Assessing stand structure of beech and spruce from measured spectral radiation properties and modeled leaf biomass parameters

Agricultural and Forest Meteorology 165 (2012) 82–91

Contents lists available at SciVerse ScienceDirect

Agricultural and Forest Meteorology journal homepage: www.elsevier.com/locate/agrformet

Assessing stand structure of beech and spruce from measured spectral radiation properties and modeled leaf biomass parameters Christian Hertel a , Michael Leuchner a,∗ , Thomas Rötzer b , Annette Menzel a a b

Chair of Ecoclimatology, Technische Universität München, Germany Chair of Forest and Yield Science, Technische Universität München, Germany

a r t i c l e

i n f o

Article history: Received 7 November 2011 Received in revised form 16 May 2012 Accepted 14 June 2012 Keywords: Light quality Norway spruce European beech Red/far-red ratio Blue/red ratio Green/red ratio PPFR PAR

a b s t r a c t Solar radiation is crucial for growth and competition within forest ecosystems. The spectral waveband between 400 and 700 nm is mainly responsible for photosynthesis and thus for plant growth. Within this spectral waveband, single spectral ratios (e.g. blue/red, green/red, red/far-red) influence and trigger different processes like leaf expansion, germination, stem growth and flowering. Spectral irradiance and biomass are heavily interrelated. Spectral radiation measurements covering the range 360–1020 nm were carried out with 130 sensors in six stand levels in a mixed plantation of Norway spruce (Picea abies [L.] Karst) and European beech (Fagus sylvatica L.). However, direct measurements of vertical and horizontal distributions of foliage are very complex and time-consuming and for this reason foliage biomass parameters of leaf area index (LAI) and specific leaf area (SLA) modeled by the growth model BALANCE are useful parameters to give complete stand representations. The interaction between modeled biomass parameters of European beech and Norway spruce and measured radiation profiles through all stand levels in this unique spectral and spatiotemporal resolution was the aim of this study. Both species showed the typical response to variation in light availability for both modeled parameters. Results exhibit a significant negative relationship between LAI and the photosynthetical photon fluence rate (PPFR) for both species. The blue/red (B/R) ratio showed significant negative relationships to LAI of both species. The vertical distribution of green/red and red/far-red in respect to LAI varied depending on the species and their morphological crown habit. Analyses of SLA and radiation under spruce showed no significant relationship at all. In contrast, beech showed significant relationships of SLA and the spectral ratios of red/far-red and blue/red. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Among various foliage biomass factors, light quality and quantity play important roles within forest ecosystems. Spectral irradiance within forest stands is strongly influenced by the architecture and distribution of foliage biomass in the canopy. The reverse is also true; with radiation influencing production of foliage biomass. This interdependence underpins the complexity of merging leaf biomass parameters and radiation datasets to understand their interactions better. Various studies (e.g. Hertel et al., 2011; ˜ Leuchner et al., 2007; Navrátil et al., 2007; Serrano and Penuelas, 2005) have shown that highly resolved spectral measurements are a tool to assess the complex canopy structure. The assessment of the distribution of the whole-tree foliage biomass is an important prerequisite to analyze its interaction with light. However, direct

∗ Corresponding author at: Hans-Carl-von-Carlowitz-Platz 2, 85354 Freising, Germany. Tel.: +49 8161 714747; fax: +49 8161 714753. E-mail address: [email protected] (M. Leuchner). 0168-1923/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.agrformet.2012.06.008

measurements of vertical and horizontal distributions of foliage are very complex and time-consuming in mature forest stands. Another crucial point is that optical foliage biomass measurements (e.g. Li-Cor LAI 2000, hemispherical fisheye photographs) deliver only useable results from ground level, with uncertainties at other levels. For this reason forest growth models can be a useful alternative to direct leaf collection and optical measurements. Furthermore modeled parameters and radiation measurements with optical fibers in different stand heights (Leuchner et al., 2005) are non-destructive and are able to capture long and consistent time series. One of the key parameters used to describe canopy structure and give stand representations is the leaf area index (LAI). The LAI (m2 m−2 ) quantifies the total amount of leaves or needles within a canopy and is defined as the total one-sided area of leaves per unit of ground surface area (Pokorny et al., 2008; Watson, 1947). The LAI represents the foliage which intercepts photons, transfers energy, and exchanges carbon and water vapor between forest stands and the atmosphere (Pokorny et al., 2008). Various studies (Reiter, 2004; Wang et al., 2004; Pretzsch et al., 1998) report mean LAI values for beech ranging between 4 and 6 during the period of

