Identification of Arceuthobium globosum using unmanned aerial vehicle images in a high mountain forest of central Mexico

The identification of forests infested by parasitic plants is important for the design of appropriate control and prevention strategies. Satellite images and geographic information systems are commonly used to analyze the presence of pest and parasitic plants in the forests. However there is a need for finer resolution. In the last decade,the use of unmanned aerial vehicles has become increasingly common for capturing aerial images(10 cm per pixel). The objective of the study was to obtain RGB values(Red, Green and Blue) through the colorimetric ranges for use in identification of Yellow Dwarf Mistletoe(YDM)(Arceuthobium globosum) in aerial images taken in a forest of central Mexico via a programmed algorithm.Three tonalities of YDM were classified according to its phenological stages, viz. green(young stage), yellow(adult stage), and brown(senescence stage), considering two light intensities, sunny and cloudy. Non-parametric tests were used in statistical analyses. The Spearman test showed significant negative correlation(p 0.001) between phenological stage and colour, indicating that lower RGB values were associated with greater age. The differences between groups were analysed using Kruskal–Wallis(p 0.01) and Mann–Whitney tests(p 0.01). The applied algorithm identified the presence and predominant colours of YDM according to its phenological stage.     

Monitoring of eucalyptus sprouts control using digital images obtained by unmanned aerial vehicle

This study evaluated the effectiveness of chemical control for eradicating eucalyptus sprouts using images obtained with Unmanned Aerial Vehicle (UAV). The study was carried out in eucalyptus plantations in Itabela, BA, in replanting areas during pre-planting application of herbicide. Aerial images obtained by a UAV were used to evaluate the effectiveness of the herbicide application for sprouts control. After the images were acquired, they were processed to calculate the normalized difference vegetation index (NDVI) and submitted to a supervised classification to quantify the percentage reduction of sprout green matter. The percentage data were submitted to analysis of variance and the means were compared by Tukey test at 5%. The differences, both visual and average percentages were observed only during the evaluation period with no evident effect of the treatments on sprout control. The images provided by the UAV allowed to monitor and identify visually the plots where biomass reduced or increased and to evaluate the effectiveness of chemical control of eucalyptus sprouts, indicating areas where it was nonexistent, partial or total.     

Tree species classification using deep learning and RGB optical images obtained by an unmanned aerial vehicle

The diversity of tree species and the complexity of land use in cities create challenging issues for tree species classification.The combination of deep learning methods and RGB optical images obtained by unmanned aerial vehicles(UAVs) provides a new research direction for urban tree species classification.We proposed an RGB optical image dataset with 10 urban tree species,termed TCC10,which is a benchmark for tree canopy classification(TCC).TCC10 dataset contains two types of data:tree canopy images with simple backgrounds and those with complex backgrounds.The objective was to examine the possibility of using deep learning methods(AlexNet,VGG-16,and ResNet-50) for individual tree species classification.The results of convolutional neural networks(CNNs) were compared with those of K-nearest neighbor(KNN) and BP neural network.Our results demonstrated:(1) ResNet-50 achieved an overall accuracy(OA) of 92.6% and a kappa coefficient of 0.91 for tree species classification on TCC10 and outperformed AlexNet and VGG-16.(2) The classification accuracy of KNN and BP neural network was less than70%,while the accuracy of CNNs was relatively higher.(3)The classification accuracy of tree canopy images with complex backgrounds was lower than that for images with simple backgrounds.For the deciduous tree species in TCC10,the classification accuracy of ResNet-50 was higher in summer than that in autumn.Therefore,the deep learning is effective for urban tree species classification using RGB optical images.     

Assessing the impacts of species composition, top height and density on individual tree height prediction of quaking aspen in boreal mixedwoods

In this study we developed an individual tree height prediction model for quaking aspen (Populus tremuloides Michx.) grown in boreal mixedwood forests in Alberta using the nonlinear mixed model (NLMM) approach. We examined the impacts of density, species composition, and top height on aspen height predictions. Statistically significant stand level variables were incorporated into the base height–diameter model to increase the predictive ability and accuracy of the model at both the population and subject-specific levels. Our analyses showed that top height and density impacted height growth, but species composition did not. More importantly, we found that the inclusion of additional variables into the base model, despite improving model fitting statistics on the modelling data, did not improve the model's predictive ability and accuracy when cross-validated and when tested on an independent testing data set. Under the NLMM framework the base model performed as well as or better than the expanded models that contained other stand level variables. This has important theoretical and practical implications because, other than for biological reasons, more accurate local tree height predictions for aspen can be achieved simply by using the base height–diameter model fitted with the NLMM approach without the inclusion of other variables.     

