiration was estimated according to the Thornthwaite method (Thornthwaite 1948; Allen et al. 1998) using Esterri d'Àneu corrected meteorological data and estimated daily extraterrestrial radiation. To determine the water balance, we carried out a soil profile. The soil texture is clay loam, and we estimate a field capacity (Rmax) of 250 l/m2. The soil profile allowed us to estimate the root depth between 45–65 cm, we used 50 cm to estimate the water balance. To estimate the water balance we used the monthly and daily average temperature, the monthly and daily total precipitation (P), the real evapotranspiration (ETR), the potential evapotranspiration (ETP), the soil water deficit (D = ETP-ETR), soil water reserve (R), changes of soil water reserve (ΔR) and surface runoff (S = P-(ETR + ΔR).
To determine the accumulated heat, growing degree days (dd), until the opening of the buds, the development of the leaves and the beginning of the stem growth in diameter, we calculated the sum of the degree days from January 1 over the base temperature of + 5ºC (dd5) (see Rousi and Heinonen 2007). The model used (Eq. 3) is that of Spring Warming (SW) by Chuine et al. (2013), where t0, is the date of the year, Julian days, from which degree-days, \({R}_{f}\left({x}_{t}\right)\), begin to accumulate, F* is the critical sum of degree-days until growth takes place in date y according to equations (3) and (4).
\({S}_{f}={\sum }_{{t}_{0}}^{\text{y}}{R}_{f}\left({x}_{t}\right)\ge {F}^{*}\) (Eq. 3)
where
\({R}_{f}\left({x}_{t}\right)=\left\{\begin{array}{c}0, if {x}_{t}<{T}_{b1}\\ {x}_{t}-{T}_{b1}, if {x}_{t}\ge {T}_{b1}\end{array}\right.\) (Eq. 4)
where xt is the average daily temperature and Tb1 is the minimum or base temperature required to accumulate degree-days.
Daily growth rates and cumulative growth in dimeter at breast height (DBH) were used to determine the growth dynamics of birches during the 2-year of study.
We obtained the cumulative growth by adding the increment in diameter produced between successive samplings, and to characterize the pattern of cumulative growth we have used the 3-parameter Gompertz sigmoidal model (Winsor 1932). This function is characterized by its asymmetry, with an initial concave phase of accelerated growth (exponential) that decelerates from the point of inflection becoming convex. The expression of the function is
\(y=k{e}^{-{e}^{a-bx}}\) (Eq. 5)
where y is the accumulated growth in diameter of the trunk (mm); x is the time, (day of the year, DOY); k is the asymptote when y = k that remains constant over time, it is the growth produced in a period of one year (mm/year) that corresponds to the thickness of the annual ring; e is Euler's number (the base of natural logarithms, e = 2.71828); and the parameters a and b are positive constants related to the growth rate and displacement along the x-axis. From the adjustment of this function, we have calculated the maximum growth rate (MGR, mm/day) and the x and y coordinates of the position of the inflection point, where x is the DOY value and y the attained increment in diameter since the beginning of growth for that year.
\(MGR= \frac{bk}{e}\) (Eq. 6)
\(inflection point \left\{{}_{y=\frac{k}{e}}{}^{x=\frac{a}{b}} \right.\) (Eq. 7)
We have also characterized daily trunk growth rates in diameter by differentiating Eq. (5), so that
\(\frac{dy}{dx}=by{e}^{a-bx}\) (Eq. 8)
We have fitted the model with the Sigmaplot12 program (See the Gompertz function in on line resource 1 and on line resource 2 Table S3 for equivalences of function expression and parameters).
The calendar of the different phenological variables recorded for the birches during the 2-year period is shown in Fig. 1,Table 1 and in on line resource 2 Tables S1 and S2.
| Altitude (m asl) | Bud burst DOY(dd/mm) | dd5 | |
|---|---|---|---|
| Arrose, 42º 36’N, 1º 6’E (2000)1 | 1536 | 93(2/4) | 43.91 |
| Arrose, 42º 36’N, 1º 6’E (2001)1 | 1536 | 84(25/3) | 49.33 |
| 61°48’N (1992–1997)2 | 85 | 131(11/5) | 33.3 |
| 61°48’N (1993–1996)2 | 85 | 130(10/5) | 31.8 |
| 61°48’N 29°19’E (1997–2005)3 | 85 | 129(9/5) | 44.8 |
The date of buds opening is not coincident in the two years of study. The advancement of bud opening in 2001 compared to 2000 coincides with a faster accumulation of dd5 (shorter thermal time) in 2001, by March 25 (DOY84) 49.33 dd5 had already been accumulated, while by the same dates in 2000 there were barely accumulated 43.91 dd5 (Fig. 2 and Table 1). Low temperatures prevented enough dd5 until April 9, 2000 (DOY100). But the progressive reduction of night hours meant that, despite the lack of accumulation of dd5, on April 2, 2000 (DOY93), bud opening finally began (Table 1 and on line resource 2 Table S1).
