Estimation of leaf area index and its sunlit portion from DSCOVR EPIC data: Theoretical basis.

This paper presents the theoretical basis of the algorithm designed for the generation of leaf area index and diurnal course of its sunlit portion from NASA's Earth Polychromatic Imaging Camera (EPIC) onboard NOAA's Deep Space Climate Observatory (DSCOVR). The Look-up-Table (LUT) approach implemented in the MODIS operational LAI/FPAR algorithm is adopted. The LUT, which is the heart of the approach, has been significantly modified. First, its parameterization incorporates the canopy hot spot phenomenon and recent advances in the theory of canopy spectral invariants. This allows more accurate decoupling of the structural and radiometric components of the measured Bidirectional Reflectance Factor (BRF), improves scaling properties of the LUT and consequently simplifies adjustments of the algorithm for data spatial resolution and spectral band compositions. Second, the stochastic radiative transfer equations are used to generate the LUT for all biome types. The equations naturally account for radiative effects of the three-dimensional canopy structure on the BRF and allow for an accurate discrimination between sunlit and shaded leaf areas. Third, the LUT entries are measurable, i.e., they can be independently derived from both below canopy measurements of the transmitted and above canopy measurements of reflected radiation fields. This feature makes possible direct validation of the LUT, facilitates identification of its deficiencies and development of refinements. Analyses of field data on canopy structure and leaf optics collected at 18 sites in the Hyytiälä forest in southern boreal zone in Finland and hyperspectral images acquired by the EO-1 Hyperion sensor support the theoretical basis.

Quantitative characterization of clumping in Scots pine crowns.

BACKGROUND AND AIMS: Proper characterization of the clumped structure of forests is needed for calculation of the absorbed radiation and photosynthetic production by a canopy. This study examined the dependency of crown-level clumping on tree size and growth conditions in Scots pine (Pinus sylvestris), and determined the ability of statistical canopy radiation models to quantify the degree of self-shading within crowns as a result of the clumping effect. METHODS: Twelve 3-D Scots pine trees were generated using an application of the LIGNUM model, and the crown-level clumping as quantified by the crown silhouette to total needle area ratio (STAR(crown)) was calculated. The results were compared with those produced by the stochastic approach of modelling tree crowns as geometric shapes filled with a random medium. KEY RESULTS: Crown clumping was independent of tree height, needle area and growth conditions. The results supported the capability of the stochastic approach in characterizing clumping in crowns given that the outer shell of the tree crown is well represented. CONCLUSIONS: Variation in the whole-stand clumping index is induced by differences in the spatial pattern of trees as a function of, for example, stand age rather than by changes in the degree of self-shading within individual crowns as they grow bigger.

Hyperspectral remote sensing of foliar nitrogen content.

A strong positive correlation between vegetation canopy bidirectional reflectance factor (BRF) in the near infrared (NIR) spectral region and foliar mass-based nitrogen concentration (%N) has been reported in some temperate and boreal forests. This relationship, if true, would indicate an additional role for nitrogen in the climate system via its influence on surface albedo and may offer a simple approach for monitoring foliar nitrogen using satellite data. We report, however, that the previously reported correlation is an artifact--it is a consequence of variations in canopy structure, rather than of %N. The data underlying this relationship were collected at sites with varying proportions of foliar nitrogen-poor needleleaf and nitrogen-rich broadleaf species, whose canopy structure differs considerably. When the BRF data are corrected for canopy-structure effects, the residual reflectance variations are negatively related to %N at all wavelengths in the interval 423-855 nm. This suggests that the observed positive correlation between BRF and %N conveys no information about %N. We find that to infer leaf biochemical constituents, e.g., N content, from remotely sensed data, BRF spectra in the interval 710-790 nm provide critical information for correction of structural influences. Our analysis also suggests that surface characteristics of leaves impact remote sensing of its internal constituents. This further decreases the ability to remotely sense canopy foliar nitrogen. Finally, the analysis presented here is generic to the problem of remote sensing of leaf-tissue constituents and is therefore not a specific critique of articles espousing remote sensing of foliar %N.

Analysis of the sensitivity of absorbed light and incident light profile to various canopy architecture and stand conditions.

