Oecologia

, Volume 169, Issue 4, pp 915–925 | Cite as

Variation in foliar nitrogen and albedo in response to nitrogen fertilization and elevated CO2

  • Haley F. Wicklein
  • Scott V. Ollinger
  • Mary E. Martin
  • David Y. Hollinger
  • Lucie C. Lepine
  • Michelle C. Day
  • Megan K. Bartlett
  • Andrew D. Richardson
  • Richard J. Norby
Physiological ecology - Original research

Abstract

Foliar nitrogen has been shown to be positively correlated with midsummer canopy albedo and canopy near infrared (NIR) reflectance over a broad range of plant functional types (e.g., forests, grasslands, and agricultural lands). To date, the mechanism(s) driving the nitrogen–albedo relationship have not been established, and it is unknown whether factors affecting nitrogen availability will also influence albedo. To address these questions, we examined variation in foliar nitrogen in relation to leaf spectral properties, leaf mass per unit area, and leaf water content for three deciduous species subjected to either nitrogen (Harvard Forest, MA, and Oak Ridge, TN) or CO2 fertilization (Oak Ridge, TN). At Oak Ridge, we also obtained canopy reflectance data from the airborne visible/infrared imaging spectrometer (AVIRIS) to examine whether canopy-level spectral responses were consistent with leaf-level results. At the leaf level, results showed no differences in reflectance or transmittance between CO2 or nitrogen treatments, despite significant changes in foliar nitrogen. Contrary to our expectations, there was a significant, but negative, relationship between foliar nitrogen and leaf albedo, a relationship that held for both full spectrum leaf albedo as well as leaf albedo in the NIR region alone. In contrast, remote sensing data indicated an increase in canopy NIR reflectance with nitrogen fertilization. Collectively, these results suggest that altered nitrogen availability can affect canopy albedo, albeit by mechanisms that involve canopy-level processes rather than changes in leaf-level reflectance.

Keywords

Albedo Nitrogen Leaf structure Nitrogen fertilization Free air CO2 enrichment 

Introduction

The relationship between mass-based foliar nitrogen (N) and leaf-level photosynthetic capacity has been globally documented across a wide range of plant species (Field and Mooney 1986; Evans 1989; Reich et al. 1997, 1999; Wright et al. 2004), and a similar trend has been observed at the canopy level for boreal and temperate forests (Ollinger et al. 2008). Recently, Ollinger et al. (2008) and Hollinger et al. (2010) demonstrated that both of these variables are also significantly and positively correlated with full spectrum canopy albedo. Given the importance of even small changes in albedo on surface heat exchange, the occurrence of an N effect on albedo would bear interesting and potentially important consequences for climate. However, resolving underlying mechanisms of the N–albedo relationship will be important because different mechanisms can carry different implications for a response to altered N availability.

Sources of variation in vegetation albedo occur at scales ranging from internal leaf structures to tree crowns and whole plant canopies (Ollinger 2011). At the leaf level, reflectance is dominated by photosynthetic pigments in the visible part of the spectrum (400–700 nm), leaf structure in the near infrared (NIR, 700–1,350 nm), and water content in the mid-infrared (Mid IR, 1,350–2,500 nm) (Gates et al. 1965; Slaton et al. 2001; Jacquemoud et al. 2009). At the stem and canopy level, the number of scattering or absorbing surfaces that photons encounter is further influenced by structural traits such as shoot architecture, leaf angle distribution, and crown geometry (e.g., Chen and Cihlar 1995; Asner 1998; Rautiainen et al.2004). The greatest variation in reflectance at both leaf and canopy levels often occurs in the NIR region because there are few, if any, compounds that absorb NIR radiation (Gates et al. 1965; Sánchez and Canton 1999; Ollinger 2011). Therefore, if structural changes that occur within or between leaves covary with foliar N, they could explain the nature of the canopy-level N–albedo relationship.

Hollinger et al. (2010) hypothesized that the association between leaf N concentration (Nmass) and canopy albedo may be due to covariation between N-containing photosynthetic enzymes and internal leaf structures necessary to support different rates of photosynthesis. Particularly important are changes in the ratio of the mesophyll surface area exposed to intercellular air spaces to the area of the leaf (Ames/A), which has been shown to be positively correlated with both photosynthetic rates (Nobel et al. 1975; Longstreth et al. 1985) and NIR reflectance (Slaton et al.2001). Given the different refractive indices of hydrated mesophyll cells and the intercellular airspace (Woolley 1971; Gausman et al. 1974), a higher Ames/A value should lead to more opportunities for radiation scattering, and correspondingly higher reflectance.

The objectives of this study were (1) to test the hypothesis that the canopy level association between foliar Nmass and canopy albedo stems from a similar relationship occurring at the leaf level, and (2) to determine whether two forms of disturbance that are known to affect N availability—N fertilization and elevated CO2—can also influence leaf and canopy albedo. We measured leaf reflectance and transmittance, as well as leaf chemical and structural traits (N concentration, leaf mass per area, equivalent water thickness, and water content), for three deciduous species in the eastern US that have been subjected to either long-term N or CO2 fertilization. Data from multiple free air CO2 enrichment (FACE) sites have shown that elevated CO2 leads to an increase in leaf mass per unit area (LMA, Norby et al. 2003) and the consequent dilution of foliar N due to the accumulation of carbohydrates (Oren et al. 2001; Ellsworth et al. 2004; Norby and Iversen 2006). Therefore, if the canopy albedo–N relationship is driven by changes at the leaf level, we would expect to see higher leaf-level leaf albedo in plots subjected to N fertilization and lower leaf albedo in plots exposed to elevated CO2, relative to those receiving ambient CO2 and N deposition. We also obtained canopy reflectance data from the airborne visible/infrared imaging spectrometer (AVIRIS) to examine whether canopy-level spectral responses were consistent with leaf-level results.

