Seasonal Photosynthetic Activity in the Crown Compartments of European Ash (Fraxinus excelsior)

Abstract

Leaves facing different directions (north, south, east, and west) receive differing levels of illumination, resulting in spatial differences in photosynthesis P_N in the crowns of mature trees. We measured diurnal trends in P_N for a semi-solitary European ash (Fraxinus excelsior) over spring, summer, and autumn and compared these data with leaf biometric traits and leaf area distribution. The highest light-saturated P_N (P_Nmax) was to the south and west, and the lowest to the north. Likewise, intrinsic water use efficiency, defined as the ratio (P_N:g_S) of photosynthetic rate (P_N) and stomatal conductance (g_S), was also lowest to the north. The thickest leaves were found on the northern face and the thinnest in the south, suggesting differences in leaf anatomy may have contributed to differences in P_N. The greatest leaf area was recorded in the southern crown quadrant, which contributed more than 50% of the tree’s accumulated P_N. Our research emphasises the importance of choosing representative leaves for gas exchange measurements. In-depth studies into the spatial distribution of leaves and their traits will be necessary for accurate upscaling of leaf-level photosynthesis to whole tree and canopy levels.

1. Introduction

Canopy photosynthesis is the primary driver for plant growth and biomass production in numerous ecosystems, including forest stands [1]. However, photosynthesis is commonly measured at the leaf level, and upscaling from leaf level to tree and/or canopy level can present a significant challenge [2]. Of particular importance is canopy heterogeneity, e.g., tree dimensions, differently aged and developed foliage, 3D arrangement of shoots and leaves within the canopy [3], and the dependence of photosynthesis on variations in environmental conditions, e.g., temperature, radiation, or vapour pressure deficit (VPD), whether diurnally or within the canopy. The structure of the canopy will also influence local environmental conditions, providing crucial feedback between structure, environment, and growth. Several studies have shown that canopy structure can affect the absorption of photosynthetically active radiation (PAR), whether in the crown or its compartments, in both broadleaf [4] and coniferous species [5]. In other words, canopy structure and the 3D arrangement of its components can affect both the potential and actual rate of photosynthesis in particular crown compartments. Several 3D canopy architecture models have been developed to evaluate the relationship between canopy compartments and short- and long-term variations in intra-canopy microclimate, light interception, rate of transpiration, and carbon assimilation [6,7]. In solitary and semi-solitary trees, the foliage in different crown compartments may comprise different phenotypes to address environmental gradients that are heterogeneous in space and time [8]. This distinct expression of leaf phenotypes between crown compartments could lead to complementary patterns in light interception, photosynthesis, and production capacity. For example, Granado-Yela et al. (2011) identified temporal disparity between crown compartments in solitary European olive trees (Olea europea) derived from functional specialisation in photosynthetic behaviour at different functional and spatial scales, i.e., architectural structure (crown level) and carbon budget (leaf level). The same authors also reported differences in photosynthetic rate in leaves from different crown compartments, both in the actual rate of photosynthesis controlled by incident PAR and parameters derived from photosynthetic light-response curves [9].

The European ash Fraxinus excelsior is frequently found as a semi-solitary tree, and, consequently, optimisation of photosynthetic rates within the species’ canopy and its compartments may represent an important aspect of its growth and production strategy. In recent years, in situ measurements alongside modelling approaches have been used to assess within-canopy variation in PAR absorption, transpiration, and photosynthesis in solitary trees [10]. To date, however, net photosynthesis in F. excelsior has only been studied to a limited extent, with only fragmentary studies on seedlings, e.g., [11,12], while knowledge on photosynthetic rates as affected by leaf category (sun, semi-shade, shade), age, or intra-canopy microclimate remains missing. It is now well-established that maximum net photosynthetic rate (P_Nmax) is species-specific within the Fraxinus genus, with previous studies reporting a wide range, e.g., the south European flowering ash F. ornus 8.0 μmol (CO₂) m⁻² s⁻¹ [13], F. excelsior 3.8 μmol (CO₂) m⁻² s⁻¹ [14], F. excelsior 18.0 μmol (CO₂) m⁻² s⁻¹ [15], F. excelsior—seedlings 12.26 μmol (CO₂) m⁻² s⁻¹ [16], American ash F. americana 20.0–30.0 μmol (CO₂) m⁻² s⁻¹ [17], East-Asian ash F. rhynchophylla 8.0 μmol (CO₂) m⁻² s⁻¹ [18] and green ash F. pennsylvanica 14.6 μmol (CO₂) m⁻² s⁻¹ [19].