In the year 2000, the leaves began to unfold on April 2 (DOY93) and we would have to wait until May 5 (DOY126) to find all the leaves fully formed. The first yellow leaves appeared on October 3 (DOY277) and all leaves turned yellow by October 8 (DOY282). In 2001, the leaves began to unfold on March 25 (DOY84), nine days earlier than in 2000. Unfortunately, we do not have the date for the unfolded leaves for 2001. We observed the first yellow leaves on August 27 (DOY239) that date is 38 days earlier than in 2000, and by September 11 (DOY254), and all leaves were yellow and falling that is 28 days earlier than in 2000 (on line resource 2 Table S1).
To determine the period of growth in diameter at breast height (DBH), we have considered that the trees start to grow when they reached at least 2.5% of total final growth and that the trees stop growing when the asymptotic value is reached (Fig. 1), which is equivalent to the final tree-ring width (on line resource 3 Fig. S2).
In the year 2000, we found a significant increase in diameter from May 12 (DOY133), 7 days after total leaf formation and 40 days from bud burst. The increase in diameter extended until August 4 (DOY217) (Fig. 1, on line resource 2 Table S2). From the beginning of August, the growth in diameter did not increase, having reached the maximum growth, only oscillations are observed around the asymptotic size of the diameter due to processes of hydration and dehydration of the trunk. The growing period of wood increments is 84 days in the year 2000.
In the year 2001, we measured significant increases in diameter from May 25 (DOY145), 61 days after bud burst. The increase in diameter extended until July 24 (DOY205). The growth period was only 60 days, 24 days shorter than in 2000. In this year the increase in diameter started 12 days earlier and ended 12 days later than in the year 2001 (Fig. 1, on line resource 2 Table S2), which translated into a wider growth ring (on line resource 3 Fig. S2).
Regarding the accumulation of degree-days (Fig. 2), it is observed that in the year 2000 the start of growth in diameter took place when 156.13 dd5 had been accumulated (DOY133). For the same dates in 2001, the degree-days accumulated were only about 115.88 dd5. The growth of the year 2001 was delayed until heat accumulation was 177.04 dd5 (DOY145) (Fig. 2).
The evolution throughout the year of the daily trunk growth rates is synchronous between the trees; the Spearman correlation coefficients (mean ± 1SD) for the years 2000 and 2001 are rS = 0.785(± 0.072), p-value = 0.000, n = 34, and mean rS = 0.871(± 0.029), p-value = 0.0001, n = 23. The smallest tree, B2, is the one that shows the lowest correlations with the other two trees (Fig. 1).
The dynamics of the accumulated wood increments throughout the growth period significantly fit the sigmoid Gompertz model, r = 0.998, p < 0.0001 and r = 0.996, p < 0.0001 for 2000 and 2001 (Fig. 3, on line resource 2 Table S3). The total increase in diameter at the end of the annual growth (asymptotic value, k) predicted by the model was 1.405 and 0.970 mm (mean of the 3 trees) for the years 2000 and 2001. And the total observed mean annual increment was 1.362 and 1.011 mm for the years 2000 and 2001, values very similar to those predicted by the model. There are differences in the total annual increase in diameter among the tree birches (Fig. 1, Table 2). Tree B2, the smallest one, always presented a lower cumulative increase in DBH, for the years 2000 and 2001 the increases were 0.531 and 0.528 mm/year, respectively. On the other hand, trees B1 and B3 with the largest initial diameter (21.04 and 22.79 cm, respectively) presented higher cumulative increase (2.075, 1.480 mm/year in 2000, 1.675, and 0.821 mm/year in 2001) (Table 2).
| YEAR 2000 | from the Gompertz model parameters | observed | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| B1 | B2 | B3 | average | B1 | B2 | B3 | average | ||||
| Asymptote (mm) | k | 2.161 | 0.552 | 1.501 | 1.405 | 2.075 | 0.531 | 1.480 | 1.362 | ||
| Inflection | (DOY) | x = a/b | 161 | 157 | 160 | 160 | 176 | 164 | 176 | 176 | |
| (mm) | y = k/e | 0.7951 | 0.2029 | 0.5523 | 0.5169 | 0.84 | 0.23 | 0.63 | 0,57 | ||
| Maximum growth rate (mm/day) | bk/e | 0.0426 | 0.0101 | 0.0283 | 0.0268 | 0.052 | 0.018 | 0.035 | 0.029 | ||
| YEAR 2001 | from the Gompertz model parameters | observed | |||||||||
| B1 | B2 | B3 | average | B1 | B2 | B3 | average | ||||
| Asymptote (mm) | k | 1.622 | 0.511 | 0.778 | 0.970 | 1.675 | 0.528 | 0.821 | 1.011 | ||
| Inflection | (DOY) | x = a/b | 157 | 160 | 154 | 157 | 167 | 159 | 153 | 167 | |
| (mm) | y = k/e | 0.597 | 0.188 | 0.286 | 0.357 | 0.490 | 0.270 | 0.400 | 0,679 | ||
| Maximum growth rate (mm/day) | bk/e | 0.049 | 0.012 | 0.023 | 0.028 | 0.057 | 0.018 | 0.028 | 0.031 | ||
The model predicts that the onset of growth in diameter in the year 2000, once the trees had already accumulated 2.5%, took place on May 17, 12, 15 (DOY137, 132 and 135) for birches B1, B2, and B3, respectively, the observed data were on May 13 (DOY133) for all trees. For the year 2001, the start dates were predicted to be on May 22, 12 and 18 (DOY142, 132 and 138) and the observed dates took place on May 25 (DOY145) for all trees. There is, therefore, a certain lag in the year 2001 in relation to 2000.