C. Hertel et al. / Agricultural and Forest Meteorology 165 (2012) 82–91

full foliation and for spruce between 4 and 8. Higher LAI values are not necessarily related to greater biomass production. Seasonal LAI changes in deciduous forests are experienced during leaf unfolding and senescence. Chen (1996) also investigated seasonal variations in LAI in coniferous forest stands of about 25–30% during growth cycles. Another focus is on the determination of the specific leaf area (SLA) that is strictly linked to that of LAI. SLA (m2 kg−1 ) is a measure of leaf thickness and is defined as projected leaf area per unit leaf dry mass and relates to light conditions in deciduous and coniferous tree species. SLA is therefore well suited to characterize the physiological and morphological properties of the foliage within branches, which experiences gradually changing light conditions. Light, and therefore SLA, scales with height in the canopy (Landhäuser and Lieffers, 2001; Kazda et al., 2000; Niinemets et al., 1999; Cemark, 1998; Morales et al., 1996; Matyssek, 1986). Cemark (1998) reported a strong increasing vertical SLA gradient for beech stands which was related to the cumulative LAI profile through the canopies. Thus, SLA at a particular crown location is directly or indirectly related to light availability at that point. The morphological adaptation of leaves leads to the determination and differentiation of sun and shade leaves (e.g. Reiter, 2004; Abrams and Mostoller, 1995). Sun adapted foliage shows a significantly higher photosynthetic activity compared to shade adapted foliage (Urban et al., 2007; Marek et al., 1999). While recent work has described the role of canopy structure on light transmittance in various forest stands (Leuchner et al., 2007; Pecot et al., 2005) information is lacking how different spectral properties vary in different stand layers and how they are statistically related to biomass parameters such as described above. Different radiation properties between direct and diffuse radiation at different canopy heights influence morphological processes. Overcast sky conditions (OVC) have a major effect on the spectral irradiance within canopies. Various spectral wavebands and ratios such as the red/far-red ratio (R: 655–665 nm, FR: 725–735 nm) exhibit specific behavior under OVC conditions. Studies have shown that R/FR decreases under OVC conditions and below denser canopies (Hertel et al., 2011; Leuchner et al., 2007; Pecot et al., 2005; Endler, 1993; Holmes and Smith, 1977). Below deciduous trees, R/FR declines more than under coniferous canopies triggered by a higher selective transmission of light and a higher absorption of R light by the leaves (Leuchner et al., 2007; Lieffers et al., 1999; Federer and Tanner, 1966). Furthermore, plants can use diffuse sky conditions more effectively than direct radiation for photosynthetic activities (Campbell and Norman, 1998). As solar radiation penetrates into the canopy, the spectral composition changes due to absorption, reflection and transmission e.g. for the blue/red ratio (BW : broadband blue 400–500 nm; RW : broadband red 600–700 nm) as shown by Hertel et al. (2011) and Navrátil et al. (2007). This increase of BW /RW may stimulate photosynthetic activity (Matsuda et al., 2007; Vogelmann and Martin, 1993). There is little knowledge about the green/red ratio (GW : broadband green 500–600 nm, RW : broadband red 600–700 nm). The green waveband usually has relatively higher reflectance. Grant (1997) and Navrátil et al. (2007) observed the highest relative proportion of GW photons under clear sky (CS) conditions and at solar noon. All the ratios described above are within the photosynthetically active radiation (PAR: 400–700 nm) waveband. PAR is absorbed by chlorophyll in the PSI and PSII photosystem and carotenoids (Grant, 1997). In order to address the objectives to estimate structural stand attributes in complex forest stands a combination of models and measurements were used with the help of spatially and temporally highly resolved data for mature forest stands. The importance and usefulness of spectral measurements were evaluated above.

83

Fig. 1. Scheme of the experimental setup of sensor profiles within the forest stand.

In addition the overall objectives of the current study were (1) the description of stand structure by measured vertical radiation profiles from the period of full foliation until after defoliation to track seasonal changes for the implementation in further studies and (2) the comparison of a measured spectral radiation dataset with modeled leaf biomass parameters to detect detailed relationships between these factors. 2. Materials and methods 2.1. Radiation data The intensive investigation plot ‘Kranzberger Forst’, 35 km NE of Munich, Germany (48◦ 25 08 N, 11◦ 39 41 E, 485 m a.s.l.), where all radiation measurements were obtained is a mixed plantation of Norway spruce (Picea abies [L.] Karst) and European beech (Fagus sylvatica L.). In 2005 the spruce stand was 54 years old and the beech stand 61 years old with a single storey structure. For the main research area a number of 194 trees and a single story structure were given with a total area of 30 m × 30 m. The shortest distance between a sensor profile and a small clearing was about 15 m, thus avoiding significant influence of clearings on the data (Hertel et al., 2011). The maximum leaf area density for beech was situated in the upper third and for spruce in the lower half of the canopy (Häberle et al., 2003). The basal area is 46.4 m2 /ha the stand density 764 stems/ha (Wipfler et al., 2005). The measurements were performed by a self-built 130 sensor system consisting of a miniature multichannel spectrometer (Zeiss MCS module UV–VIS–NIR) with quartz glass optical fibers (0.6 mm core diameter) and sensor heads of white 10 mm polyoxymethylene spheres (Leuchner et al., 2005). The system was developed with a total measuring cycle of 2 min 15 s for all 130 locations providing a quasi-simultaneous acquisition of the entire three-dimensional radiation regime in the stand. The spectral range obtained was 360–1020 nm with a high spectral resolution of 0.8 nm. An extensive description of the single components and the calibration procedure is described by Leuchner et al. (2005). Spectral measurements were performed on a regular grid consisting of 25 vertical profiles in five different stand layers (3 m: z/H = 0.11, 14 m: z/H = 0.54, 17 m: z/H = 0.65, 20 m: z/H = 0.77 and 23 m: z/H = 0.88; z/H represents the relative height to the apex) and above the canopy (26 m: z/H = 1.0) (Fig. 1). From this experimental setup, it was possible to produce a very detailed description of the spatial distribution of solar radiation for both beech and spruce. In order to address the objectives of this study, analyses were performed for beech and spruce for all weather conditions during the different phenological stages of full foliation (August,

