Mapping beech (Fagus sylvatica L.) forest structure with airborne hyperspectral imagery

Estimating forest structural attributes using multispectral remote sensing is challenging because of the saturation of multispectral indices at high canopy cover. The objective of this study was to assess the utility of hyperspectral data in estimating and mapping forest structural parameters including mean diameter-at-breast height (DBH), mean tree height and tree density of a closed canopy beech forest (Fagus sylvatica L.). Airborne HyMap images and data on forest structural attributes were collected from the Majella National Park, Italy in July 2004. The predictive performances of vegetation indices (VI) derived from all possible two-band combinations (VI_(i,j) = (R_i − R_j)/(R_i + R_j), where R_i and R_j = reflectance in any two bands) were evaluated using calibration (n = 33) and test (n = 20) data sets. The potential of partial least squares (PLS) regression, a multivariate technique involving several bands was also assessed. New VIs based on the contrast between reflectance in the red-edge shoulder (756–820 nm) and the water absorption feature centred at 1200 nm (1172–1320 nm) were found to show higher correlations with the forest structural parameters than standard VIs derived from NIR and visible reflectance (i.e. the normalised difference vegetation index, NDVI). PLS regression showed a slight improvement in estimating the beech forest structural attributes (prediction errors of 27.6%, 32.6% and 46.4% for mean DBH, height and tree density, respectively) compared to VIs using linear regression models (prediction errors of 27.8%, 35.8% and 48.3% for mean DBH, height and tree density, respectively). Mean DBH was the best predicted variable among the stand parameters (calibration R² = 0.62 for an exponential model fit and standard error of prediction = 5.12 cm, i.e. 25% of the mean). The predicted map of mean DBH revealed high heterogeneity in the beech forest structure in the study area. The spatial variability of mean DBH occurs at less than 450 m. The DBH map could be useful to forest management in many ways, e.g. thinning of coppice to promote diameter growth, to assess the effects of management on forest structure or to detect changes in the forest structure caused by anthropogenic and natural factors.     

Tree species classification using UAS-based digital aerial photogrammetry point clouds and multispectral imageries in subtropical natural forests

Tree species composition of forest stand is an important indicator of forest inventory attributes for assessing ecosystem health, understanding successional processes, and digitally displaying forest biodiversity. In this study, we acquired high spatial resolution multispectral and RGB imagery over a subtropical natural forest in southwest China using a fixed-wing UAV system. Digital aerial photogrammetric (DAP) technique was used to generate multi-spectral and RGB derived point clouds, upon which individual tree crown (ITC) delineation algorithms and a machine learning classifier were used to identify dominant tree species. To do so, the structure-from-motion method was used to generate RGB imagery-based DAP point clouds. Then, three ITC delineation algorithms (i.e., point cloud segmentation (PCS), image-based multiresolution segmentation (IMRS), and advanced multiresolution segmentation (AMRS)) were used and assessed for ITC detection. Finally, tree-level metrics (i.e., multispectral, texture and point cloud metrics) were used as metrics in the random forest classifier used to classify eight dominant tree species. Results indicated that the accuracy of the AMRS ITC segmentation was highest (F₁-score = 82.5 %), followed by the segmentation using PCS (F₁-score = 79.6 %), the IMRS exhibited the lowest accuracy (F₁-score = 78.6 %); forest types classification (coniferous and deciduous) had a higher accuracy than the classification of all eight tree species, and the combination of spectral, texture and structural metrics had the highest classification accuracy (overall accuracy = 80.20 %). In the classification of both eight tree species and two forest types, the classification accuracies were lowest when only using spectral metrics, indicated that the texture metrics and point cloud structural metrics had a positive impact on the classification (the overall accuracy and kappa accuracy increased by 1.49–4.46 % and 2.86–6.84 %, respectively).     