Materials and methods

Study sites

We measured spectral, chemical, and structural characteristics of leaf samples from two sites in the eastern US: Harvard Forest, Petersham, MA (42.5°N, 72°W) and Oak Ridge National Laboratory, Roane County, TN (35.9°N, 84.3°W). These sites were chosen because they contain deciduous tree species relevant to our objectives and because they allowed us to examine the spectral response of leaves to N fertilization and elevated CO2, both of which represent important environmental change agents that are known to alter leaf Nmass.

Harvard Forest, MA

Harvard Forest (HF) is located in Central Massachusetts and has been a long-term ecological research site since 1988. A chronic N fertilization experiment was established at HF in 1988 (Magill et al. 2004). Two stands were chosen for N additions: a mixed hardwood stand that regenerated naturally after a clearcut around 1945, and an even-aged red pine (Pinus resinosa Aiton) stand that was heavily disturbed by an ice storm in December of 2008, and not used in this study. In each stand, four plots were established: control (no added N), low N (additions of 50 kg N ha−1 year−1), low N plus sulfur (not included in this study), and high N (additions of 150 kg N ha−1 year−1). Each plot measures 30 × 30 m. Additions of ammonium nitrate (NH4NO3) began in 1988 and are distributed over six equal applications during the growing season (May–September). Mean annual precipitation at HF is 1,100 mm, distributed evenly throughout the year. Ambient N deposition averages 8 kg N ha−1 year−1 (Ollinger et al. 1993). The dominant soil types are stony- to sandy-loams formed from glacial till. Elevation is 385 m above sea level.

Oak Ridge National Environmental Research Park, TN

The free air CO2 enrichment (FACE) facility at the Oak Ridge National Laboratory (ORNL) Environmental Research Park is located in a sweetgum (Liquidambar styraciflua L.) monoculture that was established in 1988 (Norby et al. 2001). In 1996, five 25-m-diameter plots were established: two with FACE apparatus emitting elevated CO2, two with FACE apparatus but ambient CO2, and one with no FACE apparatus. Exposure to elevated CO2 began in the spring of 1998. Average daytime CO2 concentration for 2009 was maintained at approximately 565 ppm for the enriched plots and 401 ppm for the ambient plots (Riggs et al. 2009). In 2004, an N fertilization experiment was initialized in a sweetgum stand located approximately 150 m from the FACE site and planted at the same time. An 85 m × 50 m area was fertilized in a block pattern, with each block containing two control plots and two fertilized plots (both 12 × 16 m). The fertilized plots received 200 kg N ha−1 year−1, applied as urea each year before leaf out (Iversen and Norby 2008). Mean annual temperature at Oak Ridge is 14°C, and mean annual precipitation is 1,371 mm, distributed evenly throughout the year. Ambient N deposition averages between 10 and 15 kg N ha−1 year−1 (Johnson et al. 2004). The dominant soil type is an Aquic Hapludult, a silty clay loam.

Data collection and analysis

Field sampling: Harvard Forest

At HF, sampling of foliage from the control and treated hardwood plots was conducted between 20 and 23 July 2009. Within each plot, five red maple and seven black oak trees were randomly selected and sampled. Green leaves were collected from the top, middle, and bottom of the canopy using a 12-gauge shotgun. Sample collection heights were determined using a digital hypsometer (Haglöf Vertex). Leaves were placed in plastic Ziploc bags and stored on ice until analysis, which was carried out within 36 h of collection.

Field sampling: Oak Ridge National Environmental Research Park

At ORNL, field sampling was conducted between 28 and 30 July 2009. Using a slingshot canopy sampler (N-fertilized site) or tower climbing with pole pruners (FACE site), we collected green leaves from the top, middle, and bottom of the canopy. Heights were determined using either a digital hypsometer (Haglöf Vertex) or measuring tapes deployed by climbers. Within each CO2 fertilization treatment (ambient and elevated CO2), we sampled ten sweetgum trees. From each N-fertilized plot, we collected 12 upper canopy and 12 lower canopy samples. The adjacent FACE ambient CO2 plots were used as the control treatment for the N-fertilized samples in statistical analysis. In all cases, leaves were placed in plastic Ziploc bags and stored on ice until analysis.

Collection of field spectra

We measured hemispherical reflectance and transmittance spectra for healthy leaves from each individual using a portable spectrometer (ASD FieldSpec 3; analytical Spectral Devices, Boulder, CO, USA) connected to an integrating sphere with an 8° near-normal incidence port to capture both diffuse and specular reflectance (SphereOptics, Concord, NH, USA) and a halogen bulb light source. The ASD spectrophotometer was used in conjunction with an integrating sphere, which holds leaf samples in a fixed position and is intended to eliminate angular effects by creating perfectly diffuse radiation. It measures reflectance from 350 to 2,500 nm, in 1-nm intervals. For each sample, measurements were taken for a single leaf and stacks of two, four, and eight leaves. Reflectance and transmittance spectra of leaf stacks were taken to create optically dense layers of leaf tissue, in order to examine how reflectance varies as a function of the total thickness of the leaf material present. Each leaf spectrum was determined as the average of 50 individual scans. The spectra were corrected for dark current, and a white reference standard was measured prior to each set of reflectance or transmittance measurements of one growing stack of leaves. For reflectance measurements, both the sample and the reference standard were mounted on the sphere and their positions were switched between two successive scans. Spectral reflectance was then calculated as the ratio of the value obtained from the sample to the value obtained from the reference.