Estimation of forest stem volume using multispectral optical satellite and tree height data in combination

Abstract

The accuracy of forest stem volume estimation at stand level was investigated using multispectral optical satellite and tree height data in combination. The stem volumes for the investigated coniferous stands, located in southern Sweden, were in the range of 15–585 m³ ha^?1 with an average stem volume of 266 m³ ha^?1. The results from regression analysis showed a substantial improvement for the combined stem volume estimates compared with using satellite data only. The accuracy in terms of root mean square error (RMSE) was calculated to 11.2% of the average stem volume using SPOT-4 data and tree height data in combination compared with 23.9% using SPOT-4 data only. By replacing SPOT-4 data with Landsat TM data the RMSE was improved from 25.2% to 12.2%. In addition, a sensitivity analysis was performed on the combined stem volume estimates by adding random errors, normally distributed with zero expectations, with standard deviations of 1, 1.5 and 2 m to tree height data. The results showed that the RMSE increased with increasing random tree height error to 15.4%, 18.0% and 19.9% using SPOT-4 data and 16.3%, 19.2% and 21.2% using Landsat TM data. The results imply that multispectral optical satellite data in combination with accurate tree height data could be used for standwise stem volume estimation in forestry applications.     

Prediction of tree height,basal area and stem volume in forest stands using airborne laser scanning

The aim of this study was to develop prediction models using laser scanning for estimation of forest variables at plot level, validate the estimations at stand level (area 0.64 ha) and test the effect of different laser measurement densities on the estimation errors. The predictions were validated using 29 forest stands (80×80 m²), each containing 16 field plots with a 10 m radius. For the best tested case, mean tree height, basal area and stem volume were predicted with a root mean square error of 0.59 m (3% of average value), 2.7 m² ha^?1 (10% of average value) and 31 m³ ha^?1 (11% of average value), respectively, at stand level. There were small differences in terms of prediction errors for different measuring densities. The results indicate that mean tree height, basal area and stem volume can be estimated in small stands with low laser measurement densities producing accuracies similar to traditional field inventories.     

Deciduous forest mapping using change detection of multi-temporal canopy height models from aerial images acquired at leaf-on and leaf-off conditions

Discrimination of deciduous trees using spectral information from aerial images has only been partly successfully due to the complexity of the reflectance at different view angles, times of acquisition, phenology of the trees and inter-tree radiance. Therefore, the objective was to evaluate the accuracy of estimating the proportion of deciduous stem volume (P) utilizing change detection between canopy height models (CHMs) generated by digital photogrammetry from leaf-on and leaf-off aerial images instead of using spectral information. The study was conducted at a hemi-boreal study area in Sweden. Using aerial images from three seasons, CHMs with a resolution of approximately 0.5?m were generated using semi-global matching. For training plots, metrics describing the change between leaf-on and leaf-off conditions were calculated and used to model the continuous variable P, using the Random Forest approach. Validated at sub-stands, the estimation accuracy of P in terms of root mean square error and bias was found to be 18% and ?6%, respectively. The overall classification accuracy, using four equally wide classes, was 83% with a kappa value of 0.68. The validation plots in classes of high proportion of coniferous or deciduous stem volume were well classified, whereas the mixed forest classes showed lower classification accuracies.     

Changes in photosynthesis and leaf characteristics with tree height in five dipterocarp species in a tropical rain forest

Variations in leaf photosynthetic, morphological and biochemical properties with increasing plant height from seedlings to emergent trees were investigated in five dipterocarp species in a Malaysian tropical rain forest. Canopy openness increased significantly with tree height. Photosynthetic properties, such as photosynthetic capacity at light saturation, light compensation point, maximum rate of carboxylation and maximum rate of photosynthetic electron transport, all increased significantly with tree height. Leaf morphological and biochemical traits, such as leaf mass per area, palisade layer thickness, nitrogen concentration per unit area, chlorophyll concentration per unit dry mass and chlorophyll to nitrogen ratio, also changed significantly with tree height. Leaf properties had simple and significant relationships with tree height, with few intra- and interspecies differences. Our results therefore suggest that the photosynthetic capacity of dipterocarp trees depends on tree height, and that the trees adapt to the light environment by adjusting their leaf morphological and biochemical properties. These results should aid in developing models that can accurately estimate carbon dioxide flux and biomass production in tropical rain forests.     