机载LiDAR数据估算样地和单木尺度森林地上生物量

利用机载激光雷达点云数据,结合大量实测单木结构信息,分别从样地和单木尺度估算了森林地上生物量AGB。首先,利用局部最大值单木提取算法提取了每个样地内的单木结构参数,并针对样地和单木尺度分别计算了一组激光雷达变量。然后,利用激光雷达变量和地上生物量及其两者的对数形式,从样地和单木尺度分别构建了估算模型。最后,针对两种尺度估算过程中存在的不确定性进行了详细讨论。结果表明:(1)样地和单木尺度模型估算的森林地上生物量与地面实测值都具有明显的相关性,且对数模型估算效果要优于非对数模型;(2)样地尺度模型估算效果(R2=0.84,rRMSE=0.23)明显优于单木尺度模型(R2=0.61,rRMSE=0.46);(3)按树木类型分别进行估算可以提高单木地上生物量的估算精度;(4)不论是样地还是单木尺度地上生物量估算都存在一定的不确定性,与样地尺度相比,单木尺度估算过程的不确定性更大,这种不确定性主要来自单木识别过程。     

Can regional aerial images from orthophoto surveys produce high quality photogrammetric Canopy Height Model? A single tree approach in Western Europe

Forest monitoring tools are needed to promote effective and data driven forest management and forest policies. Remote sensing techniques can increase the speed and the cost-efficiency of the forest monitoring as well as large scale mapping of forest attribute (wall-to-wall approach). Digital Aerial Photogrammetry (DAP) is a common cost-effective alternative to airborne laser scanning (ALS) which can be based on aerial photos routinely acquired for general base maps. DAP based on such pre-existing dataset can be a cost effective source of large scale 3D data. In the context of forest characterization, when a quality Digital Terrain Model (DTM) is available, DAP can produce photogrammetric Canopy Height Model (pCHM) which describes the tree canopy height. While this potential seems pretty obvious, few studies have investigated the quality of regional pCHM based on aerial stereo images acquired by standard official aerial surveys. Our study proposes to evaluate the quality of pCHM individual tree height estimates based on raw images acquired following such protocol using a reference filed-measured tree height database. To further ensure the replicability of the approach, the pCHM tree height estimates benchmarking only relied on public forest inventory (FI) information and the photogrammetric protocol was based on low-cost and widely used photogrammetric software. Moreover, our study investigates the relationship between the pCHM tree height estimates based on the neighboring forest parameter provided by the FI program.Our results highlight the good agreement of tree height estimates provided by pCHM using DAP with both field measured and ALS tree height data. In terms of tree height modeling, our pCHM approach reached similar results than the same modeling strategy applied to ALS tree height estimates. Our study also identified some of the drivers of the pCHM tree height estimate error and found forest parameters like tree size (diameter at breast height) and tree type (evergreenness/deciduousness) as well as the terrain topography (slope) to be of higher importance than image survey parameters like the variation of the overlap or the sunlight condition in our dataset. In combination with the pCHM tree height estimate, the terrain slope, the Diameter at Breast Height (DBH) and the evergreenness factor were used to fit a multivariate model predicting the field measured tree height. This model presented better performance than the model linking the pCHM estimates to the field tree height estimates in terms of r² (0.90 VS 0.87) and root mean square error (RMSE, 1.78 VS 2.01 m). Such aspects are poorly addressed in literature and further research should focus on how pCHM approaches could integrate them to improve forest characterization using DAP and pCHM. Our promising results can be used to encourage the use of regional aerial orthophoto surveys archive to produce large scale quality tree height data at very low additional costs, notably in the context of updating national forest inventory programs.     

The spatial prediction of tree species diversity in savanna woodlands of Southern Africa

In this study, we tested the utility of remotely sensed data in predicting tree species diversity in savanna woodlands. Specifically, we developed linear regression functions based on a combination of the coefficient of variation of near infrared (NIR) radiance and the soil-adjusted vegetation index (SAVI), both derived from advanced space-borne thermal emission and reflection radiometer satellite imagery. Using the regression functions in a Geographic Information System (GIS), we predicted the spatial variations in tree species diversity. Our results showed that tree species diversity can be predicted using a combination of the coefficient of variation of NIR radiance and SAVI. We conclude that remotely sensed data can be used to spatially predict tree species diversity in savanna woodlands.     