To calculate full spectrum (FULL, 400–2,500 nm) spectrally weighted optical values, each reflectance or transmittance spectra was weighted by the solar spectrum energy (ASTM G173-03 Reference Spectra received by a 37° tilted equator-facing surface through an air mass of 1.5, derived from SMARTS v. 2.9.2; Gueymard 2004) to obtain a value representing the reflected (denoted as αFULL) or transmitted (denoted as τFULL) energy as a proportion of incident. These values approximate albedo values measured using broad-band radiometric instruments. For visible (αvis or τvis, 400–700 nm), NIR (αnir or τnir, 700–1,300 nm), and Mid IR (αmidir or τmidir, 1,350–2,500 nm) portions of the spectrum, the solar-weighted reflectance and transmittance values were calculated by following the same process, but using only wavelengths from the regions of interest.

Leaf chemical and structural analysis

After spectral measurements were obtained, two circular disks (2.035 cm2 area) were removed from each leaf and weighed. Leaf disks were then oven-dried at 70°C for at least 72 h and reweighed to determine water content (% leaf fresh weight), equivalent water thickness (EWT, g water per cm2 leaf), and LMA (g leaf per m2 leaf).

The remaining leaf sample was then dried at 70°C, ground using a Wiley mill, and passed through a 1-mm mesh screen. Prior to N concentration analysis, the ground samples were dried for 24 h at 70°C. We measured mass-based foliar N concentration (Nmass, g of N per 100 g of dry leaf matter) using a Costech Elemental Analyzer. We determined N per unit leaf area (Narea) by multiplying Nmass by the LMA of the sample (Narea = Nmass × LMA).

Aircraft remote sensing

For the ORNL site, high spectral resolution remote sensing images were acquired from NASA’s AVIRIS (Airborne Visible/InfraRed Imaging Spectrometer) instrument within 10 days of our field sampling. AVIRIS measures upwelling radiation in 224 contiguous optical bands from 400 to 2,500 nm with a spectral resolution of 10 nm. AVIRIS was flown on an ER-2 at an altitude of 5,500 m giving a spatial resolution of 4.6 m. The image dataset was orthorectified by NASA JPL, and atmospherically corrected with ImSpec LLC’s Atmospheric Correction Now (ACORN) v.6, transforming data from calibrated sensor radiance to apparent surface reflectance. After removing strong water adsorption bands that contained no usable data, we determined canopy reflectance for each treatment by averaging reflectance values within each waveband for pixels falling within each of the plots. Due to the spatial resolution of the image and the size of the treatment plots, we were able to obtain 36 pixels in the ambient CO2 plot, 12 pixels in the elevated CO2 plot, and 24 pixels in the N-fertilized.

Statistical analysis

Summary statistics (means and standard error) were computed for all optical properties and leaf traits. The significances of the mean differences between treatments were determined by analysis of variance (ANOVA), with pair-wise comparisons tested using Tukey’s ‘Honest Significant Difference’ method. Regression analysis was used to determine relationships between optical properties and leaf traits. For multiple regressions the adjusted r2 was considered instead of r2 because this statistic penalizes the model for an increased number of parameters, thereby decreasing the likelihood of overfitting.

The Shapiro–Wilk statistic was used to test for normality in all models, and where needed variables were power-transformed to correct for skew. Single leaf αFULL and αnir were non-normal due to a small number of outliers (determined by 1.5 × interquartile range). Statistical tests were performed with and without the outliers, and, because the resulting explained variances were similar, outliers were removed to follow assumptions of normality in linear regression models. All statistical analysis was completed using the software R, v.2.8.1 (R Foundation for Statistical Computing, 2008). Reported results are for single leaves unless otherwise specified.

Results

Treatment differences

Across all species and sites, Nmass was higher in the high N treatment than in the low N and control treatments, which were not significantly different from each other (P < 0.001; Fig. 1a–c). For red maple and black oak at HF, there were no differences between nitrogen treatments in αFULL (Fig. 1d, e), τFULL, αnir, τnir, αmidir, τmidir, or FULL absorption [1 − (α + τ)] (P > 0.1 in all cases). For the sweetgum at ORNL, there were no differences between treatments in αFULL (except for higher mean values in the ORNL elevated CO2 treatment than ORNL N-fertilized treatment; Fig. 1f), τFULL, αnir, or FULL absorption (P > 0.1 in all cases). There was higher τnir in the ORNL N-fertilized treatment than in the ambient or elevated CO2 treatments (P < 0.01). Both αmidir and τmidir were higher for N-fertilized sweetgums (P < 0.01), likely due to differences in LMA and EWT. For all species, αvis and τvis followed differences in N between treatments, with lower values corresponding to higher N fertilization (P < 0.05 in all cases). There was no difference in LMA between nitrogen treatments for black oak or red maple (P > 0.1; Fig. 1g, h). However, the N-fertilized treatment at ORNL had lower LMA than the ambient or elevated CO2 treatments (P < 0.001; Fig. 1i). EWT did not differ between treatments for any species (P > 0.1).
Fig. 1

ANOVA results (means ± standard error) for Nmass (ac), solar weighted full spectrum reflectance (αFULL; df), and LMA (gi) treatment differences. Results for black oak (site: HF) are depicted with black bars, red maple (site: HF) in gray bars, and sweetgum (site: ORNL) in white bars. Means with different letters were significantly different in pair-wise comparisons (Tukey’s multiple comparison test)