Likewise, there have been differences reported in net photosynthesis between sun and shade leaves [20], though some other studies have reported no substantial difference, as was the case for F. ornus seedlings [21]. In solitary trees, such as the kermes oak Quercus coccifera [22], foliage on the outer crown surface (the envelope) facing east, south, or west is considered sun leaf. The sun leaves of broadleaf tree species tend to be smaller and thicker [23] and to have higher photosynthetic rates per unit leaf area than shade leaves [24]. Another ash species, F. ornus, exhibits environmental plasticity in leaf structure (leaf mass per area (LMA), bulk tissue density, and thickness), which, in solitary trees, results in a markedly higher quantum yield in sun leaves than shade leaves [25]. It has been suggested that degree of shading may play an important role in such cases, with LMA in a moderately shaded F. pennsylvanica, for example, decreasing only slightly compared to an unshaded control, whereas shoot length and total leaf area increased significantly [26]. As such, variations in photosynthetic response and leaf architecture may be observed, depending on the degree of self-shading on the northern side of solitary trees (Fraxinus).

In general, the light environment within a solitary tree crown will be heterogeneous due to both a self-shading effect and the amount of incident light, determined by the elevation of the sun and the azimuth direction of incident light [27]. Consequently, photosynthesis and transpiration in foliage located within the crown will be controlled by both available light at a distinct point and microclimate drivers, such as air temperature and humidity. Nevertheless, our knowledge of how different crown compartments in solitary or semi-solitary Fraxinus respond photosynthetically to diel changes in PAR, temperature differences, and co-acting environmental factors remains very limited.

The main aim of this study was to perform a series of ecophysiological measurements in a solitary F. excelsior and evaluate variations in net photosynthetic activity in differently oriented (north, south, east, and west) crown compartments. Gas exchange measurements were used to evaluate diurnal and seasonal dynamics in P_N and transpiration within the tree canopy. Finally, we describe leaf area distribution and leaf morphology in relation to tree orientation. In doing so, we test the hypothesis that, on days of full sunshine, P_Nmax will be highest in east- and south-facing compartments due to leaf acclimation to direct sunshine and light-driven microclimatic parameters. We also hypothesise that the east- and south-facing compartments will contribute most to total carbon assimilation throughout the day. Finally, we hypothesise that leaf thickness and LMA will be lowest in the north-facing crown compartment.

2. Material and Methods

2.1. Site Description

This research took place at the 23 ha Michovky experimental tree nursery (VUKOZ Průhonice, Czech Republic; 49.9919031 N, 14.5765778 E), which was planted in 2004 with a range of experimental tree plots, comprising maples (Acer spp.), linden (Tilia cordata, Fraxinus spp.), rowan (Sorbus aucuparia), and Turkish hazel (Corylus colurna). Local climate data (precipitation, air temperature and humidity, and soil temperature and soil humidity at different depths) have been collected every 10 min. since 1997 at the Fiedler-Mágr automatic meteorological station, located ca. 100 m from the nursery. VPD was calculated from these data using the method outlined in [28].