机载激光雷达及高光谱的森林乔木物种多样性遥感监测

利用机载LiDAR和高光谱数据并结合37个地面调查样本数据,基于结构差异与光谱变异理论,通过相关分析法分别筛选了3个最优林冠结构参数和6个最优光谱指数,在单木尺度上利用自适应C均值模糊聚类算法,在神农架国家自然保护区开展森林乔木物种多样性监测,实现了森林乔木物种多样性的区域成图。研究结果表明,(1)基于结合形态学冠层控制的分水岭算法可以获得较高精度的单木分割结果(R~2=0.88,RMSE=13.17,P0.001);(2)基于LiDAR数据提取的9个结构参数中,95%百分位高度、冠层盖度和植被穿透率为最优结构参数,与Shannon-Wiener指数的相关性达到R~2=0.39—0.42(P0.01);(3)基于机载高光谱数据筛选的16个常用的植被指数中,CRI、OSAVI、Narrow band NDVI、SR、Vogelmann index1、PRI与Shannon-Wiener指数的相关性最高(R~2=0.37—0.45,P0.01);(4)在研究区,利用以30 m×30 m为窗口的自适应模糊C均值聚类算法可预测的最大森林乔木物种数为20,物种丰富度的预测精度为R~2=0.69,RMSE=3.11,Shannon-Wiener指数的预测精度为R~2=0.70,RMSE=0.32。该研究在亚热带森林开展乔木物种多样性监测,是在区域尺度上进行物种多样性成图的重要实践,可有效补充森林生物多样性本底数据的调查手段,有助于实现生物多样性的长期动态监测及科学分析森林物种多样性的现状和变化趋势。     

Comparative performance of convolutional neural network,weighted and conventional support vector machine and random forest for classifying tree species using hyperspectral and photogrammetric data

ABSTRACT

The classification of tree species can significantly benefit from high spatial and spectral information acquired by unmanned aerial vehicles (UAVs) associated with advanced classification methods. This study investigated the following topics concerning the classification of 16 tree species in two subtropical forest fragments of Southern Brazil: i) the potential integration of UAV-borne hyperspectral images with 3D information derived from their photogrammetric point cloud (PPC); ii) the performance of two machine learning methods (support vector machine – SVM and random forest – RF) when employing different datasets at a pixel and individual tree crown (ITC) levels; iii) the potential of two methods for dealing with the imbalanced sample set problem: a new weighted SVM (wSVM) approach, which attributes different weights to each sample and class, and a deep learning classifier (convolutional neural network – CNN), associated with a previous step to balance the sample set; and finally, iv) the potential of this last classifier for tree species classification as compared to the above mentioned machine learning methods. Results showed that the inclusion of the PPC features to the hyperspectral data provided a great accuracy increase in tree species classification results when conventional machine learning methods were applied, between 13 and 17% depending on the classifier and the study area characteristics. When using the PPC features and the canopy height model (CHM), associated with the majority vote (MV) rule, the SVM, wSVM and RF classifiers reached accuracies similar to the CNN, which outperformed these classifiers for both areas when considering the pixel-based classifications (overall accuracy of 84.4% in Area 1, and 74.95% in Area 2). The CNN was between 22% and 26% more accurate than the SVM and RF when only the hyperspectral bands were employed. The wSVM provided a slight increase in accuracy not only for some lesser represented classes, but also some major classes in Area 2. While conventional machine learning methods are faster, they demonstrated to be less stable to changes in datasets, depending on prior segmentation and hand-engineered features to reach similar accuracies to those attained by the CNN. To date, CNNs have been barely explored for the classification of tree species, and CNN-based classifications in the literature have not dealt with hyperspectral data specifically focusing on tropical environments. This paper thus presents innovative strategies for classifying tree species in subtropical forest areas at a refined legend level, integrating UAV-borne 2D hyperspectral and 3D photogrammetric data and relying on both deep and conventional machine learning approaches.     