Relationships between N and leaf optical properties

The relationships between Nmass or Narea and the α and τ of all optical regions were qualitatively similar across sites and treatments, although Narea explained less of the variance in foliar optical properties than did Nmass (Table 1). Variation in Narea can be caused by changes in LMA or Nmass, although Nmass and LMA generally exhibited opposite trends (e.g., LMA is positively correlated with αvis, while the relationship between Nmass and αvis is negative). Thus, it is likely that the similarities between the Nmass and Narea relationships are driven by changes in Nmass, but dampened by the opposing effects of LMA. Because the changes in Narea can be accounted for by these two variables, subsequent analysis and discussion of foliar N and optical properties will focus on Nmass. Furthermore, Nmass was not related to canopy position (height from which foliage was sampled) for any of the species or treatments in our analysis (P > 0.1 in all cases), except for a weak, negative correlation in the elevated sweetgum treatment (r2 = 0.19, P < 0.05). As a result, we used mean values from all heights within a tree for subsequent analyses involving Nmass.
Table 1

Comparison of regression statistics for spectrally weighted reflectance and transmittance foliar N on a mass (Nmass, %) and area basis (Narea, g m−2), reporting the coefficient of determination (r2), P value, and the sign of the slope of the regression line (‘Sign’ column) for each model

Response variable

Source of variation

r2

Nmass

r2

Narea

P value

Sign

P value

Sign

FULL reflectance

0.26

<0.001

0.15

<0.001

FULL transmittance

 

ns

  

ns

 

NIR reflectance

0.17

<0.001

0.06

<0.01

NIR transmittance

0.10

<0.001

+

0.05

<0.01

+

Mid IR reflectance

 

ns

 

0.09

<0.001

Mid IR transmittance

0.21

<0.001

+

 

ns

 

VIS reflectance

0.51

<0.001

0.17

<0.001

VIS transmittance

0.45

<0.001

0.31

<0.001

Reflectance and transmittance values are all weighted by the solar spectrum

ns insignificant trends

Across all species, Nmass was the best predictor of FULL, NIR, and visible α and τ for all species combined (Fig. 2), with the exception of τFULL, which was not correlated with Nmass (P > 0.05). However, contrary to our expectations, the relationships between Nmass and α were negative (Fig. 2a–c; P < 0.001 in all cases). There was a positive relationship between both τFULL and τnir and Nmass, whereas the relationship between Nmass and τvis was negative (Fig. 2d–f; P < 0.001 in all cases). The relationships between Nmass and α in the FULL and NIR regions are primarily due to differences between species [within species, these relationships were not significant (P > 0.1)], with the exception of sweetgum αFULL which showed a slight negative relationship with Nmass (P < 0.05, r2 = 0.06), whereas the relationships between Nmass and the visible region are significant even within species (P < 0.05 in all cases).
Fig. 2

Regression between foliar Nmass and a solar-weighted full spectrum reflectance (αFULL; 400–2,500 nm), b solar-weighted NIR reflectance (αnir; 700–1,300 nm), and c solar-weighted visible reflectance (αvis; 400–700 nm; data log-transformed), d solar-weighted full spectrum transmittance (τFULL), e solar-weighted NIR transmittance (τnir), and f solar-weighted visible transmittance (τvis; data log-transformed). The nonlinear relationship between Nmass and both αvis and τvis appears to be due to differences in slope between species that vary in their N concentrations. All correlations are significant at P < 0.001. Black oak (BO), red maple (RM) and sweetgum (SG) were all included in the regression analysis

Although Nmass was not a good predictor of αmidir or τmidir, the relationship was significant and positive for τmidir (r2 = 0.21, P < 0.001). There was no relationship between Nmass and αmidir (P > 0.1). Absorption was positively correlated with Nmass in the visible (r2 = 0.52, P < 0.001) and the Mid IR (r2 = 0.17, P < 0.001), but not in the NIR (P > 0.1), leading to a weak, positive relationship between total FULL absorption and Nmass (r2 = 0.07, P < 0.01). There was some colinearity between spectral response variables, which is not surprising, as they are all part of the same overall spectra. However, the different regions responded differently, because absorbers are located primarily in the visible (pigment absorption) and Mid IR (water absorption), with the NIR being primarily influenced by scattering (Ollinger 2011).

The above results describe how Nmass varied with α and τ for a single leaf. In leaf stacks, the correlation coefficients for the relationships between Nmass and foliar optical properties decreased as the number of leaves increased. Despite this, the negative slopes of the relationship between Nmass and αFULL remained intact for all leaf stacks (Fig. 3; P < 0.001 in all cases).
Fig. 3

Foliar Nmass versus solar weighted full spectrum reflectance (αFULL) for stacks of one (r2 = 0.25, P < 0.001), two (r2 = 0.18, P < 0.001), four (r2 = 0.15, P < 0.001), and eight (r2 = 0.10, P < 0.001) leaves. Black oak (BO), red maple (RM) and sweetgum (SG) were all included in the regression analysis. Although the y-intercept differs, all stacks show similar negative relationships between foliar N and αFULL

Relationships between leaf traits and leaf optical properties

Across all species and treatments, there was a weak, positive correlation between LMA and αFULL (P < 0.05), and weak, negative correlation between LMA and both τFULL and τnir (Fig. 4; P < 0.001 in both cases). There were stronger negative correlations between LMA and αmidir (r2 = 0.38, P < 0.001) and τmidir (r2 = 0.40, P < 0.001), although this is likely due to the influence of EWT. LMA was positively correlated with αvis (r2 = 0.20, P < 0.001), whereas there was no relationship between τvis and LMA (P > 0.1).
Fig. 4

Regression between LMA and a solar-weighted full spectrum reflectance (αFULL; P < 0.05), b solar-weighted NIR reflectance (αnir; P > 0.1), c solar-weighted full spectrum transmittance (τFULL; P < 0.001), and d solar-weighted NIR transmittance (τnir; P < 0.001). Black oak (BO), red maple (RM) and sweetgum (SG) were all included in the regression analysis

