Estimation Of Canopy Gap Fraction Research Articles

Estimation of Canopy Gap Fraction from Terrestrial Laser Scanner Using an Angular Grid to Take Advantage of the Full Data Spatial Resolution

This paper develops an algorithm to estimate vegetation canopy gap fraction (GF), taking advantage of the full Terrestrial Laser Scanner (TLS) resolution. After calculating the TLS angular resolution, the algorithm identifies the missing laser hits (gaps) within an angular grid in the azimuthal and zenithal directions. The algorithm was first tested on angular data simulations with random (R), cluster (C) and random and cluster together (RC) gap pattern distributions. Noise introduced in the simulations as a percentage of the resolution accounted for the effect of TLS angular uncertainty. The algorithm performs accurately if angular noise is <6% of the angular resolution. To assess the impact of the change in projection, this study compared these GF estimates from angular grid simulations to their transformation into simulated hemispherical images (SHI). SHI with C patterns perform accurately, but R and RC patterns underestimate GF, especially for GF values below 0.6. The SHI performance to estimate GF was always far below the algorithm developed here with the angular grid simulations. When applied to actual TLS data acquired over individual Quercus ilex L. trees, the algorithm rendered a GF between 0.26 and 0.40. TLS had an angular noise <6%. Converting the angular grid into simulated HI (TLS-SHI) provided a better agreement with actual HI acquired in the same location as the TLS data, since they are in the same projection. The TLS-SHI underestimated GF by an average of 4% compared to HI. HI and TLS-SHI presented 14% and 17% lower values than the GF calculated from the angular grids, respectively. Nevertheless, the results from the simulations indicate that the algorithm based on the angular grid should be closer to the actual GF of the tree canopy.

Continental Estimates of Canopy Gap Fraction by Active Remote Sensing

ABSTRACTThis study estimates canopy gap fraction (GF) from Geoscience Laser Altimeter System (GLAS) waveform data retrieved from a cross section of the Australian ecosystem, represented by 4 sites. However, as the literature supports, waveform profiles from systems such as GLAS require scaling in order to retrieve more accurate estimates of GF; the degree of scaling required is not a priori known, which has hindered large-scale estimates of GF by such systems to date. We employ spatially coincident GLAS footprints and airborne laser scanning (ALS) point cloud data from each site to scale initial estimates of GLAS GF to match ALS estimates, which allows the inference of scale factors to all footprints across the Australian forest landscape via the Random Forest (RF) technique. These refined estimates of GF are submitted to a second RF model to yield regional predictions of GF across the forest landscape. These predictions are validated against ALS data from 20 sites throughout Australia (R2 = 0.56, RMSE = 25.95%) and compared to MODIS, and also compared to ALS data (R2 = 0.13, RMSE = 66.46%). The developed method for producing large-scale estimates of GF shows promise for future studies and refinement, and is particularly pertinent as future spaceborne waveform systems come online (ICESat-2 and the Global Ecosystem Dynamics Investigation LiDAR).

Estimating Canopy Gap Fraction Using ICESat GLAS within Australian Forest Ecosystems

Spaceborne laser altimetry waveform estimates of canopy Gap Fraction (GF) vary with respect to discrete return airborne equivalents due to their greater sensitivity to reflectance differences between canopy and ground surfaces resulting from differences in footprint size, energy thresholding, noise characteristics and sampling geometry. Applying scaling factors to either the ground or canopy portions of waveforms has successfully circumvented this issue, but not at large scales. This study develops a method to scale spaceborne altimeter waveforms by identifying which remotely-sensed vegetation, terrain and environmental attributes are best suited to predicting scaling factors based on an independent measure of importance. The most important attributes were identified as: soil phosphorus and nitrogen contents, vegetation height, MODIS vegetation continuous fields product and terrain slope. Unscaled and scaled estimates of GF are compared to corresponding ALS data for all available data and an optimized subset, where the latter produced most encouraging results (R2 = 0.89, RMSE = 0.10). This methodology shows potential for successfully refining estimates of GF at large scales and identifies the most suitable attributes for deriving appropriate scaling factors. Large-scale active sensor estimates of GF can establish a baseline from which future monitoring investigations can be initiated via upcoming Earth Observation missions.