The effect of leaf-on and leaf-off forest canopy conditions on LiDAR derived estimations of forest structural diversity

Forest structural diversity metrics describing diversity in tree size and crown shape within forest stands can be used as indicators of biodiversity. These diversity metrics can be generated using airborne laser scanning (LiDAR) data to provide a rapid and cost effective alternative to ground-based inspection. Measures of tree height derived from LiDAR can be significantly affected by the canopy conditions at the time of data collection, in particular whether the canopy is under leaf-on or leaf-off conditions, but there have been no studies of the effects on structural diversity metrics. The aim of this research is to assess whether leaf-on/leaf-off changes in canopy conditions during LiDAR data collection affect the accuracy of calculated forest structural diversity metrics. We undertook a quantitative analysis of LiDAR ground detection and return height, and return height diversity from two airborne laser scanning surveys collected under leaf-on and leaf-off conditions to assess initial dataset differences. LiDAR data were then regressed against field-derived tree size diversity measurements using diversity metrics from each LiDAR dataset in isolation and, where appropriate, a mixture of the two. Models utilising leaf-off LiDAR diversity variables described DBH diversity, crown length diversity and crown width diversity more successfully than leaf-on (leaf-on models resulted in R² values of 0.66, 0.38 and 0.16, respectively, and leaf-off models 0.67, 0.37 and 0.23, respectively). When LiDAR datasets were combined into one model to describe tree height diversity and DBH diversity the models described 75% and 69% of the variance (R² of 0.75 for tree height diversity and 0.69 for DBH diversity). The results suggest that tree height diversity models derived from airborne LiDAR, collected (and where appropriate combined) under any seasonal conditions, can be used to differentiate between simple single and diverse multiple storey forest structure with confidence.     

Predicting forest carbon stocks from high resolution satellite data in dry forests of Zimbabwe: exploring the effect of the red-edge band in forest carbon stocks estimation

In this study, we tested whether the inclusion of the red-edge band as a covariate to vegetation indices improves the predictive accuracy in forest carbon estimation and mapping in savanna dry forests of Zimbabwe. Initially, we tested whether and to what extent vegetation indices (simple ratio SR, soil-adjusted vegetation index and normalized difference vegetation index) derived from high spatial resolution satellite imagery (WorldView-2) predict forest carbon stocks. Next, we tested whether inclusion of reflectance in the red-edge band as a covariate to vegetation indices improve the model's accuracy in forest carbon prediction. We used simple regression analysis to determine the nature and the strength of the relationship between forest carbon stocks and remotely sensed vegetation indices. We then used multiple regression analysis to determine whether integrating vegetation indices and reflection in the red-edge band improve forest carbon prediction. Next, we mapped the spatial variation in forest carbon stocks using the best regression model relating forest carbon stocks to remotely sensed vegetation indices and reflection in the red-edge band. Our results showed that vegetation indices alone as an explanatory variable significantly (p < 0.05) predicted forest carbon stocks with R² ranging between 45 and 63% and RMSE ranging from 10.3 to 12.9%. However, when the reflectance in the red-edge band was included in the regression models the explained variance increased to between 68 and 70% with the RMSE ranging between 9.56 and 10.1%. A combination of SR and reflectance in the red edge produced the best predictor of forest carbon stocks. We concluded that integrating vegetation indices and reflectance in the red-edge band derived from high spatial resolution can be successfully used to estimate forest carbon in dry forests with minimal error.     