Equivalent water thickness was negatively correlated with αmidir and τmidir and was the best overall predictor of Mid IR foliar optical properties (Fig. 5; P < 0.001 in all cases). EWT was weakly, but positively, correlated with αFULL (r2 = 0.06, P < 0.01), αnir (r2 = 0.04, P < 0.01), and αvis (r2 = 0.28, P < 0.001), and negatively correlated with τFULL (r2 = 0.10, P < 0.001) and τnir (r2 = 0.09, P < 0.001). Relative water content (g H2O g−1 fresh leaf tissue) was not correlated with any foliar optical parameter we considered (P > 0.1 in all cases).
Fig. 5

Both a solar-weighted reflectance (α) and b solar-weighted transmittance (τ) declined in the Mid IR region (700–1,350 nm) as leaf EWT increased. This was due to the strong positive correlation between Mid IR absorption [1 − (α + τ)] and EWT (c), which demonstrates the importance of water abosrption in the Mid IR in influencing scattering from this region. All correlations are significant at P < 0.001. Black oak (BO), red maple (RM) and sweetgum (SG) were all included in the regression analysis

Multiple regression models using all combinations of leaf traits did not improve predictions for αFULL, τFULL, αnir, or τnir above that which was obtained by Nmass alone.

Relationships between leaf traits

The relationship between LMA and EWT was significant and positive both within (black oak: r2 = 0.73; red maple: r2 = 0.36; sweetgum: r2 = 0.46; P < 0.001 in all cases) and across (r2 = 0.54, P < 0.001) all species. LMA and Nmass were negatively correlated within sweetgums (r2 = 0.22, P < 0.001), but the relationship was not significant in black oaks, red maples, or across all species. EWT and Nmass were not significantly correlated within any species (P > 0.1 in all cases), but with all species combined they were slightly negatively related (r2 = 0.05, P < 0.01).

Canopy-level reflectance

Despite the absence of differences in leaf-level reflectance between treatments, reflectance data obtained from the AVIRIS sensor showed that whole-canopy NIR reflectance was higher for the N-fertilized sweetgum treatment than the ambient CO2 sweetgum treatment (Fig. 6; P < 0.001). Canopy NIR reflectance from the elevated CO2 sweetgum treatment was lower than the ambient CO2 sweetgum treatment, although the trend was not significant (Fig. 6; P > 0.1) and the difference was small compared to that between the N-fertilized and ambient CO2 treatments. There was little to no difference in canopy reflectance in the visible or MIR portions of the spectrum.
Fig. 6

AVIRIS canopy reflectance across the full spectrum for N-fertilized, ambient CO2, and elevated CO2 sweetgum treatments at ORNL

Discussion

A primary goal of this study was to examine relationships between leaf traits and leaf optical properties that might help explain the positive correlation between canopy Nmass and full spectrum canopy albedo seen in temperate and boreal forests (Ollinger et al. 2008; Hollinger et al. 2010). Our results do not support the hypothesis that the canopy-level trend was caused by a similar trend occurring at the leaf level. There were no differences in leaf αFULL or τFULL between comparable N or CO2 treatments, despite changes in leaf Nmass caused by N fertilization and increases in whole canopy NIR reflectance, as measured using AVIRIS, in the N-fertilized plots at ORNL. Earlier work showed a reduction in leaf Nmass in CO2-treated plots relative to controls at ORNL (Norby and Iversen 2006). Although we also saw lower Nmass in CO2-treated plots, the difference was smaller and not significant. As a result, our prediction that foliar α would decrease with increased CO2 fertilization was neither supported nor refuted.

Whereas the slope of the previously observed relationship between Nmass and canopy full spectrum albedo is positive, we observed a negative relationship at the leaf level, when all species and treatments were combined. This pattern was driven in part by the visible region, where high N foliage is expected to absorb more light due to increased pigment concentration. Nevertheless, the relationship was also negative, albeit weaker, even when restricted to the NIR region.

To what can we attribute these results? Previous studies have linked scattering and NIR reflectance to leaf structural parameters such as leaf thickness, percent intercellular air space (%IAS), or the Ames/A ratio (Knapp and Carter 1998; Gausman et al. 1970; Slaton et al.2001). In our own data, there was no correlation between αnir and LMA, although this could be due to the confounding effects on reflectance of leaf thickness and leaf density, two variables that determine LMA, but have opposing effects on other leaf traits such as the fraction of leaf volume composed of intercellular air space (increased leaf thickness increases %IAS, whereas increased leaf density decreases %IAS; Niinemets 1999, 2001). A number of studies have observed positive relationships between NIR reflectance and %IAS (Castro-Esau et al. 2006; Gausman et al.1970; Slaton et al. 2001), while others suggest varying relationships between %IAS and foliar Nmass. In a study involving fertilization of tall fescue (Festuca arundinacea Schreb.), Rademacher and Nelson (2001) found that the ratio of IAS to total mesophyll space was lower in high N foliage because N fertilization enhanced mesophyll area to a greater degree than leaf volume. Conversely, Niinemets (1999), across a broad range of woody vegetation, observed a negative relationship between leaf density and both Nmass and the fraction of leaf mesophyll as intercellular airspace. This implies that leaves with high foliar N should be less dense and have a greater %IAS.

In contrast to our leaf level results, remote sensing data collected by the AVIRIS sensor showed an increase in whole-canopy NIR reflectance with N fertilization in the ORNL sweetgum plantation, a response that is consistent with observations along natural gradients in plant N status (Ollinger et al. 2008; Hollinger et al.2010). Collectively, these results suggest that the association between Nmass and canopy NIR reflectance is driven by stem or canopy level properties that are either influenced by, or covary with, leaf Nmass.