84

C. Hertel et al. / Agricultural and Forest Meteorology 165 (2012) 82–91

September 2005), leaf fall (October 2005) and during the defoliated period of beech (November, December 2005). The grid of vertical profiles was divided into beech (12 profiles) and spruce (13 profiles) sections. Data from the 25 sensors per level were assigned accordingly to the species in or under whose crown the sensor was located. The mutual influence of the species was negligible because of high crown density and sufficient distance between the species. All sensors for a species were averaged for each stand height and only measurements during positive solar elevation angles were considered. For the spectral analysis an extended selection of three different spectral ratios and the PAR waveband was chosen. Broadband blue (BW ) and broadband red (RW ) were used for the determination of the BW /RW ratio. Similarly broadband green (GW ) and RW were used for the GW /RW ratio. R/FR was calculated by dividing the total irradiance of red light (R, 655–665 nm) by the total irradiance of far-red light (FR, 725–735 nm) (Pecot et al., 2005; Capers and Chazdon, 2004). The entire photosynthetically active components (PAR) are measured as the photosynthetic photon fluence rate (PPFR). Sensors defined different layers of the canopy as follows. The sensor at 3 m was used to represent the stem and forest floor, those at 14 m and 17 m for the shade crown, and those at 20 m and 23 m for the sun crown. For beech, only the sensor at 23 m was used to represent the sun crown due to the experimental setup and the stand structure of the adult species. The performed radiation measurements consider radiation out of all angles of incidence (4 sr) in equal fractions in contrast to the frequently used cosine response of planar sensors to better account for the biologically relevant amounts and spectral ratios actually seen by the phytoelements (e.g. Björn and Vogelmann, 1996; Smith and Morgan, 1981). 2.2. Modeled biomass data Leaf biomass for all trees that are needed for this study has not been measured, thus, simulated data of the model BALANCE were used. LAI and SLA data are the results from a simulation of the physiological growth model BALANCE (Grote and Pretzsch, 2002) based on initial information of the spruce and beech trees, on soil information and on meteorological data collected at the same site. The model BALANCE is able to calculate the 3-dimensional development of individual trees or forest stands and to estimate the consequences of environmental changes. As an individual tree based model, BALANCE simulates growth responses at the single tree level, which enables an estimation of the influence of competition, stand structure and species mixture. Tree development is described as a response to individual environmental conditions and environmental conditions change with individual tree development. For growth simulations carbon, water and nutrient balances of the individual trees form the fundamental processes. Biomass increase is the result of the interaction of these physiological processes, which depends on a tree’s microenvironment that is itself influenced by the stand structure. Each tree is divided into crown and root layers, which are in turn divided into eight crown and root sectors. For each layer, each sector’s microclimate and water balance are computed daily, whereas the physiological processes assimilation, respiration, nutrient uptake, growth, senescence, and allocation are calculated at ten day or monthly time steps from the aggregated driving variables. Photosynthesis within a segment is calculated using the method of Haxeltine and Prentice (1996) and related to the water balance via stomatal closure. Carbon and nitrogen allocation to roots, branches, foliage and stem is computed on the basis of functional balance and pipe model principles (e.g. Mäkela, 1990; Shinozaki et al., 1964). Foliage biomass, which forms the basis of the LAI calculations of a segment, is estimated from the segment volume, the actual foliated segment, the foliage density and the SLA of the segment. The SLA

of a segment is computed based on the parameterized minimum and maximum specific foliage area, a given scaling parameter for the SLA and the segment’s competition factor (Grote and Pretzsch, 2002). To depict biomass, LAI and SLA values and the annual cycle of foliage development must be known. From the beginning of leaf bud burst biomass, leaf area, light availability and radiation absorption change. Thus, the date of foliage emergence of a tree determines its assimilation and respiration rate but also affects the environmental conditions of the trees in its vicinity. In BALANCE the bud burst of a tree is modeled using a temperature sum model (Rötzer et al., 2004), while foliage senescence is estimated according to the respiration for each segment of a tree (Rötzer et al., 2010). A more detailed description of BALANCE can be obtained in Grote and Pretzsch (2002), and Rötzer et al. (2005, 2009, 2010). 2.3. Statistical analysis The response of light availability to variation in leaf biomass parameter distribution was investigated for each species, three spectral ratios and PPFR using linear and logarithmic regression analyses representing the months August and September 2005 for the period of full foliation depending on the relation to describe the main attenuation of the radiation. Probability values of p < 0.05 were considered significant and those at p < 0.001 highly significant by using t-test. The logarithmic regression was only chosen for the PPFR–LAIcum analysis, for a better explanation of the distribution within the upper stand layers. Different definitions of LAI were used for the analyses. For the stand description depicted by boxplots (Fig. 2) and for the linear and logarithmic regressions a cumulated LAI (LAIcum ) from all modeled heights between the apex and the forest ground was used. For analyses of the vertical distribution an individual LAI (LAIlay ) for each layer with LAI sums from 1 m above to 1 m below the respective sensor was used. The respective definition of LAI was used depending on whether a layer specific analysis or the linkage to the whole stand structure was performed. SLA values for each layer were determined as the arithmetic mean from 1 m above to 1 m below the respective sensor. Both parameters were also averaged separated by species and their affiliation to the respective sensor profile by month. Ten modeled single trees for spruce and 29 single trees for beech were taken into account of the analyses. Monthly mean values especially of LAIcum were explicitly chosen, because values are better comparable and realistic for understanding the stand structure. Statistical analyses were conducted using SPSS 14.0 and Sigma Plot, version 6.0. 3. Results 3.1. Distribution of biomass during the period of full foliation Fig. 2 shows boxplots of the modeled vertical pattern of SLA and LAIcum through the stand separately for beech and spruce. The percentile distributions of the layer mean values of both variables were calculated for August and September 2005, the period of full foliation based on a parameterization obtained for the research site (Grote and Pretzsch, 2002). Differences in crown morphology for spruce and beech are apparent through all stand heights. Fig. 2A and B reflects the species related shape of the crown. The modeled cumulative LAIcum for spruce (Fig. 2A) increased from the apex (LAIcum = 1.16, z/H = 1.00) to the ground layer (LAIcum = 3.8, z/H = 0.11). Furthermore it showed a stronger gradual increase from 23 m (z/H = 0.88) down to 17 m (z/H = 0.65) than beech, which is in good agreement with the morphological crown habit of spruce. The modeled mean value of LAIcum for the spruce stand was 3.80 m2 m−2 . Fig. 2B displays the LAIcum for beech. From the height of 23 m (z/H = 0.88) down to 20 m (z/H = 0.77) a rapid

C. Hertel et al. / Agricultural and Forest Meteorology 165 (2012) 82–91

LAIcum[m2 m-2]

SLA [m2 kg-1]

SLA [m2 kg-1]

height [z/H]

height [z/H]

LAIcum[m2 m-2]

85

Fig. 2. Boxplots of the modeled mean distributions of SLA and LAIcum for spruce and beech. Means for the period August and September 2005 ((A) LAIcum spruce, (B) LAIcum beech, (C) SLA spruce, and (D) SLA beech).