We analyzed the effect of simplifying assumptions in canopy representation of radiation transfer models, comparing modeled diffuse non-interceptance and photosynthetic photon flux density with measurements at different layers of complex pine-broadleaved canopy with large seasonal variation of leaf area index. The most detailed model included clumping of trees (i.e., stand density) and a vertical specification of leaf angle distribution and shoot clumping. A less detailed model replaced the vertically specified variables with their means. The most parsimonious model accounted for neither shoot clumping nor stand density. The vertical specification of shoot clumping and leaf angle distribution only slightly improved vertical and seasonal openness and light estimates over using mean values. Further simplification had little effect on total absorbed light but was more risky for estimates of the vertical distributions of openness and light absorbed by the canopy, which will affect photosynthesis estimates due to the non-linearity of photosynthetic light response. Including woody surfaces in winter, when leaf area was low, was essential for reproducing the measurements correctly. A sensitivity analysis showed that ignoring (i) shoot clumping could result in a substantial overestimation of total absorbed light with errors increasing with decreasing leaf area and (ii) stand density in sparse stands could lead to substantial overestimation of total absorbed light, and the effect is largely independent of leaf area. Also, (iii) the effect of changing leaf angle distribution increased with decreasing leaf area, and was larger and more persistent along the leaf area range with increasing shoot clumping. Overall, accounting for the effect of tree clumping on absorbed light is most important in stands composed of species where leaves are not very clumped (e.g., broadleaved). However, even in forests with highly clumped shoots (i.e., coniferous), an accurate estimation of absorbed light distribution in stands requires incorporation of stand density in the model.

Fertilization has little effect on light-interception efficiency of Picea abies shoots.

We investigated effects of nutrient availability on shoot structure and light-interception efficiency based on data from control (C) and irrigated + fertilized (IL) trees of Norway spruce (Picea abies (L.) Karst.). The sampling of 1-year-old shoots was designed to cover the variation in canopy exposure within the live crown zone, where current-year shoots were still found. Canopy openness was used as a measure of light availability at the shoot's position. Openness values for the sample shoots ranged from 0.02 to 0.77 on the IL plot, and from 0.10 to 0.96 on the C plot. Among needle dimensions, needle width increased most with canopy openness. At fixed canopy openness, needle width was larger, and the ratio of needle thickness to width was smaller in IL trees than in C trees. Specific needle area (SNA) and the ratio of shoot silhouette area to total needle area (STAR) decreased with canopy openness, so that the combined effect was a threefold decrease in the ratio of shoot silhouette area to unit dry mass (SMR = STAR x SNA) along the studied range of openness values. This means that the light-interception efficiency of shoots per unit needle dry mass was three times higher for the most shaded shoots than for sun shoots. A test of the effect of fertilization on the relationships of SNA, STAR and SMR indicated statistically significant differences in both slope and intercept for SNA and STAR, and in the intercept for SMR. However, the differences partly cancelled each other so that, at medium values of canopy openness, differences between treatments in predicted SNA, STAR and SMR were small. At 0.5 canopy openness, predicted STAR of IL shoots was 6.1% larger than STAR of C shoots, but SMR of IL shoots was 10% smaller than that of C shoots. The results suggest that light-interception efficiency per unit needle area or mass of the shoots is not greatly affected by fertilization.

Shoot structure, light interception, and distribution of nitrogen in an Abies amabilis canopy.

We studied the effects of variation in shoot structure and needle morphology on the distributions of light and nitrogen within a Pacific silver fir (Abies amabilis (Dougl.) Forbes) canopy. Specifically, we investigated the role of morphological shade acclimation in the determination of resource use efficiency, which is claimed to be optimal when the distribution of nitrogen within the canopy is directly proportional to the distribution of intercepted photosynthetically active radiation (PAR). Shoots were collected from different heights in the crowns of trees representing four different size classes. A new method was developed to estimate seasonal light interceptance (SLI, intercepted PAR per unit needle area) of the shoots using a model for the directional distribution of above-canopy PAR, measurements of shoot silhouette area and canopy gap fraction in different directions. The ratio SLI/SLI(o), where the reference value SLI(o) represents the seasonal light interceptance of a spherical surface at the shoot location, was used to quantify the efficiency of light capture by a shoot. The ratio SLI/SLI(o) doubled from the top to the bottom of the canopy, mainly as a result of smaller internal shading in shade shoots than in sun shoots. Increased light-capturing efficiency of shade shoots implies that the difference in intercepted light by sun shoots versus shade shoots is much less than the decrease in available light from the upper to the lower canopy. For example, SLI of the five most sunlit shoots was only about 20 times greater than the SLI of the five most shaded shoots, whereas SLI(o) was 40 times greater for sun shoots than for shade shoots. Nitrogen content per unit needle area was about three times higher in sun needles than in shade needles. This variation, however, was not enough to produce proportionality between the amounts of nitrogen and intercepted PAR throughout the canopy.