In our study, the only canopy structural variable we were able to evaluate was leaf area index (LAI, leaf area per unit area of ground), which was higher in the ORNL N-fertilized plots than in the control plots (LAI ~ 5.5 and 4.5, respectively; Richard Norby and Colleen Iversen, personal communication). To evaluate the potential effect of this difference on the canopy reflectance values observed by AVIRIS, we used SAIL-2 (Scattering from Arbitrarily Inclined Leaves; Verhoef 1984; Braswell et al. 1996), a radiative transfer model that has been used extensively in the literature (e.g., Huemmrich and Goward 1997; Andrieu et al. 1997; Daughtry et al. 2000; Zhang et al. 2006). Although SAIL is known to be sensitive to changes in LAI, especially in low-LAI systems (e.g., Asner 1998), model runs for the control and N-fertilized plots, conducted using measured LAI and measured leaf and background reflectance and transmittance spectra, produced a negligible difference in canopy reflectance (Fig. 7). This suggests that canopy properties other than LAI are driving the observed changes in canopy reflectance between the stands.
Fig. 7

Results of SAIL-2 model runs for the ambient CO2 and N-fertilized plots, conducted using measured LAI and measured leaf and background reflectance and transmittance spectra. Resulting spectra show a negligible difference in canopy reflectance

Other canopy properties that may provide a link between plant N status and canopy reflectance include shoot architecture (Smolander and Stenberg 2003; Malenovský et al.2008), crown geometry (e.g., Rautiainen et al. 2004)- and leaf angle distribution (Asner 1998; Close and Beadle 2006). Each of these are influenced by a variety of resource optimization mechanisms that are known to vary with plant nutrient status, perhaps resulting in associations between leaf traits and canopy structural properties that have yet to be explored (e.g., Ollinger 2011). As an example, N fertilization experiments involving eucalyptus (Close and Beadle 2006), wheat (Brooks et al. 2000)- and rice (Tari et al. 2009) have all demonstrated reductions in mean leaf angle (i.e.- leaves become more horizontal) with improved N nutrition. Because low leaf angle is associated with increased reflectance, particularly in the NIR region (Asner 1998; Ollinger 2011), such a response could contribute to a positive relationship between Nmass and canopy albedo if it were to occur over broad gradients within native ecosystems. Canopy leaf angle is a challenging parameter to measure, and at present, we lack the data needed to evaluate the extent to which variation in leaf angle, or other related parameters, may have contributed to our results.

Conclusions

This study investigated the importance of leaf-level reflectance properties as a potential cause of the relationship between foliar Nmass and canopy full spectrum albedo. Although we cannot completely rule out the importance of mechanisms involving leaf-level radiation scattering, our results suggest that they are not the dominant influence on this relationship. Instead, we suggest that the principal mechanism lies in associations between leaf-level Nmass and optically important canopy structural properties that have yet to be thoroughly explored. Although measuring canopy structural properties is challenging, future work should consider variables such as leaf arrangement and foliage clumping, leaf angle distribution, and crown geometry, in conjunction with leaf traits and canopy spectral properties. If met with success, such investigations would improve our overall understanding of how biogeochemical processes influence biophysical properties that are relevant to surface energy exchange and interactions between ecosystems and climate.

Notes

Acknowledgments

We thank G. James Collatz for helpful comments on a draft of this manuscript, Rob Braswell for providing the SAIL-2 model code, and Richard Norby, Colleen Iversen, and Jeffery Warren for support at ORNL. We are indebted to Michael Eastwood, ER-2 pilots Denis Steel, Tim Williams, and the rest of the AVIRIS team for aircraft data acquisition. This work was funded by a grant from the North American Carbon Program (NACP) NASA’s Terrestrial Ecology and Carbon Cycle Science Programs and a graduate fellowship provided by the Research and Discover program. The ORNL FACE experiment was supported by the US Department of Energy, Office of Science, Biological and Environmental Research Program. A.D.R. and M.K.B. acknowledge support, through the Harvard Forest REU program, from the National Science Foundation (Grant DBI-04-52254).