increase of LAIcum occurred within the foliated canopy. Below 20 m (z/H = 0.77) LAIcum values were very similar, since almost no leaves and branches were found in the lower layers. Mean values of modeled LAIcum at the forest floor (z/H = 0.11) mirrored an LAI for beech at 2.31 m2 m−2 . The vertical distribution of SLA in spruce and beech is shown in Fig. 2C and D. The modeled values of SLA for spruce ranged between 2.47 m2 kg−1 at 14 m (z/H = 0.54) and 4.90 m2 kg−1 at 17 m (z/H = 0.65). An increase of SLA from the apex down to 17 m (z/H = 0.65) was investigated for spruce. In general, the foliage in the shade crown was characterized by higher SLA than in the sun crown at 23 m (z/H = 0.88). SLA of beech was consistently much higher than for spruce at the same height, and ranged between 26.96 and 39.40 m2 kg−1 (Fig. 2D). The highest mean values were found at a height of 17 m (z/H = 0.65). Foliage in the lower crown layers at 14 m (z/H = 0.54) had higher SLA, and did not reach the minimum values of the sun foliage (Reiter, 2004; Pokorny and Marek, 2000).

3.2. Vertical distribution of spectral ratios and biomass parameters The vertical distribution and availability of light is one of the crucial factors for the existence and production of biomass. As already shown in various studies (e.g. Petritian et al., 2009; Leuchner et al., 2007) there is a strong correlation between light and height in forest stands. More interesting is the interaction of biomass with different spectral ratios. Fig. 4A, B, D, and E displays the modeled layer specific vertical distribution of LAIlay and SLA separately for beech and spruce and in all investigated periods. LAIlay for beech was higher in the sun crown, especially in August and September. Modeled results of biomass

distribution in spruce resulted in higher LAIlay values within 17–23 m (Fig. 3D). September LAIlay values shown in Fig. 3A and D were nearly equal to the August values but maxima differed as a result of the beginning of senescence of the leaves. For beech, the distinctive decrease of LAIlay in October at 23 m (z/H = 0.88) is visible as a result of senescence and so the LAIlay for 20 m (z/H = 0.77) was modeled as nearly equal (Fig. 3A). After defoliation beech has no leaf biomass left and smaller changes in biomass for spruce were modeled by BALANCE. Other crown dimensions determined foliage biomass within this period for spruce especially at the height of the stem story and shade crown. The graph in Fig. 3B shows that the distribution of SLA matched the LAIlay distribution below the dense broadleaved canopy for the species beech. SLA increased with canopy depth in the shade crown of both species and had its highest value of 38.50 m2 kg−1 at 17 m (z/H = 0.65) for beech. Values of spruce SLA were always lower in comparison to values of beech at the same height (Fig. 3E). The higher values of SLA at 14 m and 17 m matched the LAIlay distribution (Fig. 3B and E). PPFR of beech showed a strong decrease down to 17 m (z/H = 0.65; Fig. 3C). Slightly increasing values at the forest floor were observed in August. September values of PPFR were similar to August but no increase within the ground layer was visible for beech. PPFR for spruce decreased gradually from the apex down to the ground layer (z/H = 0.11). The gradient of PPFR in October for both species was similar to the stages of full foliation but showed lower overall values. Differences were observable even within the month of leaf fall. Lowest PPFR values were observed within the months November/December and values are, e.g. 58% lower for the height of 26 m, 69% lower for the height of 20 m and 53% lower for the height of 3 m compared to August.

C. Hertel et al. / Agricultural and Forest Meteorology 165 (2012) 82–91

height [z/H]

height [z/H]

height [z/H]

height [z/H]

86

Fig. 3. Vertical distributions of modeled LAIlay , SLA, PPFR, BW /RW , GW /RW and R/FR for 2005 separately for beech and spruce. Defoliated period (November to December) was summed up.

The spectral ratio of BW /RW in August increased within the areas of higher biomass, especially in the sun crown of beech (Fig. 3G). Within the shade crown the ratios decreased obviously. The spectral ratio of BW /RW for spruce showed also an increase within the areas of higher biomass (e.g. for August 2005 from BW /RW of 0.74 above the canopy up to 0.81 at 20 m (z/H = 0.77). Below that height of z/H = 0.77, values decreased down to the forest floor. The BW /RW gradients under beech and spruce for September and October were also comparable to the previous month except lower overall values were obvious throughout all layers. Maximum values were found at 23 m (0.84) under beech and 20 m (0.93) under spruce in September. After defoliation of beech changes were also visible for

the course of the spectral ratio. No peak of BW /RW within the height of the sun crown was observed. Slightly lower values of BW /RW at 17 m (z/H = 0.65) were apparent (Fig. 3G). The GW /RW gradients showed quite similar behavior. In August increased GW /RW was observed within the areas of higher biomass, especially in the sun crown of beech (Fig. 3H). Results for September and October increasingly converged to the gradient of BW /RW in the same months. GW /RW results under beech for the period November/December showed decreasing overall values and a less distinct peak within the heights of 23 m and 20 m. R/FR showed a pattern with a similar course as PPFR. R/FR also decreased downwards through the canopy and only at ground level

C. Hertel et al. / Agricultural and Forest Meteorology 165 (2012) 82–91

87

Fig. 4. Dependency of spectral radiation properties (BW /RW , GW /RW , PPFR, R/FR) and LAIcum for beech (A) and spruce (B); when no significance was existent, the regression line was omitted in the plot.