Canopy gap fraction estimation from digital hemispherical images using sky radiance models and a linear conversion method

Thresholding methods for estimating canopy indices from gamma corrected digital hemispherical images are influenced both by the image interpreter and applicability of the automatic interpretation algorithm. The matrix output data of camera sensors (i.e., raw data) can be used to avoid gamma correction. Nevertheless, the impact of an interpreter or a thresholding algorithm is still present. In this paper, we present a new methodology which applies a previously published linear conversion method of raw data to a single below-canopy image to obtain unbiased estimates of canopy gap fraction. The new methodology derives an above-canopy reference image from a below-canopy image using sky radiance information available in gaps. A sky radiance model and interpolation are used to create the above-canopy image which could alternatively be obtained with a second sensor. To evaluate the new method, we conducted test measurements in three mature forest stands. The comparison of image processing output to LAI-2000 Plant Canopy Analyzer data showed high agreement and repeatability. The proposed method can be applied in versatile future versions of hemispherical image analysis programs to create unbiased comparison data for other manual or automatic image processing methods designed for plant canopies.

Methodology comparison for slope correction in canopy leaf area index estimation using hemispherical photography

The slope effect and correction methods for estimation of canopy gap fraction, leaf area index (LAI), mean leaf angle and clumping index using hemispherical photography, were investigated. The evaluation was carried out in tropical cloud forest and plantations in South-East Kenya in order to consider a range of canopy architecture and slopes up to 65%. The aim was to compare two acquisition techniques and various correction procedures. All estimates assume uniform slope, canopy parallel to ground and homogeneous canopy structure at the photo site level. (1) Photographs oriented to local zenith (levelled acquisition). Calculation and removal of sky parts of the hemisphere obstructed by topography. Azimuthal inversion of gap fraction without prior averaging, deriving local LAI estimates (quasi-random model). (i) Fixed path lengths over azimuths. Zenith reference axis. LAI referred to horizontal and corrected for topographic shading. (ii) Variable path lengths over azimuths. Normal to slope reference axis. LAI adjusted to horizontal by dividing by the slope cosine. (2) Photographs oriented parallel to slope (tilted acquisition). Fixed path lengths over azimuths. Normal to slope reference axis. LAI adjusted to horizontal by dividing by the slope cosine. Azimuthal inversion of gap fraction without prior averaging, deriving local LAI estimates (quasi-random model). Gap fractions present a stronger upslope/downslope asymmetry if retrieved from levelled acquisition. As a result, gap dispersion index and clumping index proved to be significantly higher for levelled acquisition ( P < 0.001). LAI estimates adjusted to horizontal are not significantly different, whether retrieved from levelled or tilted acquisitions, up to 30% slopes. From levelled acquisition, fixed and variable path length do not yield significantly different LAI estimates along the whole slope gradient. From tilted acquisition, LAI values were systematically higher than from levelled acquisitions, the stronger the slope, the higher the difference. Mean leaf angles do not differ significantly ( P > 0.05) for fixed vs. variable path lengths along the slope gradient up to 30%. For more severe slopes, variable path lengths yield lower mean leaf angle values. The interpretation of results from tilted acquisition remains uncertain. As a preliminary study, no preference is suggested for the levelled or tilted acquisition technique. Further investigation is needed and indirect optical derived estimates should be checked against direct reference measures, which are almost entirely lacking for mountainous areas.

Indirect estimates of canopy gap fraction based on the linear conversion of hemispherical photographs: Methodology and comparison with standard thresholding techniques

One of the crucial steps in image processing is the thresholding, meaning the image segmentation needed to separate the foreground (canopy) from the background (sky). Due to the plethora of factors affecting image grey levels (lens vignetting, gamma correction, heterogeneity of sky irradiance, etc.), with current thresholding procedures the estimation of canopy gap fraction is rather uncertain. Both manual and automatic segmentation methods are highly dependent on exposure, operator experience or assumptions of thresholding algorithms, and therefore produce rather subjective and non-repeatable results. However, since CCD and CMOS sensors used in digital cameras respond linearly to light, it is nowadays possible to obtain images in which digital numbers are proportional to canopy brightness. For this purpose the standard in-camera logarithmic conversion (gamma correction) has to be replaced with a linear conversion of the sensor analog signal. In this work, the potential of the linear conversion of digital images in the estimation of canopy gap fraction is assessed and discussed, comparing the estimates obtained with digital cameras and with the LI-COR LAI-2000. In addition, the performance of the linear conversion is compared with standard automatic or manual thresholding methods, both in terms of canopy gap fraction and of plant architectural parameters. The ability to acquire hemispherical photographs with digital numbers proportional to radiance opens new possibilities in the analysis of canopy structure and microclimate, with important feedbacks in ecology and remote sensing. In particular, using two synchronized cameras above and below the canopy, it is possible to calculate high-resolution gap fraction data, which provide an accurate and objective estimation of the intercepted radiation and of typical architectural parameters, such as LAI. In addition, the effectiveness of this methodology for the estimation of linear and logarithmic averages of canopy gap fraction, makes this technique an important step forward in the estimation of leaf area index of non-random canopies.