References

  1. Andrieu B, Baret F, Jacquemoud S, Malthus T, Steven M (1997) Evaluation of an improved version of SAIL model for simulating bidirectional reflectance of sugar beet canopies. Remote Sens Environ 60:247–257CrossRefGoogle Scholar
  2. Asner GP (1998) Biophysical and biochemical sources of variability in canopy reflectance. Remote Sens Environ 64:234–253CrossRefGoogle Scholar
  3. Braswell BH, Schimel DS, Privette JL, Moore B III, Emery WJ, Sulzman EW, Hudak AT (1996) Extracting ecological and biophysical information from AVHRR optical data: an integrated algorithm based on inverse modeling. J Geophys Res 101:23335–23348CrossRefGoogle Scholar
  4. Brooks TJ, Wall GW, Pinter PJ Jr, Kimball BA, LaMorte RL, Leavitt SW, Matthias AD, Adamsen FJ, Hunsaker DJ, Webber AN (2000) Acclimation response of spring wheat in a free-air CO2 enrichment (FACE) atmosphere with variable soil nitrogen regimes. 3. Canopy architecture and gas exchange. Photosynth Res 66:97–108PubMedCrossRefGoogle Scholar
  5. Castro-Esau KL, Sanchez-Azofeifa GA, Rivard B, Wright AJ, Quesada M (2006) Variability in leaf optical properties of Mesoamerican trees and the potential for species classification. Am J Bot 94:517–530CrossRefGoogle Scholar
  6. Chen JM, Cihlar J (1995) Quantifying the effect of canopy architecture on optical measurements of leaf area index using two gap size analysis methods. IEEE Transact Geosci Remote Sens 33:777–787CrossRefGoogle Scholar
  7. Close DC, Beadle CL (2006) Leaf angle responds to nitrogen supply in eucalypt seedlings. Is it a photoprotective mechanism? Tree Physiol 26:743–748PubMedCrossRefGoogle Scholar
  8. Daughtry CST, Walthall CL, Kim MS, Brown de Colstoun E, McMurtrey JE III (2000) Estimating corn leaf chlorophyll concentration from leaf and canopy reflectance. Remote Sens Environ 74:229–239CrossRefGoogle Scholar
  9. Ellsworth DS, Reich PB, Naumburg ES, Koch GW, Kubiske ME, Smith SD (2004) Photosythesis, carboxilation and leaf nitrogen responses of 16 species to elevated pCO2 across four free-air CO2 enrichment experiments in forest, grassland and desert. Glob Change Biol 10:2121–2138CrossRefGoogle Scholar
  10. Evans JR (1989) Photosynthesis and nitrogen relationships in leaves of C3 plants. Oecologia 78:9–19CrossRefGoogle Scholar
  11. Field C, Mooney HA (1986) The photosynthesis-nitrogen relationship in wild plants. In: Givinsh TI (ed) On the economy of form and function. Cambridge University Press, Cambridge, pp 25–55Google Scholar
  12. Gates DM, Keegan HJ, Schleter JC, Weidner VR (1965) Spectral properties of plants. Appl Opt 4:11–20CrossRefGoogle Scholar
  13. Gausman HW, Allen WA, Cardenas R, Richardson AJ (1970) Relation of light reflectance to historical and physical evaluations of cotton leaf maturity. Appl Opt 9:545–552PubMedCrossRefGoogle Scholar
  14. Gausman HW, Allen WA, Escobar DE (1974) Refractive index of plant cell walls. Appl Opt 13:109–111PubMedCrossRefGoogle Scholar
  15. Gueymard C (2004) The sun’s total and spectral irradiance for solar energy applications and solar radiation models. Sol Energy 76:423–453CrossRefGoogle Scholar
  16. Hollinger DY, Ollinger SV, Richardson AD, Meyers TP, Dail DB, Martin ME, Scott NA, Arkebauer TJ, Baldocchi DD, Clark KL, Curtis PS, Davis KJ, Desai AR, Dragoni D, Goulden ML, Gu L, Katul GG, Pallardy SG, Paw UKT, Schmid HP, Stoy PC, Suyker AE, Verma SB (2010) Albedo estimates for land surface models and support for a new paradigm based on foliage nitrogen concentration. Glob Change Biol 16:696–710CrossRefGoogle Scholar
  17. Huemmrich KF, Goward SN (1997) Vegetation canopy PAR absorptance and NDVI: an assessment for ten tree species with the SAIL model. Remote Sens Environ 61:254–269CrossRefGoogle Scholar
  18. Iversen CM, Norby RJ (2008) Nitrogen limitation in a sweetgum plantation: implications for carbon allocation and storage. Can J For Res 38:1021–1032CrossRefGoogle Scholar
  19. Jacquemoud S, Verhoef W, Baret F, Bacour C, Zarco-Tejada PJ, Asner GP, François C, Ustin SL (2009) PROSPECT + SAIL models: a review of use for vegetation characterization. Remote Sens Environ 113:S56–S66CrossRefGoogle Scholar
  20. Johnson DW, Cheng W, Joslin JD, Norby RJ, Edwards NT, Todd DE Jr (2004) Effects of elevated CO2 on nutrient cycling in a sweetgum plantation. Biogeochemistry 69:379–403CrossRefGoogle Scholar
  21. Knapp AK, Carter GA (1998) Variability in leaf optical properties among 26 species from a broad range of habitats. Am J Bot 85:940–946PubMedCrossRefGoogle Scholar
  22. Longstreth DJ, Bolaños JA, Goddard RH (1985) Photosynthetic rate and mesophyll surface area in expanding leaves of Alternanthera philoxeroides grown at two light levels. Am J Bot 72:14–19CrossRefGoogle Scholar
  23. Magill AH, Aber JD, Currie WS, Nadelhoffer KJ, Martin ME, McDowell WH, Melillo JM, Steudler P (2004) Ecosystem response to 15 years of chronic nitrogen additions at the Havard Forest LTER, Massachusetts, USA. For Ecol Manag 196:7–28CrossRefGoogle Scholar
  24. Malenovský Z, Martin E, Homolová L, Gastellu-Etchegorry J-P, Zurita-Milla R, Schaepman ME, Pokorný R, Clevers JGPW, Cudlín P (2008) Influence of woody elements of a Norway spruce canopy on nadir reflectance simulated by the DART model at very high spatial resolution. Remote Sens Environ 112:1–18CrossRefGoogle Scholar
  25. Niinemets U (1999) Research review: components of leaf dry mass per area-thickness and density-alter leaf photosynthetic capacity in reverse directions in woody plants. New Phytol 144:35–47CrossRefGoogle Scholar