88

C. Hertel et al. / Agricultural and Forest Meteorology 165 (2012) 82–91

(z/H = 0.11) an increase was observed in August (Fig. 3L) for both species. Stronger variations within the gradient of beech compared to spruce could be observed for the R/FR (Fig. 3I). The increase of R/FR in September at the ground layer was less distinctive for beech and spruce compared to the gradient observed for August. Measured results of R/FR for beech in the period of defoliation showed a gradual decrease from the apex down to the ground layer. This was in contrast to the R/FR gradient within the period of full foliation. Results for spruce differed only in the overall extent of the values compared to the period of full foliation. 3.3. Interdependency of radiation and biomass The interaction of radiation properties and biomass parameters during the period of full foliation (August, September 2005) is shown in Figs. 4 and 5. The LAIcum and BW /RW for spruce (Fig. 4B) showed a significant negative relationship (R2 = 0.40, p = 0.02). Higher variability in the data and no significant relationship between decreasing LAIcum and increasing GW /RW ratio was observed (p = 0.64). There was a highly significant negative logarithmic relationship between LAIcum and PPFR with a coefficient of determination of 0.88 (p < 0.0001). The relationship between LAIcum and R/FR was also significant (R2 = 0.69, p = 0.003). LAIcum characteristics and the relationship with beech radiative parameters are displayed in Fig. 4A. For beech a significant negative correlation between LAIcum and BW /RW (p = 0.01) was observed as well with a higher coefficient of determination (R2 = 0.58) than for spruce. In contrast to the results for spruce, LAIcum and GW /RW showed a significant R2 of 0.62 (p = 0.007). Significant results were also observed for the PPFR waveband under beech (R2 = 0.28, p < 0.0001). In contrast to spruce, the lower LAIcum with increasing R/FR was less pronounced and showed no significance (p = 0.72). The relationships between SLA and different spectral ratios are displayed in Fig. 5. For spruce SLA no significant relationship for any of the radiative parameters could be observed (Fig. 5B). However, beech showed different characteristics (Fig. 5A). A significant relationship between decreasing SLA and increasing BW /RW (R2 = 0.74, p = 0.005) was found. A negative relationship between SLA and R/FR was significant (R2 = 0.63, p = 0.02). On the other hand GW /RW showed stronger variability within leaf biomass and therefore no statistical significance was observed (p = 0.06). The dataset for PPFR was also not significant (p = 0.08). 4. Discussion 4.1. Vertical distribution and variation of biomass Coniferous stands such as spruce have the highest values of LAIcum (Pokorny et al., 2008). This was also investigated for the modeled foliage biomass distribution and availability. Modeled LAIcum spruce values for the specific single trees in our research site were on average higher than equivalent values for beech. The relatively high SLA values for beech at a height of 14 m, where usually not a high amount of leaf biomass is found, were caused by the smaller and younger trees within the investigation site and influenced mean values only marginally and suggest secondary crop trees with sun leaf-like foliage. This small influence can also be seen in the modeled vertical LAIlay values for September. SLA for spruce corresponds well to the cone shaped morphological crown habit. Both modeled parameters represent the natural morphological shape and biomass distribution of beech and spruce as shown in validation studies (e.g. Grote and Reiter, 2004). Overall the modeled results made a good representation of the vertical distribution and morphological crown shape of both species. Foliage density varies within the canopy and depends on competition (Sinoquet and Le

Roux, 2000). Leaf area is generally considered as a major parameter and as the main attribute controlling light interception (Bartelink, 1997). 4.2. Vertical distribution of spectral ratios related to biomass Figs. 3–5 depict the influence of biomass on light interception. The vertical gradients of LAIlay and SLA show the typical course within a dense broadleaf and a clumped coniferous species. The vertical profile of PPFR showed that ca. 1% of the radiation from above the canopy reached the forest floor as shown in detail by Leuchner et al. (2007). Earlier results also show that PPFR was extinguished by absorption and reflection within the upper parts of the canopy. Leuchner et al. (2005) showed a stronger effect under clear sky conditions, while under overcast sky conditions a higher fraction of diffuse radiation penetrates into the forest stand. The higher the LAIcum , the lower the PPFR as a result of increasing shade below the crown represented by a strong negative logarithmic relationship for spruce. These excellent correlation results for spruce may result from its cone shaped morphological crown habit because radiation can penetrate continuously deeper into the stand. On the other hand, beech shows lower coefficients of determination due to the dense broadleaved canopy where most radiation is filtered and attenuated within the uppermost layer of the crown. Even the logarithmic regression was not able to represent the relationship between LAIcum and PPFR in a proper way for beech. Only the availability of sunflecks within the shade crown may influence this result (Leuchner et al., 2005, 2011). The fact of low correlations between the BALANCE-simulated competition indices and measured radiation characteristics at fixed heights may also show that some aspects of the radiation–foliage relations may be not adequately represented in the BALANCE model. In the future an implemented 3D radiation model may improve these results. Higher LAIcum values involve a greater amount of shaded leaves. With decreasing light availability, SLA of beech increased significantly from the apex downwards to the height of z/H = 0.54 (14 m) (Fig. 3A and C). Strong significant relationships between SLA and PPFR as investigated by, e.g. Chen et al. (1996), Messier and Puttonen (1995) and Walters et al. (1993) for PPDF were not supported by our findings. The spherical nature of PPFR measurements may influence this result especially under overcast sky conditions and low solar elevation angles, because photons can penetrate from all directions. The increase of BW /RW and GW /RW for beech and spruce within the sun crown, especially during the months of full foliation, was influenced by their respective morphological crown shapes. Our results show that a higher fraction of red light in comparison to blue and green light was absorbed by the sun leaves. Studies from Navrátil et al. (2007) and Woodward (1983) showed similar behavior of blue light. Reflectance and light scattering may be influenced by leaf wax and bark reflection (e.g. Grant et al., 2003). The higher overall values of BW /RW and GW /RW under spruce were likely due to the morphological crown shape, the clumping and the physiological properties of the foliage like leaf wax which is used as reflectionscattering of excessive incident radiation (Gordon et al., 1998). Below the height of the sun crown (23 m, z/H < 0.88 under beech and 20 m, z/H < 0.77 under spruce), BW /RW and GW /RW decreased in both species. This gradient could be observed in all investigated months. The result for BW /RW was also found by Woodward (1983) who described an exponential decline of B light in Buddleia davidii in shaded areas. Further the characteristics of BW /RW during CS conditions should be emphasized (Hertel et al., 2011). Direct radiation leads to a peak of the ratio within the areas of higher leaf biomass in the upper canopy. On the other side the BW /RW ratio was relatively stable in case of diffuse radiation through all stand layers (Hertel et al., 2011).