Preparation of volume table of SAL (Shorea robusta) - an approach using satellite data

Sal (Shorea robusta) is an important forest tree species in north and north-eastern India. Large-scale plantations of this species have been raised there under taungya and coppice system of management. The conventional volume table prepared for high sal forest is referred to infer the volume of production of this species. Earlier workers have used aerial remote sensing data to develop volume tables of this species. In the present study a volume table for sal is developed based on remotely sensed satellite data using a regression technique. A two-step method was developed to estimate mean tree volume from satellite data. In step 1, mean crown diameter — an intermediate variable - was estimated from satellite data. In step 2, the estimated mean crown diameter was used to estimate the mean tree volume. Addition of age of the crop as an independent variable improved the predictive ability of the regression equation.     

Single tree biomass modelling using airborne laser scanning

Accurate forest biomass mapping methods would provide the means for e.g. detecting bioenergy potential, biofuel and forest-bound carbon. The demand for practical biomass mapping methods at all forest levels is growing worldwide, and viable options are being developed. Airborne laser scanning (ALS) is a promising forest biomass mapping technique, due to its capability of measuring the three-dimensional forest vegetation structure. The objective of the study was to develop new methods for tree-level biomass estimation using metrics derived from ALS point clouds and to compare the results with field references collected using destructive sampling and with existing biomass models. The study area was located in Evo, southern Finland. ALS data was collected in 2009 with pulse density equalling approximately 10 pulses/m². Linear models were developed for the following tree biomass components: total, stem wood, living branch and total canopy biomass. ALS-derived geometric and statistical point metrics were used as explanatory variables when creating the models. The total and stem biomass root mean square error per cents equalled 26.3% and 28.4% for Scots pine (Pinus sylvestris L.), and 36.8% and 27.6% for Norway spruce (Picea abies (L.) H. Karst.), respectively. The results showed that higher estimation accuracy for all biomass components can be achieved with models created in this study compared to existing allometric biomass models when ALS-derived height and diameter were used as input parameters. Best results were achieved when adding field-measured diameter and height as inputs in the existing biomass models. The only exceptions to this were the canopy and living branch biomass estimations for spruce. The achieved results are encouraging for the use of ALS-derived metrics in biomass mapping and for further development of the models.     

Image matching as a data source for forest inventory – Comparison of Semi-Global Matching and Next-Generation Automatic Terrain Extraction algorithms in a typical managed boreal forest environment

Image matching is emerging as a compelling alternative to airborne laser scanning (ALS) as a data source for forest inventory and management. There is currently an open discussion in the forest inventory community about whether, and to what extent, the new method can be applied to practical inventory campaigns. This paper aims to contribute to this discussion by comparing two different image matching algorithms (Semi-Global Matching [SGM] and Next-Generation Automatic Terrain Extraction [NGATE]) and ALS in a typical managed boreal forest environment in southern Finland. Spectral features from unrectified aerial images were included in the modeling and the potential of image matching in areas without a high resolution digital terrain model (DTM) was also explored. Plot level predictions for total volume, stem number, basal area, height of basal area median tree and diameter of basal area median tree were modeled using an area-based approach. Plot level dominant tree species were predicted using a random forest algorithm, also using an area-based approach. The statistical difference between the error rates from different datasets was evaluated using a bootstrap method.Results showed that ALS outperformed image matching with every forest attribute, even when a high resolution DTM was used for height normalization and spectral information from images was included. Dominant tree species classification with image matching achieved accuracy levels similar to ALS regardless of the resolution of the DTM when spectral metrics were used. Neither of the image matching algorithms consistently outperformed the other, but there were noticeably different error rates depending on the parameter configuration, spectral band, resolution of DTM, or response variable. This study showed that image matching provides reasonable point cloud data for forest inventory purposes, especially when a high resolution DTM is available and information from the understory is redundant.     