  26. Niinemets U (2001) Global-scale climatic controls of leaf dry mass per area, density, and thickness in trees and shrubs. Ecology 82:453–469CrossRefGoogle Scholar
  27. Nobel PS, Zaragoza LJ, Smith WK (1975) Relation between mesophyll surface area, photosynthetic rate, and illumination level during development for leaves of Plectranthus parviflorus Henckel. Plant Physiol 55:1067–1070PubMedCrossRefGoogle Scholar
  28. Norby RJ, Iversen CM (2006) Nitrogen uptake, ditribution, turnover, and efficiency of use in a CO2-enriched sweetgum forest. Ecology 87:5–14PubMedCrossRefGoogle Scholar
  29. Norby RJ, Todd DE, Fults J, Johnson DW (2001) Allometric determination of tree growth in a CO2-enriched sweetgum stand. New Phytol 150:477–487CrossRefGoogle Scholar
  30. Norby RJ, Sholtis JD, Gunderson CA, Jawdy SS (2003) Leaf dynamics of a deciduous forest canopy: no response to elevated CO2. Oecologia 136:574–584PubMedCrossRefGoogle Scholar
  31. Ollinger SV (2011) Sources of variability in canopy reflectance and the convergent properties of plants. New Phytol 189:375–394PubMedCrossRefGoogle Scholar
  32. Ollinger SV, Aber JD, Lovett GM, Millham SE, Lathrop RG, Ellis JE (1993) A spatial model of atmospheric deposition for the Northeastern US. Ecol Appl 3:459–472CrossRefGoogle Scholar
  33. Ollinger SV, Richardson AD, Martin ME, Hollinger DY, Frolking S, Reich PB, Plourde LC, Katul GG, Munger JW, Oren R, Smith ML, Paw UKT, Bolstad PV, Cook BD, Day MC, Martin TA, Monson RK, Schmid HP (2008) Canopy nitrogen, carbon assimilation and albedo in temperate and boreal forests: functional relations and potential climate feedbacks. Proc Nat Acad Sci 105:19335–19340CrossRefGoogle Scholar
  34. Oren R, Ellsworth DS, Johnsen KH, Phillips N, Ewers BE, Maier C, Schäfer KVR, McCarthy H, Hendrey G, McNulty SG, Katul GG (2001) Soil fertility limits carbon sequestration by forest ecosystems in a CO2-enriched atmosphere. Nature 411:469–472PubMedCrossRefGoogle Scholar
  35. Rademacher IF, Nelson CJ (2001) Nitrogen effects on leaf anatomy within the intercalary meristems of tall fescue leaf blades. Ann Bot 88:893–903CrossRefGoogle Scholar
  36. Rautiainen M, Stenberg P, Nilson T, Kuusk A (2004) The effect of crown shape on the reflectance of coniferous stands. Remote Sens Environ 89:41–52CrossRefGoogle Scholar
  37. Reich PB, Walters MB, Ellsworth DS (1997) From tropics to tundra: global convergence in plant functioning. Proc Nat Acad Sci 94:13730–13734PubMedCrossRefGoogle Scholar
  38. Reich PB, Ellsworth DS, Walters MB, Vose JM, Gresham C, Volin JC, Bowman WD (1999) Generality of leaf train relationships: a test across six biomes. Ecology 80:1955–1969CrossRefGoogle Scholar
  39. Riggs JS, Tharp ML, Norby RJ (2009) ORNL FACE CO2 data. carbon dioxide information analysis center, Oak Ridge, TN, USA. http://cdiac.ornl.gov. Accessed 15-July-2010
  40. Sánchez J, Canton MP (1999) Space imaging processing. CRC Press LLC, Boca RatonGoogle Scholar
  41. Slaton MR, Hunt ER, Smith WK (2001) Estimating near-infrared leaf reflectance from leaf structural characteristics. Am J Bot 88:278–284PubMedCrossRefGoogle Scholar
  42. Smolander S, Stenberg P (2003) A method to account for shoot scale clumping in coniferous canopy reflectance models. Remote Sens Environ 88:363–373CrossRefGoogle Scholar
  43. Tari DB, Gazanchian A, Pirdashti HA, Nasiri M (2009) Flag leaf morphophysiological response to different agronomical treatments in a promising line of rice (Oryza sativa L.). Am-Euras J Agric Environ Sci 5:403–408Google Scholar
  44. Verhoef W (1984) Light scattering by leaf layers with application to canopy reflectance modeling: the SAIL model. Remote Sens Environ 16:125–141CrossRefGoogle Scholar
  45. Woolley JT (1971) Reflectance and transmittance of light by leaves. Plant Physiol 47:656–662PubMedCrossRefGoogle Scholar
  46. Wright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers F, Cavender-Bares J, Chapin T, Cornelissen JHC, Diemer M, Flexas J, Garnier E, Groom PK, Gulias J, Hikosaka K, Lamont BB, Lee T, Lee W, Lusk C, Midgley JJ, Navas M-L, Niinemets U, Oleksyn J, Osada N, Poorter H, Poot P, Prior L, Pyankov VI, Roumet C, Thomas SC, Tjoelker MG, Veneklaas EJ, Villar R (2004) The worldwide leaf economics spectrum. Nature 428:821–827PubMedCrossRefGoogle Scholar
  47. Zhang Q, Xiao X, Braswell BH, Linder E, Ollinger S, Smith ML, Jenkins JP, Baret F, Richardson AD, Moore B III, Minocha R (2006) Characterization of seasonal variation of forest canopy in a temperate deciduous broadleaf forest, using daily MODIS data. Remote Sens Environ 105:189–203CrossRefGoogle Scholar

Copyright information

© Springer-Verlag 2012

Authors and Affiliations

  • Haley F. Wicklein
    • 1
  • Scott V. Ollinger
    • 1
  • Mary E. Martin
    • 1
  • David Y. Hollinger
    • 2
  • Lucie C. Lepine
    • 1
  • Michelle C. Day
    • 1
  • Megan K. Bartlett
    • 3
  • Andrew D. Richardson
    • 4
  • Richard J. Norby
    • 5
  1. 1.Complex Systems Research Center, Morse Hall, Institute for the Study of Earth, Oceans, and SpaceUniversity of New HampshireDurhamUSA
  2. 2.Northern Research StationUS Department of Agriculture Forest ServiceDurhamUSA
  3. 3.Department of Environmental Science, Policy, and ManagementUniversity of CaliforniaBerkeleyUSA
  4. 4.Department of Organismic and Evolutionary Biology, Harvard University HerbariumHarvard UniversityCambridgeUSA
  5. 5.Environmental Sciences DivisionOak Ridge National LaboratoryOak RidgeUSA

Personalised recommendations