The local soil is classified as Haplic Luvisol [29], with a higher content of loess in the upper soil layers, giving the soil a distinct heavy character. For a more detailed analysis of soil parameters, see [30]. The nursery understory was originally overgrown with an expansive lawn of native bush reed grass, Calamagrostis epigejos, though more recent intensive lawn management (weeding and removal of some trees) has resulted in a more varied plant species pool comprising 93 species, 44 of which are non-native and 13 classified as invasive. Understory conditions depend on season and soil water condition, with dry patches evident during hot and dry periods (from late spring to early autumn), especially on sunny exposures, and extensive moss growth (e.g., Brachythecium sp., Calliergonella cuspidate, Oxyrrhynchium hians) during the wet autumn–winter season. The most common grasses in the spring–summer seasons include common meadow grass (Poa pratensis), bush reed, and common mouse-ear chickweed (Cerastium holosteoides).

2.2. Experimental Trees and Measuring Periods

A 14-year-old, semi-solitary F. excelsior, 8.2 m high and 151 mm in diameter (DBH, d_1,3), growing at the southern edge of the nursery was selected for experimental measurements (Figure 1).

Experimental monitoring took place over three periods in 2017, each representing a different part of the vegetation season. The first measurements were undertaken in late spring (after full leaf canopy development) between 7 and 10 June, the second in summer between 23 and 29 July, and the final measurements in autumn between 27 September and 5 October.

2.3. Gas Exchange Measurements

Measurements of net photosynthesis took place in situ using a portable open-system LiCor-6400XT infrared gas analyser (Li-Cor, Inc., Lincoln, NE, USA) equipped with a transparent leaf chamber, with settings as follows: CO₂ mixer adjusted to natural CO₂ concentration before each measurement—varying between 385 μmol (CO₂) mol (air)⁻¹ and 405 μmol (CO₂) mol (air)⁻¹; flow controller set for 500 μmol s⁻¹ before each measurement; temperature and air humidity left unregulated to be as close to the surrounding environment as possible.

Each hour, a random leaf was selected from the edge and bottom crown sections from the four cardinal directions (north, south, east, and west) and measured. Five points were recorded on each leaf to improve accuracy and for statistical representation, meaning that 20 measurement points were taken each hour. As the whole period from morning to evening was covered, this means that a different number of points are measured each day over the whole measurement period due to changes in day length. Daily trends in net photosynthesis activity (P_N) were selected for the most representative sunny days over each measuring period (i.e., 8 June, 28 July, and 1 October), with the following measuring intervals: 9.5, 9, and 6.3 h, respectively.

2.4. Biometric Measurements

As a first step, crown length in the cardinal and inter-cardinal directions was measured from ground level using a geodetic tape measure. In October, all leaves from the tree crown were collected during an afternoon with total cloud coverage. Then, each branch was cut and wrapped in a waterproof plastic bag for immediate transport to the laboratory for further measurement. At the time of collection, the tree’s DBH and height were measured, along with the horizontal angle (azimuth) of each branch and the stem diameter at the height of each branch, to place the branch in Cartesian dimensions.

After collection, three leaf morphology characteristics were measured. First, we measured the total leaf area for each of the four geographic quadrants using a Leaf Area scanner (Masaryk University in Brno, Czech Republic) linked with a computer-calculated leaf area based on the contrast difference between shaded areas of the leaf and background. Next, three representative leaf samples from the two crown levels in each geographic quadrant were collected and transported to the laboratory, where we measured fresh mass and leaf area (as average with no variation) and dry mass after 12 h of drying at 105 °C. Prior to drying, the leaf thickness was measured using digital optical microscopy. First, the fresh leaves were cut to obtain a cross section across the widest part of the leaf blade. The cross-sections were then placed on a glass plate and observed under a VHX-5000 digital microscope with a maximum resolution of 18 megapixels (Keyence, Japan). Images of representative sections were then used to measure the leaf thickness (typically as 10 replicates for each cross section, performed for ca. 15 leaves in each crown compartment) using the linear distance measurements routine in the inbuilt Keyence software package (Keyence, Ver.1.7.0.4, 2014). In total, 150 measurements were taken for each leaf category, i.e., sun and semi-shade.