C. Hertel et al. / Agricultural and Forest Meteorology 165 (2012) 82–91

89

Fig. 5. Dependency of spectral radiation properties (BW /RW , GW /RW , PPFR, R/FR) and SLA for beech (A) and spruce (B); when no significance was existent, the regression line was omitted in the plot.

90

C. Hertel et al. / Agricultural and Forest Meteorology 165 (2012) 82–91

The increase of both spectral ratios seen in temporal scale, especially from August to September, is due to changing solar elevation as previously shown by Hertel et al. (2011). The statistical results for BW /RW showed the same characteristics for beech and spruce, and support the descriptive findings of a relationship between radiation and biomass. The BW /RW ratio showed a higher negative relationship to LAIcum for beech and displayed a steeper slope of the regression line. The shallower slope of the spruce regression line is a consequence of its morphological crown habit and allows higher BW /RW due to the possibility of more R-light absorption in lower stand layers. The GW /RW ratio as a descriptive parameter of a forest stand was statistically significant only for beech. Higher LAIcum values involve a greater amount of shaded leaves and increase the mean SLA values. With decreasing light availability, SLA of beech increased significantly from the apex downwards to the shade crown. This relationship was previously observed in many other studies (Beaudet and Messier, 1998; Chen et al., 1996). Klooster et al. (2007) and Bonosi (2006) found that the higher the shade tolerance of a species, the more pronounced the SLA increase. Our investigation supports this finding. An excellent strong negative relationship between SLA and BW /RW under beech showed the leaf level morphological response to variations in light. The course of R/FR in August showed an increase of the ratio at the forest floor under beech (Fig. 3I). This was the result of more diffuse unattenuated radiation penetrating omnidirectionally from the upper hemisphere into the forest stand and foliage gaps during more overcast radiation days compared to September (data not shown). The discoloring of the beech leaves in October and leaf fall in November/December affected the spectral ratio. Changes only occurred in the absolute values of the spectral ratios and not in their distribution of the vertical gradient. For November and December (after defoliation of beech) no significant changes in the gradient of e.g. GW /RW and R/FR under spruce were observed. The increased values of R/FR at the forest floor and, in contrast, the low overall PPFR values showed the existence of more diffuse radiation caused by the season. Changes in modeled LAIlay and SLA were BALANCE simulation results for the calculation periods within the phenological phases. Larger changes occurred under beech (Fig. 3). A totally different gradient of R/FR resulted from the missing leaf biomass in November/December. As an indicator of competition, R/FR (Leuchner et al., 2007; Lieffers et al., 1999; Ross et al., 1986) decreased less strongly than under foliated conditions, and only branches attenuated and reflected incoming radiation. R/FR and LAIcum for spruce showed the expected relationship as mentioned above as a parameter of competition, the higher the cumulated biomass the lower the R/FR. Interestingly, there is no significant relationship between LAIcum of beech and R/FR. The denser leaf distributions, the narrower crowns, and the laterally penetrated radiation at the ground layer are possible explanations. On the other hand, the R/FR displays a high correlation to beech SLA. The light environment influences the morphological and optical properties of leaves and results in the differentiation of sun and shaded leaves within the canopy. Bartelink (1997) found that beech presented a strong differentiation between both leaf categories that is represented by the SLA. The amount of sun leaves is limited to the upper crown layers. Larsen and Kershaw (1996) have also shown that a vertical foliage differentiation is much more important for light absorption than crown shape. This result was also supported by our investigations but we also showed the crucial influence of the morphological crown shape, especially for beech. On the other hand, spruce showed no significant relationship of SLA and radiation. This may be due to the morphological shape of the biomass. A clumped arrangement of needles and the vertically oriented multilayered crown allows maximized light interception. Furthermore

the differences of SLA values through all stand layers were less distinct than under beech. 5. Conclusions During seasonal changes in biomass there seems to be the potential to represent and reconstruct stand architecture of beech and spruce by different spectral ratios but correlations in this study partly remain at a quite low level. But the results clearly show differences between both species in modeled SLA distribution and radiation properties. Only for beech, a broadleaved species, was SLA explained by good statistical relationships with BW /RW and R/FR within the period of full foliation. Sun and shade leaves seemed to have a major effect on different spectral ratios, e.g. BW /RW . This differentiation is crucial for a description of the biomass distribution of different species within the crown. Results for LAIcum reveal the possibility of the representation of this biomass parameter for both species. Spruce showed the best results with PPFR and R/FR and less distinctly for BW /RW . In contrast to spruce, beech was represented best by BW /RW and GW /RW as a result of higher reflection of B- and G-light compared to R-light by the leaves. It is evident that the increase of BW /RW and GW /RW is shown in the sun crown. For both species, the BW /RW ratio and PPFR in combination with LAI were appropriate tools to describe stand architectures. The present results are useable, but further improvement and development of new and better models may increase the fit of both parameters. There is also a relationship between crown level morphology and light variation. Temporal variations in SLA and LAI during leaf fall influence light interception and show morphological crown changes, especially under broadleaved species. SLA should be determined not only at one particular crown location because an SLA profile can be capable of representing the stand morphology. Acknowledgements This project was part of the Sonderforschungsbereich 607 (SFB 607) funded by the Deutsche Forschungsgemeinschaft. We thank Tim Sparks for editing the language of our manuscript and two anonymous reviewers for insightful comments and suggestions that helped improve the quality of the manuscript. References Abrams, M.D., Mostoller, S.A., 1995. Gas exchange, leaf structure and nitrogen in contrasting successional tree species growing in open and understory sites during a drought. Tree Physiol. 15, 361–370. Bartelink, H.H., 1997. Allometric relationships for biomass and leaf area of beech (Fagus sylvatica L.). Ann. For. Sci. 54, 39–50. Beaudet, M., Messier, C., 1998. Growth and morphological responses of yellow birch, sugar maple, and beech seedlings growing under a natural light gradient. Can. J. For. Res. 28, 1007–1015. Björn, L.O., Vogelmann, T.C., 1996. Quantifying light and ultraviolet radiation in plant biology. Photochem. Photobiol. 64, 403–406. Bonosi, L., 2006. The influence of light and size on photosynthetic performance, light interception, biomass partitioning and tree architecture in open grown Acer pseudoplatanus, Fraxinus excelsior and Fagus sylvatica seedlings. Schriftenr. Freiburger Forstliche Forsch. 34. Capers, R.S., Chazdon, R.L., 2004. Rapid assessment of understory light availability in a wet tropical forest. Agric. For. Meteorol. 123, 177–185. Campbell, G.S., Norman, J.M., 1998. An Introduction to Environmental Biophysics. Springer, Heidelberg. Cemark, J., 1998. Leaf distribution in large trees and stands of the floodplainforest southern Moravia. Tree Physiol. 18, 727–737. Chen, J.M., 1996. Optically-based methods for measuring seasonal variation of leaf area index in boreal conifer stands. Agric. For. Meteorol. 80, 135–163. Chen, H.Y.Y., Klinka, K., Kayahara, G.J., 1996. Effects of light on growth, crown architecture, and specific leaf area for naturally established Pinus contorta var. latifolia and Pseudotsuga menziesii var. glauca saplings. Can. J. For. Res. 26, 1149–1157. Endler, J.A., 1993. The color of light in forests and its implications. Ecol. Monogr. 63, 1–27.