Predicting individual tree attributes from airborne laser point clouds based on the random forests technique

This paper depicts an approach for predicting individual tree attributes, i.e., tree height, diameter at breast height (DBH) and stem volume, based on both physical and statistical features derived from airborne laser-scanning data utilizing a new detection method for finding individual trees together with random forests as an estimation method. The random forests (also called regression forests) technique is a nonparametric regression method consisting of a set of individual regression trees. Tests of the method were performed, using 1476 trees in a boreal forest area in southern Finland and laser data with a density of 2.6 points per m². Correlation coefficients (R) between the observed and predicted values of 0.93, 0.79 and 0.87 for individual tree height, DBH and stem volume, respectively, were achieved, based on 26 laser-derived features. The corresponding relative root-mean-squared errors (RMSEs) were 10.03%, 21.35% and 45.77% (38% in best cases), which are similar to those obtained with the linear regression method, with maximum laser heights, laser-estimated DBH or crown diameters as predictors. With random forests, however, the forest models currently used for deriving the tree attributes are not needed. Based on the results, we conclude that the method is capable of providing a stable and consistent solution for determining individual tree attributes using small-footprint laser data.     

Comparison of airborne laser scanning methods for estimating forest structure indicators based on Lorenz curves

The purpose of this study was to compare a number of state-of-the-art methods in airborne laser scanning (ALS) remote sensing with regards to their capacity to describe tree size inequality and other indicators related to forest structure. The indicators chosen were based on the analysis of the Lorenz curve: Gini coefficient (GC), Lorenz asymmetry (LA), the proportions of basal area (BALM) and stem density (NSLM) stocked above the mean quadratic diameter. Each method belonged to one of these estimation strategies: (A) estimating indicators directly; (B) estimating the whole Lorenz curve; or (C) estimating a complete tree list. Across these strategies, the most popular statistical methods for area-based approach (ABA) were used: regression, random forest (RF), and nearest neighbour imputation. The latter included distance metrics based on either RF (NN–RF) or most similar neighbour (MSN). In the case of tree list estimation, methods based on individual tree detection (ITD) and semi-ITD, both combined with MSN imputation, were also studied. The most accurate method was direct estimation by best subset regression, which obtained the lowest cross-validated coefficients of variation of their root mean squared error CV(RMSE) for most indicators: GC (16.80%), LA (8.76%), BALM (8.80%) and NSLM (14.60%). Similar figures [CV(RMSE) 16.09%, 10.49%, 10.93% and 14.07%, respectively] were obtained by MSN imputation of tree lists by ABA, a method that also showed a number of additional advantages, such as better distributing the residual variance along the predictive range. In light of our results, ITD approaches may be clearly inferior to ABA with regards to describing the structural properties related to tree size inequality in forested areas.     

Mapping multiple tree species classes using a hierarchical procedure with optimized node variables and thresholds based on high spatial resolution satellite data

ABSTRACT

Tree species distribution mapping using remotely sensed data has long been an important research area. However, previous studies have rarely established a comprehensive and efficient classification procedure to obtain an accurate result. This study proposes a hierarchical classification procedure with optimized node variables and thresholds to classify tree species based on high spatial resolution satellite imagery. A classification tree structure consisting of parent and leaf nodes was designed based on user experience and visual interpretation. Spectral, textural, and topographic variables were extracted based on pre-segmented images. The random forest algorithm was used to select variables by ranking the impact of all variables. An iterating approach was used to optimize variables and thresholds in each loop by comprehensively considering the test accuracy and selected variables. The threshold range for each selected variable was determined by a statistical method considering the mean and standard deviation for two subnode types at each parent node. Classification of tree species was implemented using the optimized variables and thresholds. The results show that (1) the proposed procedure can accurately map the tree species distribution, with an overall accuracy of over 86% for both training and test stages; (2) critical variables for each class can be identified using this proposed procedure, and optimal variables of most tree plantation nodes are spectra related; (3) the overall forest classification accuracy using the proposed method is more accurate than that using the random forest (RF) and classification and regression tree (CART). The proposed approach provides results with 3.21% and 7.56% higher overall land cover classification accuracy and 4.68% and 10.28% higher overall forest classification accuracy than RF and CART, respectively.     