2.5. Physiological Data Processing

Based on the gas-exchange measurements, values for stomatal conductance (g_S) and P_N were recorded for each geographic canopy compartment (north, east, south, and west) by averaging five independent measurements per canopy compartment and time. From the daily trends obtained, we selected those measured on days of full sun or minimum cloud cover limiting solar radiation to provide representative days for each measuring period and daily g_S and P_N trends for each azimuthal canopy compartment. The relationship between g_S and P_N for each representative day was then calculated. All P_N data measured in June and July (n = 112 for each azimuth direction) were plotted against PAR at the time of measurement, as recorded by the LiCor sensor.

2.6. Statistical Analysis

After confirming the normality of the data using the Shapiro–Wilk test, a one-way ANOVA with the least significant difference test, followed by post hoc Tukey HSC tests, was used to assess differences in photosynthetic parameters. Where the normality test failed (e.g., analysis of leaf thickness), Kruskal-Wallis ANOVA on Ranks with subsequent pairwise multiple comparison procedures (Dunn’s Method) was used instead. All tests were carried out using SigmaPlot v.12.5 at a significance level of p < 0.05.

To construct P_N light-response curves, we first fitted a single rectangular hyperbola with three parameters to the data using SigmaPlot v.12.5. Equation coefficients were then estimated for the four geographic points and their standard errors compared by one-way ANOVA and post hoc Tukey HSC tests.

$$P_N=R_d+\frac{P_{Nmax}\ast PAR}{KI+PAR}$$

(1)

By using the single-rectangular hyperbola, we were able to interpret the following coefficients of the model in a physiologically meaningful way:

P_N: net photosynthesis;

R_d: dark respiration;

P_Nmax: maximum potential photosynthetic rate per individual;

PAR: a given light intensity measured by LiCor 6400 at leaf level;

KI: half-saturation constant; the light intensity at which the photosynthetic rate proceeds at ½ P_Nmax.

The following linear trapezoidal method [31] was used to integrate photosynthetic carbon gain over the course of a day:

$${\int }_a^{b}f\left(x\right)dx\approx \frac{P_{N2}-P_{N1}}{2}\{P_N\left(N1\right)+P_N(N2)\}$$

(2)

P_N2: subsequent time point of net photosynthesis;

P_N1: preceding time point of net photosynthesis.

The composite trapezoidal rule, applied to consecutive intervals, provided an approximate sum for the whole function. We used this method to calculate the accumulated P_N for the different crown compartments over selected days.

3. Results

3.1. Microclimate

The long-term (1997–2017) mean daily temperature at the site was 9.41 °C and the annual sum of precipitation was 548 mm (Figure 2A), while in 2017, the mean daily temperature was 9,9 °C and the annual sum of precipitation was 533 mm (Figure 2B). The highest PAR was recorded on 8 June 2017, when values reached 1756 μmol m⁻² s⁻¹ (Figure 3). In comparison, on 28 July 2017, scattered clouds and a lower solar angle resulted in a maximum PAR of 1644 μmol m⁻² s⁻¹, while the highest PAR on 1 October 2017 (autumn) only reached 1060 μmol m⁻² s⁻¹. At the same time, the maximum respective air temperatures were 25.9 °C (8 June 2017), 26.5 °C (28 July 2017), and 18 °C (1 October 2017). A similar decreasing trend was also observed for VPD, with maximum values of ca. 2200 Pa reached in spring, 1500 Pa in summer, and 1000 Pa in autumn (Figure 3).

3.2. Crown and Leaf Biometry

3.2.1. Crown Leaf Area and Projection

The greatest crown diameters were recorded in southerly, south-easterly, and south-westerly directions, reaching 3.4, 3.3, and 3.2 m, respectively, while the smallest crown diameters were recorded in easterly, north-easterly, and northern directions, at 2.3, 2.5, and 2.6 m, respectively (Figure 4A). The total area of leaves collected was 67.55 m², with 16.91 m² measured for northern leaves, 8.24 m² for eastern leaves, 29.05 m² for southern leaves, and 13.35 m² for western leaves (Figure 4B).