The mean maximum growth rates occurred on June 24 (DOY176) and on June 16 (DOY167) in 2000 and 2001, respectively, with a mean maximum growth rate of 0.029 and 0.031 mm/day (Table 2, Fig. 1 and Fig. 3). The maximum growth rate between trees is quite different, but the interannual values are similar. Tree B2, the smallest in DBH, is the one with the lowest maximum rate. The maximum growth rate calculated from the parameters of the Gompertz model and fitting the derivative to daily rates predicted that the maximum growth rate should occur on June 9 and 6 (DOY160 and 157) in 2000 and 2001, that is 16 and 10 days earlier than observed. The adjustment of the derivative to the daily rates predicts a maximum growth rate of 0.027 and 0028 mm/day for the years 2000 and 2001, rates similar to those observed, which are 0.029 and 0.031 mm/day (Table 2).
The percentage of the cumulative growth and growth rates of the DBH allows the comparison between years and birch growth from northern sites in Europe. In the Pyrenees, during the months of April and May, the accumulated growth was 19.71% and 10.69% in 2000 and 2001 (Fig. 4). The highest growth increase is reached in June, that was 44.12% and 65.73% and represents an accumulated growth of 63.83% and 76.41% in 2000 and 2001, respectively. During the month of July, the increase was lower, about 32.30% and 21.31%. Finally, between August and September the accumulated growth was only 5.81% and 2.27% in 2000 and 2001, respectively. Most of the growth is concentrated in June, which is very fast compared to the growth in Lithuania, but slower than that of the birch in northern Finland (Fig. 4a). The comparison of the DBH growth rates (Fig. 4b) shows that the longest period of growth is in Lithuania and the shortest in Northern Finland; nonetheless, the thickness of the rings of birch trees in the Pyrenees is similar to that of northern Finland (on line resource 2 Table S2).
In Figure S3 we depict the monthly hydrological regime together with mean birch growth for the two years of study, and in Fig. 5 the detailed relationships between growth and hydrological variables. The mean daily growth in DBH during the year 2000 shows a significant positive correlation with mean temperatures averaged between sampling dates (Fig. 5a). For this year, the strongest, positive, and most significant correlation coefficients are between the mean growth rate and mean minimum temperatures. Taking the trees individually, the correlation is significant only for trees B1 and B2. Regarding the accumulated precipitation between sampling dates, the correlation is also positive, but it is not statistically significant. On the other hand, the results for the year 2001 (Fig. 5b), show that although all the correlations between growth rates and temperatures continue to be positive, none of them is statistically significant; as for precipitations, only B2 tree growth rate, the smallest tree in diameter, shows a significant correlation.
The analysis of the association between the variables of the hydrological regime and monthly mean growth rate (Fig. 5c and Fig. 5d, Fig. S3) shows that monthly mean temperature has always a positive effect on birch growth rates, but the correlation is only significant (p < 0.025) for the year 2000. The ETP also shows a positive correlation for the two years, and its effect is significant at p < 0.01 and p < 0.05 for the years 2000 and 2001. In contrast, the soil water reserve fluctuations always show negative correlations, with the growth rates being significant (p < 0.025) in 2001. The strongest association of all is between growth rates and monthly ETR, which is positive and significant at p < 0.005 and p < 0.025 for the years 2000 and 2001, respectively.
The results of the relationships between the ETR averaged between sampling dates and the growth rate in DBH are depicted in Fig. 6a and 6b. For the year 2000, the correlation is r = 0.659, p ≤ 0.005 and for the year 2001 it is r = 0.422, p ≤ 0.025. As we show in Fig. 6c, in 2000, the growth rate in DBH started between May 5 and 12 (DOY126 and 133) when the ETR exceeded 1.7 mm/day. In 2001, growth in DBH started between May 19 and 25 (DOY139 and 145) when the ETR exceeded 1.8 mm/day (Fig. 6d). Although it is true that the previous week there were values that exceeded 2 mm/day, it should be noted that the previous fifteen days the ETR was very low, under 1.5 mm/day, due to low temperatures. This probably explains the onset delay of growth in DBH in 2001 compared to the previous year (see also Fig. 2).