Shoot structure and photosynthetic efficiency along the light gradient in a Scots pine canopy.

We examined the effects of structural and physiological acclimation on the photosynthetic efficiency of Scots pine (Pinus sylvestris L.) shoots. We estimated daily light interception (DLI) and photosynthesis (DPHOT) of a number of sample shoots situated at different positions in the canopy. Photosynthetic efficiency (epsilon) was defined as the ratio of DPHOT to the potential daily light interception (DLI(ref)) defined as the photosynthetically active radiation (PAR) intercepted per unit area of a sphere at the shoot location. To calculate DLI(ref), DLI and DPHOT, the radiation field surrounding a shoot in the canopy was first modeled using simulated directional distributions of incoming PAR on a clear and an overcast day, and estimates of canopy gap fraction in different directions provided by hemispherical photographs. A model of shoot geometry and measured data on shoot structure and photosynthetic parameters were used to simulate the distribution of PAR irradiance on the needle surface area of the shoot. Photosynthetic efficiency (epsilon) was separated into light-interception efficiency (epsilon(I) = DLI/DLI(ref)) and conversion efficiency (epsilon(PHOT) = DPHOT/DLI). This allowed us to quantify separately the effect of structural acclimation on the efficiency of photosynthetic light capture (epsilon(l)), and the effect of physiological acclimation on conversion efficiency (epsilon(PHOT)). The value of epsilon increased from the top to the bottom of the canopy. The increase was largely explained by structural acclimation (higher epsilon(I)) of the shade shoots. The value of epsilon(PHOT) of shade foliage was similar to that of sun foliage. Given these efficiencies, the clear-day value of DPHOT for a sun shoot transferred to shade was only half that of a shade shoot at its original position. The method presented here provides a tool for quantitatively estimating the role of acclimation in total canopy photosynthesis.

The robustness of canopy gap fraction estimates from red and near-infrared reflectances: A comparison of approaches

The estimation of canopy gap fraction [P o(0)] from red and near-infrared reflectance is investigated. Several approaches are compared, using classical vegetation indices ( VI) and backpropagation neural networks. The parameters of the VI-P o(0) relationships and the synaptic weights and biases of two-layer neural networks were successfully adjusted on an experimental data set and on data derived from radiative transfer model simulations. Three experimental data sets were acquired over sugar beet canopies. They expressed a large range of canopy architecture and soil background reflectance. Two of them that correspond to similar experimental procedures were merged together and then randomly split into two subsets: one for fitting the parameters of the VI-P o(0) relationships or to train the neural network, and the other for the evaluation of the predictive performances. The third experiment is used as an additional independent data set to test the robustness of the approaches. The model-simulated data set was generated using the SAIL and PROSPECT radiative transfer models, with input parameter values that were chosen to have similar distributions as observed over sugar beet canopies. As for the experimental data set, the model-simulated data set was split into a training (calibration) and a test (validation) data set. Results show that the gap fraction can be accurately estimated from the red and near-infrared reflectance without any external information except maybe the crop type (here sugar beet) and the soil line characteristics required for some of the vegetation indices. The best predictive performances were observed for the SAVI-like vegetation indices ( SAVI, TSAVI, MSAVI) and the poorest for the NDVI, PVI, and GEMI having intermediate although satisfactory results. Neural networks trained on the simulated data set appeared to be the most robust approach. It allows us to implicitly incorporate our knowledge about the physics of the radiative transfer in the interpretation of remote sensing data.