3.2.2. Leaf Biometry

LMA was highest in the east in both the upper and lower crown sections, with levels reaching 126 g m⁻², while the lowest LMA was observed in the west, reaching 98 g m⁻² in the upper crown layer and 95 g m⁻² in the lower layer (Figure 5).

Leaf thickness increased from the bottom of the crown compartment to the top in all cardinal directions except north, p < 0.001. Leaf thickness was lowest at the top of the western compartment (1274 μm) and highest at the top of the eastern compartment (1803 μm). Leaves in the southern compartment were significantly thinner than those in the north (p < 0.05).

3.3. Photosynthetic Activity

3.3.1. Maximum Rate of Photosynthesis

The highest P_Nmax was observed in the western (18.1 μmol (CO₂) m⁻² s⁻¹) and southern (17.9 μmol (CO₂) m⁻² s⁻¹) tree quadrants, while the lowest P_Nmax was recorded in the north (14.0 μmol (CO₂) m⁻² s⁻¹ (Figure 6). The fitted light response indicated that light-saturated P_Nmax differed significantly between cardinal directions (p < 0.001), with high P_Nmax levels to the west and south differing significantly from those in the north (the eastern P_Nmax did not differ significantly from any other quadrant). The half-saturation constant KI did not differ significantly between quadrants (p = 0.85) (

Table 1. Light response curve coefficients for the four cardinal quadrants. Rd = dark respiration; P_N max = maximum modelled photosynthetic rate per individual; KI = half-saturation constant. Significant values are presented in bold.

	South			East			North			West
	Coeff.	±SE	p-Value	Coeff.	±SE	p-Value	Coeff.	±SE	p-Value	Coeff.	±SE	p-Value
Rd	−3.08	1.81	0.09	−2.06	1.31	0.12	−2.68	1.74	0.13	−2.80	1.67	0.10
PN max	14.47	1.57	<0.0001	11.85	1.10	<0.0001	9.03	1.50	<0.0001	15.29	1.40	<0.0001
KI	159.58	55.96	0.01	176.34	75.27	0.02	105.98	63.88	0.10	182.43	73.81	0.02

).

3.3.2. Diurnal Trends in Photosynthesis and Photosynthetically Active Radiation

Daily P_N trends showed seasonal dynamics, with maxima reached during June—July and the lowest P_N levels recorded in October for all geographic quadrants (Figure 7). The highest P_N was recorded either before noon or in the afternoon in the eastern and southern quadrants, while the lowest values were recorded in the northern canopy compartment, irrespective of month or season. East-, south-, and west-facing leaves typically exhibit a unimodal P_N pattern, with a peak corresponding to the highest PAR (Figure 7). In contrast, north-facing leaves often showed a midday depression in P_N. While P_N was mostly positive during the day, it was negative throughout the day in the northern compartment during the autumn (due to low PAR), despite showing positive values in all other compartments.

The highest leaf illumination in the eastern crown compartment occurred during the morning, after which it decreased constantly until nightfall (Figure 7). In the southern compartment, however, PAR increased until early afternoon, reaching over 1500 µmol m⁻² s⁻¹, while in the western compartment it increased constantly until evening, though it never exceeded 1500 µmol m⁻² s⁻¹. In the northern compartment, PAR displayed a bi-modal distribution, peaking in the morning and evening with a midday depression and only exceeding 1000 µmol m⁻² s⁻¹ in June. Maximum PAR occurred in June (1711 µmol m⁻² s⁻¹) and the lowest in October (1060 µmol m⁻² s⁻¹).