C. Hertel et al. / Agricultural and Forest Meteorology 165 (2012) 82–91 Federer, C.A., Tanner, C.B., 1966. Spectral distribution of light in the forest. Ecology 47, 555–560. Gordon, D.C., Percy, K.E., Riding, R.T., 1998. Effects of UV-B radiation on epicuticular wax production and chemical composition of four Picea species. New Phytol. 138, 441–449. Grant, R.H., 1997. Partitioning of biologically active radiation in plant canopies. Int. J. Biometeorol. 40, 26–40. Grant, R.H., Heisler, G.M., Gao, W., Jenks, M., 2003. Ultraviolet leaf reflectance of common urban trees and the prediction of reflectance from leaf surface characteristics. Agric. For. Meteorol. 120, 127–139. Grote, R., Pretzsch, H., 2002. A model for individual tree development based on physiological processes. Plant Biol. 4, 167–180. Grote, R., Reiter, I.M., 2004. Competition-dependent modeling of foliage biomass in forest stands. Trees 18, 596–607. Häberle, K.H., Reiter, I.M., Nunn, A.J., Gruppe, A., Simon, U., Gossner, M., Werner, H., Leuchner, M., Heerdt, C., Fabian, P. Matyssek, R., 2003. KROCO, Freising, Germany: Canopy research in a temperate mixed forest of Southern Germany. In: Basset, Y., Horlyck, V., Wright, S.J. (Eds.), Studying forest canopies from above: The International Canopy Crane Network. Smithsonian Tropical Research Institute and UNEP, Panama, pp. 71–78. Haxeltine, A., Prentice, I.C., 1996. A general model for the light-use efficiency of primary production. Funct. Ecol. 10, 551–561. Hertel, C., Leuchner, M., Menzel, A., 2011. Vertical variability of spectral ratios in a mature mixed forest stand. Agric. For. Meteorol. 151, 1096–1105. Holmes, M.G., Smith, H., 1977. The function of phytochrome in the natural environment. 2. The influence of vegetation canopies on spectral energy distribution of natural daylight. Photochem. Photobiol. 25, 539–545. Kazda, M., Salzer, J., Reiter, I.M., 2000. Photosynthetic capacity in relation to nitrogen in the canopy of a Quercus robur, Fraximus angustifolia and Tilia cordata flood plain forest. Tree Physiol. 20, 1029–1037. Klooster, S.H.J.T., Thomas, E.J.P., Sterck, F.J., 2007. Explaining interspecific differences in sapling growth and shade tolerance in temperate forests. Ecology 95, 1250–1260. Landhäuser, S.M., Lieffers, V.J., 2001. Photosynthesis and carbon allocation of six boreal tree species grown in understory and open conditions. Tree Physiol. 21, 243–250. Larsen, D.R., Kershaw, J.A., 1996. Influence of canopy structure assumptions on predictions from Beer’s law. A comparison of deterministic and stochastic simulations. Agric. For. Meteorol. 81, 61–77. Leuchner, M., Fabian, P., Werner, H., 2005. Spectral multichannel monitoring of radiation within a mature mixed forest. Plant. Biol. 7, 619–627. Leuchner, M., Menzel, A., Werner, H., 2007. Quantifying the relationship between light quality and light availability at different phenological stages within a mature mixed forest. Agric. For. Meteorol. 142, 35–44. Leuchner, M., Hertel, C., Menzel, A., 2011. Spatial variability of photosynthetically active radiation in European beech and Norway spruce. Agric. For. Meteorol. 151, 1226–1232. Lieffers, V.J., Messier, C., Stadt, K.J., Gendron, F., Comeau, P.G., 1999. Predicting and managing light in the understory of boreal forests. Can. J. For. Res. 29, 796–811. Marek, M.V., Sprtova, M., Urban, O., Spunda, L., Kalina, J., 1999. Response of sun versus shade foliage photosynthesis to radiation in Norway spruce. Phyton 39, 131–138. Matsuda, R., Ohashi-Kaneko, K., Fukiwara, K., Goto, E., Kurata, K., 2007. Analysis of the relationship between blue-light photon flux density and the photosynthetic properties of spinach (Spinacia oleracea L.) leaves with regard to the acclimation of photosynthesis to growth irradiance. Soil Sci. Plant Nutr. 53, 459–465. Matyssek, R., 1986. Carbon, water and nitrogen relations in evergreen and deciduous conifers. Tree Physiol. 2, 177–187. Messier, C., Puttonen, P., 1995. Spatial and temporal variation in the light environment of developing Scots pine stands: the basis for a quick and efficient method of characterizing light. Can. J. For. Res. 25, 343–354. Morales, D., Gonzalez-Rodriguez, A.M., Cermak, J., Jimenez, M.S., 1996. Laurel forests in Tenerife, Canary Island: the vertical profiles of leaf characteristics. Phyton (Austria) 36, 251–263. Mäkela, A., 1990. Modelling structural–functional relationships in whole-tree growth: resource allocation. In: Dixon, R.K., Meldahl, R.S., Ruark, G.A., Warren, W.G. (Eds.), Process Modelling of Forest Growth Responses to Environmental Stress. Timber Press, Portland, OR, USA, pp. 81–95. Navrátil, M., Spunda, V., Marková, I., Janous, D., 2007. Spectral composition of photosynthetically active radiation penetrating into a Norway spruce canopy: the