Figure 6c depicts the joint evolution of the average growth rate together with daily ETR and ETP for the year 2000 (see also on line resource 3 Fig. S4a). The growth rate shows a strong decrease on June 18 (DOY170) related to the decrease in the ETR of the DOY162 and 163 up to 1.1 and 1.7 mm/day as a consequence of the drop in temperature, from 15ºC to 6 and 8ºC, respectively. The slow increase in temperature caused a delay of the ETR that did not recover above 3 mm/day until DOY175, which is reflected in the increased growth rate from DOY176 on. In DOY193, there was another drop in temperature from 18º to 10ºC, which produced a drop in the ETR to values of 2.1 mm/day, which is reflected in an abrupt drop in the growth rate in DOY207. From DOY210 up to DOY241, water deficit was strong; the ETP was much higher than the ETR due to low precipitations and a depletion of the soil water reserve that finally did not compensate for the water birches needed to transpire. Growth stops on DOY217, seven days after that water deficit began. Between DOY241 and DOY261, interspersed episodes of soil water deficit are observed. The leaves already began to yellow on DOY277, and by DOY282 all the leaves were yellow and most shed. From the first period of water deficit, small oscillations in the growth rate are observed, probably due to hydration of the wood and bark related to rainfall episodes that cause the ETR to rise occasionally. However, the effect of the bark hydration must be low, as it is only 3–5 mm thick (own measurements).
In like manner, Fig. 6d depicts the joint evolution of the mean growth rate and daily ETR and ETP for the year 2001 (see also on line resource 3 Fig. S4b). In DOY175, a decrease in the growth rate is observed to values of 0.001 mm/day, practically zero. Before that date, there is a period, from DOY161 to DOY170, of low temperatures of about 10ºC, when they should be about 15ºC. The accumulation of so many days of low temperatures resulted in a drop in the ETR to values of only 2.5 mm/day, well below the 3.5 mm/day that would be expected for those dates. In DOY200 there is another drop in ETR to 1.9 mm/day due to a drop in mean temperature from 16ºC to 8ºC that produced a drop in the growth rate on DOY205 to values of 0.006 mm/day, practically zero. In DOY214, but especially in DOY223, we find negative values of the growth rate that coincide with the beginning of the period of water deficit, which lasted until DOY264. The cessation of the DBH growth rate took place in DOY205, somewhat before the onset of the hydric deficit. On DOY239, the leaves began to turn yellow and on DOY254 they were all yellow, shortly before the end of the water deficit.
Phenological calendar of Betula pendula
We have determined the phenological calendar of Betula pendula in the Central Spanish Pyrenees for two years, 2000 and 2001. The data and the results are relevant since this is the first phenological study of B. pendula in Spanish Pyrenees considering buds, leaves, and cambium phenology. Moreover, the research is based on detailed ground phenological observations, highly valuable to assess tree life history traits and specie distribution under a warming climate (Chuine 2010; Liu et al. 2021). We have characterized the five typical phenophases, which in chronological order are i) winter rest, ii) opening of the buds, iii) unfolding of the leaves, iv) growing period in DBH, and v) the leaves shedding and the beginning of winter rest. The two years of study showed contrasting climatic conditions causing differences in the timing and duration of B. pendula phenophases.
i) The winter rest.
During the winter resting phase, with the measurements of the dendrometers, only oscillations in the DBH are observed due to processes of hydration of the trunk, all of them related to episodes of precipitation. The bark must also be involved in these oscillations, even though its thickness is only 3–5 mm (own measurements). The average mean temperatures in winter months are quite low in our study site, mean temperature is lower than 5ºC (on line resource 3 Fig. S1).
ii) The opening of the buds.
The opening of the buds seems to be determined by temperature and later regulated by the photoperiod length. In year 2000, long photoperiod reduced the time to budburst although sufficient degrees day had not been reached. Light and temperature are both the most important environmental cues whose signals are integrated into the biological clock (see Gil and Park 2018 and Xu et al. 2022). Our results show that in the year 2000 the accumulation of dd5 is lower than in 2001 and the opening of the buds is delayed with respect to the year 2001 (Fig. 2, Table 1). However, in 2000 the lower dd5 accumulation and the time delay of bud opening indicates that this phenophase is not only determined by the accumulation of heat hours (dd5) as pointed out by Linkosalo et al. (2006). Thus, as spring progresses, the longer photoperiods trigger the opening of the buds since the accumulation of dd5 is not enough as a result of a previous period of low temperatures. This result coincides with those reported for northern birch populations (Heide 1993; Myking and Heide 1995; Häkkinen et al. 1998; Junttila et al. 2012).