Unlike P_N and PAR, g_s was much more uniform throughout the tree crown, with most compartments reaching their highest levels in the morning (Figure 7). Later, as VPD increased, g_s declined in all crown compartments, with the exception of the western compartment in October, where g_s was low in the morning and peaked in the afternoon, probably due to low PAR and mutual shading by branches.

3.3.3. Water Use Efficiency

The P_N:g_S relationship showed clear differences between compartments, with the highest P_N at a specific g_S occurring in the southern compartment and the lowest in the northern compartment, with eastern and western values somewhere between with no significant differences (Figure 7). While the initial phases of the curve slopes (g_S from 0–0.2 mol (H₂O) m⁻² s⁻¹) did not differ between cardinal quadrants (p = 0.48), the side intercepts of all quadrants showed a highly significant increase in P_N with increasing g_S (p < 0.001).

Water use efficiency WUE (Figure 7) was highest in June and lowest in October. WUE changed during the day, and it was usually highest in the specific crown compartment, which was illuminated by the highest PAR. Therefore, the lowest WUE was observed in the northern compartment. The most negative WUE during the season occurred in the west compartment in October in the morning and in the evening, as a result of the most negative P_N of all compartments.

3.3.4. Whole-Tree Photosynthesis

In all cases, the south-facing crown compartment contributed most to the total tree P_N (

Table 2. Numerically integrated net photosynthesis (P_N; mmol (CO₂) day⁻¹) for the four cardinal quadrants.

Measurement	South	East	North	West
Spring	8556 ± 125	2009 ± 26	3081 ± 74	2699 ± 41
Summer	7772 ± 43	1531 ± 12	1096 ± 36	1525 ± 9
Autumn	524 ± 22	394 ± 11	−369 ± 9	−84 ± 10

). In contrast, the contribution of the northern compartment was positive in June and July and negative in October. In spring, the total integrated daily sum of P_N per crown compartment was highest in the south, reaching 8556 mmol (CO₂) day⁻¹. While the contributions of all other segments were positive, the daily accumulated P_N in each of the other three segments only accounted for a maximum of 36% of southern P_N. Consequently, the southern crown compartment assimilated more CO₂ than all other segments combined. During summer, the ratio between P_N in the southern quadrant and the other quadrants was even greater than in spring, reaching 77%. P_N was lowest at the end of the vegetation season, with accumulated autumn P_N in the southern and eastern crown compartments positive but negative in the western and northern compartments (Table 2).

4. Discussion

Daily trends in F. excelsior P_N showed clear directionality, with the highest P_N occurring on the southern, eastern, and western sides and the lowest on the northern side. While light intensity appeared to be the main driver for differences in P_N with direction, it was not the only factor. Our study demonstrated that foliage properties with respect to P_Nmax also differed between cardinal crown compartments, with light-saturated P_N lowest to the north and highest to the south and west (Figure 6). Also, P_N per unit g_S was lowest facing north and highest facing south (Figure 8). Overall, leaves were thin with low LMA in the western and southern quadrants, where P_N was high, and thickest in the eastern quadrant, where P_N and LMA were high (Figure 5).

In our study, P_Nmax reached 18 μmol (CO₂) m⁻² s⁻¹, a value matching that reported for F. excelsior in a previous study [15], marking out F. excelsior as a species with a relatively high level of photosynthesis [12]. This high P_N corresponds with the species’ high hydraulic conductance, resulting from its ring-porous anatomy, which in turn results in a high g_S [32]. A secondary reason for the high P_N, however, could have been the high fertility of the nitrogen-rich soil at the experimental site [30]. On the other hand, the observed light-saturated P_N was often lower than 18 μmol (CO₂) m⁻² s⁻¹ due to a high VPD (up to 3 kPa; Figure 3C), to which F. excelsior responded sensitively with a decline in stomatal conductance [32].