91

opposite dynamics of the blue/red spectral ration during clear and overcast days. Trees 21, 311–320. Niinemets, Ü., Kull, O., Tenhunen, J.D., 1999. Variability in leaf morphology and chemical composition as a function of canopy light environment in coexisting deciduous trees. Int. J. Plant Sci. 160, 837–848. Petritian, A.M., von Lüpke, B., Petritian, I.C., 2009. Influence of light availability on growth, leaf morphology and plant architecture of beech (Fagus sylvatica L.), maple (Acer pseudoplatanus L.) and ash (Fraxinus excelsior L.) saplings. Eur. J. Forest Res. 128, 61–74. Pecot, S.D., Horsley, S.B., Battaglia, M.A., Mitchell, R.J., 2005. The influence of canopy, sky condition, and solar angle on light quality in a longleaf pine woodland. Can. J. For. Res. 35, 1356–1366. Pokorny, R., Marek, M.V., 2000. Test of accuracy of LAI estimation by LAI-200 under artificially changed leaf to wood area proportions. Biol. Plantarum 43, 537–544. Pokorny, R., Tomaskova, I., Havrankova, K., 2008. Temporal variation and efficiency of leaf area index in young mountain Norway spruce stand. Eur. J. Forest Res. 127, 359–367. Pretzsch, H., Kahn, M., Grote, R., 1998. Die Fichten-Buchen-Mischbestände des Sonderforschungsbereiches “Wachstum oder Parasitenabwehr?” im Kranzberger Forst. Forstw. Cbl. 117, 241–257. Reiter, I.M., 2004. Space-related resource investments and gains of adult beech (Fagus sylvatica) and spruce (Picea abies) as a quantification of aboveground competitiveness. Dissertation. Munich. Ross, M.S., Flanagan, L.B., La Roi, G.H., 1986. Seasonal and successional changes in light quality and quantity in the understory of boreal forest ecosystems. Can. J. Bot. 64, 2792–2799. Rötzer, T., Grote, R., Pretzsch, H., 2004. The timing of bud burst and its effect on tree growth. Int. J. Biometeorol. 48, 109–118. Rötzer, T., Grote, R., Pretzsch, H., 2005. Effects of environmental changes on the vitality of forest stands. Eur. J. Forest Res. 124, 349–362. Rötzer, T., Seifert, T., Pretzsch, H., 2009. Modelling above and below ground carbon dynamics in a mixed beech and spruce stand influenced by climate. Eur. J. Forest Res. 128, 171–182. Rötzer, T., Leuchner, M., Nunn, A.J., 2010. Simulating stand climate, phenology, and photosynthesis of a forest stand with a process based growth model. Int. J. Biometeorol. 54, 449–464. ˜ J., 2005. Assessing forest structure and function from spectral Serrano, L., Penuelas, transmittance measurements: a case study in a Mediterranean holm oak forest. Tree Physiol. 25, 67–74. Shinozaki, K., Yoda, K., Hozumi, K., Kira, T., 1964. A quantitative analysis of plant form – the pipe model theory. I. Basic analyses. Jpn. J. Ecol. 14, 97–105. Sinoquet, H., Le Roux, X., 2000. Short term interactions between tree foliage and the aerial environment: an overview of modelling approaches available for tree structure–function models. Ann. For. Sci. 57, 477–549. Smith, H., Morgan, D.C., 1981. The spectral characteristics of the visible radiation incident upon the surface of the earth. In: Smith, H. (Ed.), Plants and the Daylight Spectrum. Academic Press, London, New York, Toronto, Sydney, San Francisco, pp. 3–20. Urban, O., Kosvancova, M., Marek, M.V., Lichtenthaler, H.K., 2007. Induction of photosynthesis and importance of limitations during the induction phase in sun and shade leaves of five ecologically contrasting tree species from temperate zone. Tree Physiol. 27, 1207–1215. Vogelmann, T.C., Martin, G., 1993. The functional significance of palisade tissue: penetration of directional versus diffuse light. Plant Cell Environ. 16, 65–72. Walters, M.B., Kruger, E.L., Reich, P.B., 1993. Growth, biomass distribution and CO2 exchange of northern hardwood seedlings in high and low light: relationships with successional status and shade tolerance. Oecologia 94, 7–16. Wang, Q., Tenhunen, J., Granier, A., Reichstein, M., Bouriaud, O., Nguyen, D., Breda, N., 2004. Long-term variations in leaf area index and light extinction in a Fagus sylvatica stand as estimated from global radiations profiles. Theor. Appl. Climatol. 79, 225–238. Watson, D.J., 1947. Comparative physiological studies in the growth of field crops. I: variation in net assimilation rate and leaf area between species and varieties, and within and between years. Ann. Bot. 11, 41–76. Wipfler, P., Seifert, T., Heerdt, C., Werner, H., Pretzsch, H., 2005. Growth of adult Norway Spruce (Picea abies [L.] Karst.) and European Beech (Fagus sylvatica L.) under free-air ozone fumigation. Plant. Biol. 7, 611–618. Woodward, F.I., 1983. Instruments for the measurement of photosynthetically active radiation and red, far-red and blue light. J. Appl. Ecol. 20, 103–115.