Update and spatial extension of strategic forest inventories using time series remote sensing and modeling

Up-to-date forest inventory information relating the characteristics of managed and natural forests is fundamental to sustainable forest management and required to inform conservation of biodiversity and assess climate change impacts and mitigation opportunities. Strategic forest inventories are difficult to compile over large areas and are often quickly outdated or spatially incomplete as a function of their long production cycle. As a consequence, automated approaches supported by remotely sensed data are increasingly sought to provide exhaustive spatial coverage for a set of core attributes in a timely fashion. The objective of this study was to demonstrate the integration of current remotely-sensed data products and pre-existing jurisdictional inventory data to map four forest attributes of interest (stand age, dominant species, site index, and stem density) for a 55 Mha study region in British Columbia, Canada. First, via image segmentation, spectrally homogenous objects were derived from Landsat surface-reflectance pixel composites. Second, a suite of Landsat-based predictors (e.g., spectral indices, disturbance history, and forest structure) and ancillary variables (e.g., geographic, topographic, and climatic) were derived for these units and used to develop predictive models of target attributes. For the often difficult classification of dominant species, two modelling approaches were compared: (a) a global Random Forests model calibrated with training samples collected over the entire study area, and (b) an ensemble of local models, each calibrated with spatially constrained local samples. Accuracy assessment based upon independent validation samples revealed that the ensemble of local models was more accurate and efficient for species classification, achieving an overall accuracy of 72% for the species which dominate 80% of the forested areas in the province. Results indicated that site index had the highest agreement between predicted and reference (R² = 0.74, %RMSE = 23.1%), followed by stand age (R² = 0.62, %RMSE = 35.6%), and stem density (R² = 0.33, %RMSE = 65.2%). Inventory attributes mapped at the image-derived unit level captured much finer details than traditional polygon-based inventory, yet can be readily reassembled into these larger units for strategic forest planning purposes. Based upon this work, we conclude that in a multi-source forest monitoring program, spatially localized and detailed characterizations enabled by time series of Landsat observations in conjunction with ancillary data can be used to support strategic inventory activities over large areas.     

Tree species identification within an extensive forest area with diverse management regimes using airborne hyperspectral data

Information on tree species composition is crucial in forest management and can be obtained using remote sensing. While the topic has been addressed frequently over the last years, the remote sensing-based identification of tree species across wide and complex forest areas is still sparse in the literature. Our study presents a tree species classification of a large fraction of the Białowieża Forest in Poland covering 62 000 ha and being subject to diverse management regimes. Key objectives were to obtain an accurate tree species map and to examine if the prevalent management strategy influences the classification results. Tree species classification was conducted based on airborne hyperspectral HySpex data. We applied an iterative Support Vector Machine classification and obtained a thematic map of 7 individual tree species (birch, oak, hornbeam, lime, alder, pine, spruce) and an additional class containing other broadleaves. Generally, the more heterogeneous the area was, the more errors we observed in the classification results. Managed forests were classified more accurately than reserves. Our findings indicate that mapping dominant tree species with airborne hyperspectral data can be accomplished also over large areas and that forest management and its effects on forest structure has an influence on classification accuracies and should be actively considered when progressing towards operational mapping of tree species composition.     