Maximum P_N values were usually recorded on those sides of the tree directly facing the sun at maximum illumination. As in [12], we recorded the highest P_N rates in June, when PAR was highest, and found no evidence for possible photoinhibition on the south-facing side of the crown. While we failed to observe a midday depression in P_N in June, there was a small midday depression in July and October (Figure 7). A similar midday photosynthesis depression in fully lit leaves in the southern crown compartment has also been reported for other broadleaf tree species, with Koike et al. [33], for example, reporting a midday decrease in P_Nmax during summer in the crown sub-layer of F. mandshurica var. japonica. In this case, the upper crown surface may have acted as a screen, reducing penetrating light to a level optimal for sub-layer leaves. Einhorn [12], however, suggested that F. excelsior was relatively insensitive to photoinhibition when compared to European beech, Fagus sylvatica. A further reason for the midday depression in P_N may be associated with the increase in VPD, to which F. excelsior responds sensitively [32] by a decline in g_S [34,35]. In such cases, changes in g_S lead to a decrease in intercellular CO₂ concentration and a subsequent drop in Rubisco carboxylation efficiency [36]. Studies undertaken on broadleaf species over the last decade (e.g., ref. [37]) suggest that P_N could be limited not only by a midday decrease in g_S but also by leaf and stem water potential. For trees in warm environments, such as Mediterranean regions, and/or during hot summer events (heat waves), non-stomatal limitations may play a role in the midday depression of photosynthesis. Grassi et al. [38], however, suggested that this tends to be associated more with drought than radiation stress. In our case, the midday (13:00) depression in P_N rate observed in July in the southern canopy compartment correlated with a midday depression in g_S. On the other hand, g_S in July was roughly double that in June (Figure 7), suggesting that stomatal limitations imposed on photosynthesis were low. Consequently, we believe that the main reason for the midday decline in P_N was scattered PAR (Figure 3A) and the subsequent decline in energy available for photosynthesis.

The lowest P_N was always recorded in the northern crown compartment, irrespective of season, the main reason being the lack of PAR (Figure 7). On the other hand, we observed a midday depression in P_N in the northern compartment that was not the result of photoinhibition but rather caused by a decline in illumination. The biochemical properties of leaves may also have contributed to the low P_N in the northern compartment. For example, light response curves for photosynthesis data obtained in spring and summer indicated that, even in full light, P_Nmax was lower in the northern compartment (Figure 6), possibly due to a low maximum rate of carboxylation (V_Cmax) caused by poor nitrogen use efficiency in north-facing leaves [39]. At the same time, a low P_N:g_S ratio was observed in the northern compartment across the whole g_S range (Figure 6), indicating poor intrinsic WUE. Similarly, Le Roux et al. (2001) [40] identified a steep WUE gradient in the crown leaves of an isolated English walnut (Juglans regia), with the lowest WUE in the north crown compartment. According to Gregoriou et al. (2007) [41], leaves that develop under low light reduce their stomatal density and g_s. In such cases, however, P_N declines even more due to differences in the leaves’ internal structure, i.e., they have fewer palisade parenchyma. As the P_N:g_S ratio was low at both low and high g_S levels in our study, we suggest large non-stomatal limitations to photosynthesis, probably as a result of low mesophyll conductance in the northern compartment [40,42,43]. Since mesophyll conductance increases with increasing PAR, reduced illumination during leaf development in the northern compartment may be one reason for the low intrinsic WUE [44]. However, instantaneous WUE, calculated from transpiration, will not be as low as the lower leaf temperatures in the northern compartment, which will result in a lower leaf VPD than leaves in the southern compartment.

For each crown compartment, leaf photosynthesis was lower in the early autumn than in July (see P_N in Figure 7), the decrease most likely attributable to changes during the early phases of leaf senescence, which is typically accompanied by altered g_S (Figure 7) and temperature-induced drought stress effects in leaves. It is now well established that the photosynthetic CO₂ fixation rate declines in broadleaf tree species with drought stress, i.e., whenever leaf water potential declines from 0 to a critical value of −2.0 to −3.0 MPa [45]. Moreover, photoperiod in the early autumn season appears to be one of the main driving factors decreasing photosynthesis in broadleaf tree species [46], caused by photoperiod- and temperature-dependent decreases in the maximum rate of CO₂ fixation by Rubisco and the rate of ribulose-bisphosphate regeneration [47].