Birches bud burst took place on DOY92 with dd5 = 43.91ºC and on DOY84 with dd5 = 49.33ºC, a month earlier than at 61º48'N in Northern Europe but with a similar thermal accumulation (Rousi and Pusenius 2005; Rousi and Heinonen 2007) (Table 1). Rousi and Pusenius (2005) experiments growing different ecotypes together, found that the buds of the ecotypes from cooler regions generally opened before than the buds of the ecotypes from warmer regions in Finland. Although our birches are from further south, the fact that bud burst thermal requirements were by similar (dd5) as those of Lithuania and Finland indicates the adaptation to the cold climate conditions of the birches of the Pyrenees due to the altitude. The bud burst date advancement in the Pyrenees compared to northern sites is because the dd5 optimal values are reached earlier. Marchand et al. (2021) study at 51º12'N reported that the 50% of buds broke out on DOY101, that is 17 days later than the Pyrenees. Juknys et al. (2012) study at 54ºN reported a large interval, ranging from DOY60 to DOY115, 24 days earlier and 23 days later than in central Pyrenees. While Schmitt et al. (2004) study sites between 66ºN and 69ºN reported opening dates ranging from DOY154 to DOY175, around 66 and 87 days later, respectively (Fig. 7, on line resource 2 Table S1). Therefore, we verified that there is a progressive delay in bud burst as latitude increases due to the latitudinal temperature gradient, despite the higher altitude of the Pyrenean locality and that gradient is not linear (Fig. 7). Pudas et al. (2005), Hakkinen et al. (1995) and Juknys et al. (2012) worked with long data series of bud burst, their results show high interannual variation, but their bud burst data aligned to the general latitudinal trend (Fig. 7).
iii) Unfolding of the leaves.
We only have data for year 2000, birch leaves were fully developed by DOY126, 33 days after bud burst and 7 days before the start of growth in DBH (Figs. 1 and 2, on line resource 2 Table S1). In the Pyrenees, bud burst break earlier than in Central and Northern Europe, but leaves unfolding is delayed in relation to Central Europe (Marchand et al. 2021) and takes place earlier than in norther Finland (Schmiit et al. 2004). At 51º12'N in Marchand et al. (2021) 50% of the leaves were fully developed on DOY106 in 2018, five days after bud burst. Schmitt et al. (2004) found that in birch stands located between 66ºN and 68ºN in northern Finland, the interval between bud burst and the start of wood formation was about two and three weeks in 1996, and the leaves were fully developed one week before the start of wood formation. In contrast, in the plots located further north, at 69ºN, the development of the leaves took place three weeks after bud burst and ended one week after the start of wood formation (Fig. 7). In any case, the latitudinal gradients from the Pyrenees to Norhern Finland are not lineal unlike the gradient described by the data from Schmitt et al. (2004) along Finland, which a much smaller spatial scale.
Rötzer and Chmielewski (2001) calculated the onset of leaves expansion of Betula pubescens from a model that takes altitude, latitude, and longitude into account. We applied their model to our Betula pendula data, a species very close to B. pubescens, obtaining the DOY116, ten days before the observed DOY126. The same authors make an average estimate for several species in the area of the Alps, located about 300 Km further to the north, for the period 1961–1998, they obtained values of DOY105 for valleys and of DOY133 for stands at altitudes higher than 1500 m asl. The authors state that there is an advance of 2.4 days for every 100 Km in latitude from north to south. Extrapolating to our site in the Pyrenees, altitude about 1500 m asl, leaves formation would take place 7.2 days earlier than in the Alps, and therefore it would be the DOY128, a date very similar to the observed DOY126 we have recorded.
iv) Phenology of the stem growth in diameter at breast height (DBH).
Onset of stem growth.
The growth in DBH started on DOY133 and DOY145 in 2000 and 2001, respectively. These dates are similar to the average of the dates reported in northeastern Lithuania by Vitas (2011) and Kairiūkštis (1963 –in Vitas 2011). On the other hand, in northern Finland, at 66.5ºN, the onset of trunk growth using the pinning technique begins one month later, on June 24, 2004 (Schmitt et al. 2004) (Fig. 7, on line resource 2 Table S2).
Our results show that the starting date of the growing period in DBH is related to the progressive increase of temperature, as well as the accumulation of dd5 (Fig. 2). An increase in average temperature also implies an increase in ETP. But if the water from the winter snowmelt and precipitation maintains the reserve of soil water, then ETP coincides with the ETR and under these environmental conditions it is possible the beginning of the growth in DBH (Fig. 6c and 6d). We attribute the delay of the growth onset in DBH in 2001 compared to 2000, to a delay in the increase in ETR due to low temperatures. Precipitation and the presence of clouds also acting negatively, but in reality what happens is that the incident radiation is lower, with lower average temperatures and therefore lower ETR values.
Cessation of stem growth.
In our study site in the Pyrenees, growth cessation in DBH took place quite soon, on DOY239 and DOY254 in 2000 and 2001, respectively, most probably due to water deficit during the summer (Fig. 6, on line resource 3 Fig. S3 and Fig. S4); precisely the years 2000 and 2001 were characterized by low precipitations and summer drought (on line resource 3 Fig. S1b and Fig. S1c) although the average climatic regime does not show summer atmospheric drought (on line resource 3 Fig. S1a). Growth cessation in DBH at higher latitudes in Europe takes place one month or even later (Fig. 7, on line resource 2 Table S2) (Kairiūkštis 1963, Schmitt et al. 2004, Vitas 2011, Marchand et al. 2021; Dox et al. 2020), although the years of the study are not the same.