Developing a gross primary production model for coniferous forests of northeastern USA from MODIS data

Accurate estimation of ecosystem carbon fluxes is crucial for understanding the feedbacks between the terrestrial biosphere and the atmosphere and for making climate-policy decisions. A statistical model is developed to estimate the gross primary production (GPP) of coniferous forests of northeastern USA using remotely sensed (RS) radiation (land surface temperature and near-infra red albedo) and ecosystem variables (enhanced vegetation index and global vegetation moisture index) acquired by the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor. This GPP model (called R-GPP-Coni), based only on remotely sensed data, was first calibrated with GPP estimates derived from the eddy covariance flux tower of the Howland forest main tower site and then successfully transferred and validated at three other coniferous sites: the Howland forest west tower site, Duke pine forest and North Carolina loblolly pine site, which demonstrate its transferability to other coniferous ecoregions of northeastern USA. The proposed model captured the seasonal dynamics of the observed 8-day GPP successfully by explaining 84–94% of the observed variations with a root mean squared error (RMSE) ranging from 1.10 to 1.64 g C/m²/day over the 4 study sites and outperformed the primary RS-based GPP algorithm of MODIS.     

用地基激光雷达提取单木结构参数——以白皮松为例

以白皮松(Pinus bungeana Zucc)为研究对象,针对地基激光雷达TLS扫描的3维点云数据在单株木垂直方向的分布特征,提出了一种基于体元化方法的树干覆盖度变化检测方法,获取单木枝下高;然后根据获取的枝下高引入2维凸包算法获取垂直方向分层树冠轮廓,并计算树冠体积和冠幅;同时获取的单木参数还有胸径与树高。结果表明:单木枝下高的估测精度较高,R2与RMSE分别为0.97 m和0.21 m;胸径估测结果的R2与RMSE分别为0.79 cm和1.07 cm;采用逐步线性回归方法建立单木树冠体积与其他单木参数的相关关系,模型变量包括冠幅、叶子填充树冠长度和胸径,样本数为20,模型的R2与RMSE分别是0.967 m3和2.64 m3。本文方法能较准确地估测枝下高,TLS数据具有对树冠结构3维建模的潜力。     

Using multi-date high spectral resolution data to assess the physiological status of macroscopically undamaged foliage on a regional scale

Forests play an important role in regulation of the global climate; moreover, they provide human beings with a whole range of ecosystem services. Forest health and ecosystem functioning have been influenced by anthropogenic activities and their consequences, such as air pollution, surface mining, heavy metal contamination, and other biotic and abiotic stress factors, which had an especially serious effect on central Europe. Many aspects of the physiological state of trees are more or less related to the concentrations of two main groups of leaf photosynthetic pigments: chlorophylls and carotenoids. Therefore, their contents can be used as non-specific indicators of the actual tree physiological status, stress and the pre-visible tree damage. Variations in leaf biochemical composition affect foliar optical properties and can be assessed remotely using high spectral resolution data (hyperspectral data). These data were successfully used in earlier studies to detect vegetation stress and damage. However, only a few approaches have dealt with the use of hyperspectral remote sensing to assess vegetation physiological status on a regional scale. Moreover, little or no research has been done on assessing vegetation health while utilizing multi-date hyperspectral images.In this study, the method for assessing forest health conditions using optical indices retrieved from hyperspectral data was applied to the two temporal HyMap date sets acquired in 07/2009 and 08/2010 to detect stress for the Norway spruce forests in Sokolov, NW Bohemia, a region affected by long-term extensive mining. The classification results were validated by ground truth data (total chlorophyll – Cab, carotenoids – Car and carotenoid to chlorophyll ratio – Car/Cab) and were associated with the geochemical conditions of the forest stands. Both biochemical analysis of the sampled foliage and classification of 2009 and 2010 hyperspectral image identified the same sites affected by vegetation stress. In addition to higher Car/Cab, which enabled detection of the stressed trees using hyperspectral image data, these sites showed critically low pH and lower values for the macronutrient parameters in both organic horizons and, in addition, both sites exhibit critically low base cation to aluminum ratios (Bc/Al) for lower organic and top mineral (0–20 cm) soil horizons.The results of this study demonstrate (i) the potential application of hyperspectral remote sensing as a rapid method of identifying tree stress prior to symptom expression, and (ii) the added value of multi-temporal approaches for hyperspectral data and its further potential for monitoring forest ecosystems.