In this study, LMA differed between F. excelsior crown compartments (Figure 5). According to Gregoriou et al. (2007) [41] and Escribano-Rocafort et al. (2017) [48], LMA reflects leaf traits related to leaf position and exposure to direct light, which vary most in crown positions. Leaf thickness, for example, increased from the bottom to the top of the crown, as also found by [49]. Importantly, leaf thickness (Table 1) did not correlate perfectly with LMA when the thickest leaves were in question. Furthermore, the highest LMA was found in the northern compartment, and the thinnest leaves were found in the south. The combination of leaf thickness and LMA provides an estimate of leaf internal density, which affects mesophyll conductance to CO₂ [42,43], with low density suggesting a low P_N [50]. A low conductance for CO₂ resulting from the difference in leaf structure in the north may be one reason for the low WUE in this cardinal quadrant. Furthermore, the high density of thin leaves in the south-facing crown compartment may be related to a higher level of drought stress in these leaves than in the northern quadrant [51].

The total contribution of each cardinal quadrant to the total tree P_N results from the combination of leaf area, illumination levels, and the biochemical properties of the leaves in each compartment. When we integrated P_N over a whole day (Table 2), the highest P_N was recorded in the east- and south-facing crown compartments. Though the south received the highest illumination levels, P_N may already have been constrained by a high VPD at midday (Figure 3C). VPD will also be the main factor limiting P_N in the western compartment compared with the east [52]. The contribution of the northern quadrant to total tree P_N varied greatly throughout the monitoring period, with the distribution of P_N being bimodal in the spring and summer when the northern compartment received solar radiation in the morning and evening. As the days became shorter in the autumn, however, mutual shading by branches caused respiration to prevail over carbon assimilation. As a result, P_N in October was negative all day in north-facing leaves, while the other cardinal quadrants all yielded a positive P_N (Figure 7). Consequently, semi-solitary trees develop an asymmetric foliage distribution, with the largest leaf area in the south-facing quadrant. This high leaf area, combined with the highest illumination and high P_Nmax, means that >50% of all carbon was assimilated in the south-facing quadrant and <50% in the other three quadrants combined (Table 2). Two questions receiving little attention to date are how carbon assimilated predominantly in the southern compartment redistributes to the shaded parts of the crown and how water for transpiration redistributes from shaded to sunlit parts of the tree following the gradient of water potentials [53]. On a whole tree scale, roots and the crown are functionally connected, and aboveground asymmetry often correlates with asymmetry belowground [54]. Manipulative experiments on seedlings suggest that carbohydrates from a specific crown compartment predominantly support the same compartment below ground [55]. Consequently, future research could be directed towards assessing how asymmetry in carbohydrate production translates to asymmetry in the root system.

5. Conclusions

In this paper, the differences in P_N in different cardinal crown quadrants of a semi-solitary F. excelsior were described. We found that south-facing foliage contributed most to overall tree P_N, with most of the difference attributable to the intensity of illumination. However, the leaves also differed in P_Nmax under saturated light, suggesting a difference in biochemical limits for photosynthesis. The difference in P_Nmax was also associated with differences in intrinsic WUE. The southern crown compartment also contributed most to total tree carbon assimilation, which was exacerbated by the southern quadrant also having the largest leaf area. Future research employing larger tree datasets than those used in the present study will be needed to generalise these findings. Based on our findings, we suggest a need for standardising the selection of leaves when measuring P_N in mature trees to capture spatial variability within the canopy. Any extrapolation of single-leaf measurements to whole-tree and canopy levels should be accompanied by a detailed study of leaf area distribution.