It is surprising that growth cessation occurred so soon if we take into account that birches are a free-growing tree species and that the temperature was still high. All seems to point out that it is not low temperatures that caused our trees to stop growing, but the water deficit. At the end of July, the water deficit together with high summer temperatures produces a sudden drop in the ETR with respect to the ETP, causing the cessation of growth in DBH. This contrasts with what happens in central and Northern Europe, where, in general, there is no water deficit, it might be that low temperatures at the end of summer could be the most important factor for growth cessation in DBH. In our study site, the Central Pyrenees, the beginning of the water deficit period causes the total cessation of growth in DBH, due to the high vulnerability of B. pendula to the lack of water causing even alterations in leaf morphology and a decline in net photosynthesis due to stomatal closure (Fernández-Martínez et al. 2016).
The duration of the stem growing period in diameter.
The period of wood formation is included within the period of the leaf phenology (on line resource 2 Table S1 and S2). In general, this is the pattern described for B. pendula, although in some studies it is reported that the growth in diameter ends after the leaves turn yellow (Schmitt et al. 2004). The growth period in diameter was shorter in 2001 than in 2000, 84 vs. 60 days, respectively, probably due to the lack of water (see drivers of birch growth in diameter below).
Birch wood growing period in our site is shorter than in Lithuania that lasts 87-133-day Lithuania as reported by Kairiũkštis (1963- in Vitas 2011) and Vitas (2011). In Northern Finland, Schmitt et al. (2004) reported the shortest growth periods of about 21 to 48 days (Fig. 7, on line resource 2 Table S2). Marchand et al. (2021) do not include the duration of the growth period, which we infer to be about 111 days. Probably, if our study in the Pyrenees had been carried out in a year without such a pronounced water deficit, the growth period would have been longer, equalling and even exceeding the period of Central Europe. In relation to climate warming, Michelot et al. (2012) predicted a lengthening of the growing period due to bud burst advance. However, water stress increase and hotter droughts in summer together with the depletion of soil water reserve can significantly shorten the growing period. Besides, episodes of irregular and heavy rainfall as occurs in our study locality in the southern Pyrenees do not favour the replenishment of water holding capacity of the soil.
v) The leaves shedding and the beginning of winter rest.
The growth cessation in DBH does not imply that birches enter into winter rest. This phenophase would begin once the trees have prepared their buds and lost their leaves. In general, low temperatures at the end of the summer are determinants to trigger the yellowing of the leaves and their progressive fall. This happened in 2000 coinciding with the first frosts at the end of September and the beginning of October (Estrella and Menzel, 2006). However, the lack of available soil water can be a stronger factor even though temperatures are high. In 2001, leaf colouring began 38 days earlier and fall 28 earlier than in 2000. According to our results, the cause of the autumn leaf events advancement in 2001 might be the drought of that year (Fig. 1, on line resource 2 Table S1 and on line resource 3 Fig. S1). The ecophysiological study by Fernández-Martínez et al. (2016) in theCcentral Pyrenees support our results; they showed that B. pendula is highly vulnerable to hydric stress in such a way that water deficit effects on this species even caused alterations in leaf morphology and a decline in net photosynthesis due to stomatal closure. Our results are in agreement with results from other authors (Juknys et al. 2012; Chen et al. 2020) but see Mariën et al. (2021); leaves colouring and shedding occurred very soon in both years and was most likely due to the pronounced hydric stress, although it was more significant in 2001. We found that water deficit in 2001 advanced the beginning of the leaves yellowing and falling, to the point that on August 27th (DOY239) the first yellow leaves appeared, and all the leaves were shed by September 11 (DOY254) long before the arrival of the first cold weather (Fig. 1 and Fig. 7, on line resource 2 Table S1).
Dynamics of stem growth in diameter at breast height.
The accumulated growth in DBH over the growing period was significantly described by the 3-parameter Gompertz sigmoidal model (Fig. 3, on line resource 2 Table S3). In 2001, tree birches reached earlier the inflection point than in 2000, although the girth growth began later; nevertheless, the attained growth in diameter at the inflection point was higher in 2000. Most probably due to the effect of temperature (see Fig. 2 and next paragraph about Drivers of growth in diameter).
The total accumulated annual wood (annual ring width), the asymptotic value, is very similar to the observed one (Table 2, on line resource 3 Fig. S2). The comparison of the accumulated growth increments in diameter with the data of other authors (Schmitt et al. 2004; Vitas 2011), allowed us to verify that the final growth is greater in Central Europe (Lithuania), being the ring width in Northern Finland very similar to that observed in the Pyrenees. The reason is that although the growing season in Northern Finland is considerably shorter, the growth rates are much higher than in the Spanish Pyrenees (Fig. 4, on line resource 2 Table S2), that might be attributed to the ‘polar day syndrome’ of northern birches (Tenkanen et al. 2022).
The stem maximum growth rate in diameter.