Simulations of the effects of shoot structure and orientation on vertical gradients in intercepted light by conifer canopies.

Coniferous tree canopies typically carry more leaf area than is necessary to intercept most of the incoming light. I postulated that an excessively large leaf area will reduce net productivity at tree level, unless the net photosynthetic production of the most shaded shoots in the canopy remains positive. The hypothesis tested was that a coniferous tree canopy maintains a large productive leaf area by increasing the efficiency of light capture as the available light decreases. The light interception efficiency of a shoot was quantified by the ratio of shoot silhouette area to total needle surface area (STAR). The STAR depends on shoot geometry and varies with shoot orientation relative to the direction of light. Shade shoots have a larger STAR and, in particular, higher values of STAR(max) than sun shoots. In addition, shade shoots tend to be horizontally inclined, which may increase the advantage of a large STAR(max) in the lower canopy, where radiation is incident from angles closer to the zenith. Adaptation to shade (changes in STAR and shoot orientation) was described on the basis of empirical data for several coniferous species, and the vertical gradient of seasonal light interception by shoots was simulated assuming different adaptive strategies. Simulations were performed at two latitudes, to account for differences in the amount and directional distribution of light during the growing season. Results support the hypothesis that increases in STAR, shoot zenith angle and shoot asymmetry (flatness) with shading increase the efficiency of light interception by deeply shaded shoots. However, because competition for light among shoots increases progressively as soon as shade acclimation occurs, there cannot exist a deep layer of shade shoots, such that the net productivity of each shoot remains positive (i.e., irradiance is above the compensation point). Therefore, if maximization of productive leaf area is the goal, the optimal strategy is to maintain an inefficient deep canopy and to increase light interception efficiency only when shading becomes severe.

Performance of the LAI-2000 plant canopy analyzer in estimating leaf area index of some Scots pine stands.

The LAI-2000 plant canopy analyzer (Li-Cor, Inc., Lincoln, NE) was tested at six experimental plots of Scots pine (Pinus sylvestris L.) in central Sweden at peak leaf area in August and after litterfall in October 1990. An independent estimate of leaf area index for August 1990 was obtained based on an empirically derived regression of needle area on stem sapwood area, and the decrease in leaf area between the two measurements was estimated from measurements of litterfall. A strong linear relationship was found between estimates by the LAI-2000 (L(Li-Cor)) and the indirect estimates of leaf area index (taken as half of total surface area) (L). The finding that L(Li-Cor) was considerably smaller than L was explained theoretically. It was shown that if shoots, instead of individual needles, are randomly distributed in the canopy, L(Li-Cor) corresponds to L multiplied by a factor (beta) characterizing the mutual shading of needles on the shoot. The shading factor, beta, was equal to the ratio of spherically projected shoot area to spherically projected needle area, where the spherically projected area is defined as the average projection (silhouette) area taken over all directions in space. The quantity betaL was defined as the shoot silhouette area index (SSAI), and an equation for the relationship between SSAI and the mean silhouette to total area ratio (mean STAR) of shoots was derived. Measured values of mean STAR for Scots pine indicated that L(Li-Cor) corresponds to SSAI rather than L. However, the decrease in leaf area index due to litterfall occurring between August and October was only partly detected by the LAI-2000, possibly because SSAI did not change to the same degree as L, i.e., there was an increase in the factor beta. This hypothesis is supported by data showing a large increase in mean STAR with shoot age.