The maximum growth rates were attained on June 24 (DOY176) and June 16 (DOY167) in 2000 and 2001, respectively (Table 2), although the DOY of the maximum growth rate estimated according to the relationship of the model parameters, does not fully coincide with the DOY observed (see below). Rossi et al. (2006) proposed that in conifers from cold and temperate climates there is a synchronization between the maximum growth rate and the length of the day, after the summer solstice the growth rate begins to decrease. Although, we studied an angiosperm, in our work the dates of the maximum growth rate are close to the summer solstice, we think that the factor that could better explain the change in trend in the growth rate (the position of the inflection point) of Pyrenean birch trees could be the ETR, because it integrates the day length, as well as the temperature and the available water. The ETR could explain the different dates of the maximum growth rates between the two years of study. On the other hand, the results depicted in Fig. 4b point to that hypothesis, i.e. the DOY of the maximum growth rate in diameter of birches from several European sites is not synchronous.
Drivers of stem growth in diameter.
According to our results, all seem to indicate that ETR is the most significant variable to better explain the interannual variability of the growth rate pattern of B. pendula in our site in the central Spanish Pyrenees (Fig. 5). B. pendula is a vulnerable species to lack of water and is growing in the southern edge of its geographical distribution where the Mediterranean climate influence, characterized by summer droughts, is frequent as it happened during the two years of study. ETR integrates the average temperature, precipitations, soil field capacity, root depth, the length of the day, and the solar declination, with temperature and rainfall being the most determining factors due to their high inter- and interannual variability. On the other hand, in northern areas under Atlantic climate influence, the ETP and the ETR tend to coincide, as these areas do not usually present a marked water deficit. The ETP integrates mean temperature, day length, and solar declination. As the temperature is the only variable that varies from one year to the next, it could be also used as an index to understand the variation of the growth rates in diameter in Central and Northern Europe.
The relationship between the stem growth rates and the ETR points out that (1) the growth rate variability observed over the years is due to temperature variability when there is no water deficit. And, (2) to water stress; if there is a marked water deficit (ETR < ETP) the growth rate slows down and might even stop if the soil water reserve is depleted by evapotranspiration. The increase in temperatures, the length of the day, and the solar declination favour the increase of the ETP, which coincides with the ETR if there is no lack of water. An increase in transpiration is followed by an increase in growth rate (Fig. 6, on line resource 3 Fig. S4). Even though there are differences in the annual precipitation regime between 2000 and 2001 (on line resource 3 Fig. S1), the lack of precipitations in summer during the two years of study favoured the depletion of the soil water reserve, leading to a water deficit, then ETR shows increasingly lower values, causing a slowdown in the growth rate to the point it stopped when the soil reserve water was exhausted.
One of the commonly used drought indices is P-ETP; nevertheless, we ruled it out in our study because it does not give any statistically significant results for any of the two years of study (Fig. 5c and 5d). This is because it does not include the reserve of soil water. Trees’ transpiration does not depend directly on precipitation, but on the reserve of soil water availability.
The latitudinal phenological gradient of Silver birch along Europe.
The results from studies on Betula ssp. by other authors (see Fig. 4 and Fig. 7, on line resource 2 Table S1 and S2,) throughout Europe allowed us to draw a phenological gradient from 42ºN to 69ºN, from the central Spanish Pyrenees to northern Finland. With the exception of the site of the Pyrenees located at 1,536 m asl, the other localities are between 20 and 200 m asl. And, even though the years of study are not the same, there is a robust latitudinal pattern of the phenological phases across Europe (Fig. 7), temperature, precipitation, and day length covary with latitude. The latitudinal gradient of the different phenological phases depicted is not lineal except for the date of the maximum growth rate in DBH; of particular interest is the growth cessation in DBH, which is attained much sooner in the Spanish Pyrenees due to water stress, thus shortening the growing period which translates into narrow tree-rings.
The bud burst gradient, breaking out progressively from the south to the north of Europe, confirms the theory that predicts that bud burst depends on the accumulation of dd5 (Rousi and Heinonen 2007; Junttila and Hänninen 2012 among other authors). We expected an earlier bud burst of birch trees in the Pyrenees, but the fact is that there are no big differences from bud burst date in Central Europe (Fig. 7). In relation to leaves unfolding, a clear non-linear gradient is also observed. The big delay of the leaves expansion in the Pyrenees compared to central Europe, makes the gradient bends even more, probably due to the altitudinal effect (Fig. 7).
A similar pattern is observed for the starting date of growth in DBH, coinciding again in the Pyrenees and Central Europe. Our research shows that leaves unfolding occurred before the start of DBH growth, as would be expected in a tree with rings of diffuse porosity (Suzuki et al. 1996). However, Schmitt et al. (2004) reported that at latitudes higher than 68º N the leaves appear after the start of growth in DBH.
The gradient of the maximum growth rates in DBH is the only one showing a linear trend. The altitude effect is no longer observed in this phenophase in the Pyrenees, which occurs earlier than in Central Europe. The verification of this fact indicates that in birches it is not true that the maximum growth rate in DBH is regulated by the photoperiod, and it is not attained in the summer solstice as reported by Rossi et al. (2006) research on conifers in cold environments. In our opinion, the date of maximum growth rate might vary among species, years, latitudes, and climatic conditions of the site, probably because the day length conditions the ETP, which in turn depends on temperatures.
There are practically no data about of yellowing and leaf shedding that allow us to define a general latitudinal pattern of birch phenology. Chuine et al. (2013) stated that phenological models cannot predict leaf colouration so far, moreover there is a great dispersion of leaf colouration dates in the same locality, for example, Juknys et al. (2012) dates range is of two months wide for leaf colouring. Under high temperatures and lack of water, the biochemical mechanisms break down, making it impossible for the biological clock to work properly. This seems to be the case of B. pendula under water stress (Fernandez-Martínez et al. 2016).
The severity of climate change is well defined in the IPCC reports and its effects on phenology were already predicted by many authors years ago (Menzel and Fabian 1999; Chmielewski and Rötzer 2001; Rötzer and Chmielewski 2001; Menzel, 2000). These authors, among others, warn about the lengthening of spring due to a progressive advance in the growth onset, an increase in minimum temperatures, and a more irregular annual regime of rainfall with more torrential rains generating greater surface runoff. If we apply this scenario to the phenology of our birches in the Pyrenees, we could envisage
a shortening of the winter dormancy period. In the short term, this would not be a problem either in the Pyrenees at altitudes above 1,500 m asl or in Fennoscandia, where the risk of a chilling deficit is small (Myking and Heide 1995). In the Pyrenees to get out of dormancy, there will be enough days of low and negative temperatures to prepare the buds for sprouting (Welling et al. 2004) but more studies are needed to be given the complexity of the orography.
bud burst advancement and risks. The increase in minimum temperatures would favour the accumulation of degree-days that would advance the bud burst, increasing the risk of late frosts typical in mountain areas. Although birch trees can resprout after these frosts, it would make birches less competitive and more sensitive to parasites and disease (Mwimba et al. 2018).
the growing period would be shorter. The increase in temperatures will increase the ETP. The growth onset in DBH would advance as well if not affected by late frosts. But since the rains would be more irregular and torrential, the ETR would decrease in summer as a result of the water deficit and that will slow down the growth rate in DBH. Higher minimum temperatures would increase nocturnal respiration, which will also negatively affect the increase in DBH. Water stress episodes would also cause a loss of leaf mass, reducing the photosynthetic capacity of trees (Fernàndez-Martínez et al. 2016).
growth arrest in DBH until leaves shedding and start of winter rest. The aforementioned hydric stress would cause and an advance in growth cessation. It would also cause a sooner leaf shedding in the south and may be a delay in the north (Piao et al. 2019) and reduce trees’ competitiveness. Consequently, the growing period could be reduced and tree-rings would be narrower and forest net primary production lower (Michelot et al. 2012).
The increased frequency of more intense and hotter droughts would make birches less competitive in front of nearby Mediterranean species, and would also make them more sensitive to pests and diseases. A reduction in its distribution area could be a fact, some small forest stands that survive residually in humid valley bottoms and ravines and near streams probably will disappear. This expected situation for the southern edge contrasts with what is expected in Northern Europe, where the increase in temperatures and the lack of hydric stress may favour greater evapotranspiration and consequently the length of the growth period, resulting in greater CO2 fixation and therefore in the formation of much wider rings.
We have established the phenological calendar (buds, leaves, and cambium) for a population of birch trees from Southern Europe, years 2000 and 2001. Likewise, we have determined the growth in diameter of the trunk of the trees.
Temperature and soil water availability are the main drivers of spring and autumn phenological events of birch trees in the Pyrenees, since there is a synchronization in the phenological calendar between individuals, and the dates vary between years. The beginning of bud opening is mainly regulated by the accumulation of dd5 and by the photoperiod. The activity of the cambium took place after the leaf unfolding and ended before leaf shedding. The beginning and end of growth in DBH, as well as the variation of the growth rate throughout the year, are regulated mainly by the ETR, which in turn depends on temperatures and the soil water availability. Leaves autumn phenology is advanced in the calendar due to the hydric stress.
Pyrenean birch trees showed a delay in the expected phenological dates for bud burst, leaf unfolding, and the start of growth in DBH with respect to the phenological gradients along Europe. The altitude effect of the birches in the Pyrenees seems to be the cause of the delay in the accumulation of dd5 necessary for the buds opening and the achievement of minimum values of ETR to start the formation of the leaves and initiate growth in DBH. On the contrary, growth cessation in DBH due to water stress occurred in similar dates as birches in the far north, thus shortening the growing period. All the birch latitudinal gradients of the different phenophases are non-linear except the one of the maximum growth rates of the DBH increments, which is lineal.
More field studies in different locations would be advisable to verify this differential pattern of birch phenology in areas under Mediterranean climate influence from the Atlantic and Nordic areas. Dendrochronological studies would also be needed to confirm the direct relationship between the ETR and tree rings.
The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
The authors have no relevant financial or non-financial interests to disclose.
The two authors, Xavier Castells Montero and Emilia Gutiérrez Merino, equally contributed to the study conception, design, material preparation, data collection and analysis. The manuscript was written by Xavier Castells Montero and Emilia Gutiérrez Merino.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
posted
You are reading this latest preprint version