Detection of individual trees and estimation of tree height using LiDAR data

For estimation of tree parameters at the single-tree level using light detection and ranging (LiDAR), detection and delineation of individual trees is an important starting point. This paper presents an approach for delineating individual trees and estimating tree heights using LiDAR in coniferous (Pinus koraiensis, Larix leptolepis) and deciduous (Quercus spp.) forests in South Korea. To detect tree tops, the extended maxima transformation of morphological image-analysis methods was applied to the digital canopy model (DCM). In order to monitor spurious local maxima in the DCM, which cause false tree tops, different h values in the extended maxima transformation were explored. For delineation of individual trees, watershed segmentation was applied to the distance-transformed image from the detected tree tops. The tree heights were extracted using the maximum value within the segmented crown boundary. Thereafter, individual tree data estimated by LiDAR were compared to the field measurement data under five categories (correct delineation, satisfied delineation, merged tree, split tree, and not found). In our study, P. koraiensis, L. leptolepis, and Quercus spp. had the best detection accuracies of 68.1% at h = 0.18, 86.7% at h = 0.12, and 67.4% at h = 0.02, respectively. The coefficients of determination for tree height estimation were 0.77, 0.80, and 0.74 for P. koraiensis, L. leptolepis, and Quercus spp., respectively.     

Estimation of mean tree height using small-footprint airborne LiDAR without a digital terrain model

In order to estimate mean tree height using small-footprint airborne light detection and ranging (LiDAR) data, a digital terrain model (DTM), which is a continuous elevation model of the ground surface, is usually required. However, generating accurate DTMs in mountainous forests using only the LiDAR data is laborious and time consuming, because it requires human-assisted methods, especially in the forests with poor laser penetration rates. Based on our previous finding that a hypothetical continuous surface model passing through the predominant tree tops (hereafter, called the “top surface model” or TSM) might be nearly parallel to a DTM, we assumed that the vertical difference between the TSM and the ground return was the mean tree height. According to this assumption, we propose a new methodology that does not require a DTM to estimate mean tree height. This method completely, automatically, and directly estimates mean tree height (MTH _E) from the LiDAR data without requiring a regression analysis using reference data. From the relationships between the MTH _E and the observed mean tree height (MTH _O) in different hinoki cypress forests, we demonstrate that this method effectively estimates the mean tree height with nearly 1-m accuracy.     

Modelling joint distribution of tree diameter and height using Frank and Plackett copulas

Bivariate distribution models are veritable tools for improving forest stand volume estimations. Their accuracy depends on the method of construction. To-date,most bivariate distributions in forestry have been constructed either with normal or Plackett copulas. In this study, the accuracy of the Frank copula for constructing bivariate distributions was assessed. The effectiveness of Frank and Plackett copulas were evaluated on seven distribution models using data from temperate and tropical forests. The bivariate distributions include: Burr III, Burr XII, Logit-Logistic, Log-Logistic, generalized Weibull,Weibull and Kumaraswamy. Maximum likelihood was used to fit the models to the joint distribution of diameter and height data of Pinus pinaster(184 plots), Pinus radiata(96 plots), Eucalyptus camaldulensis(85 plots) and Gmelina arborea(60 plots). Models were evaluated based on negative log-likelihood(-■). The result show that Frank-based models were more suitable in describing the joint distribution of diameter and height than most of their Plackett-based counterparts. The bivariate Burr III distributions had the overall best performance. The Frank copula is therefore recommended for the construction of more useful bivariate distributions in forestry.     

Variation of basic density,calorific value and volumetric shrinkage within tree height and tree age of Ugandan grown Eucalyptus grandis wood

Insufficient knowledge on Ugandan grown Eucalyptus grandis W.Hill ex Maiden wood properties,high demand,and processing challenges led to a study into its physical properties.We obtained the variation of basic density(BD),calorific value(CV) and volumetric shrinkage(VS) within tree height and tree-age of E.grandis,and its appropriate use based on these properties.Trees with good boles were harvested from Kabarole District in western Uganda to produce specimens as prescribed by British Standards and ASTM standard wood testing procedures.Secondary data reviews and statistical analysis using ANOVA,Tukey's test and multivariate analysis were done to obtain property estimates and their variation within trees and amongst tree ages.The mean BD of E.grandis is 413.6,380.5,471.0,and 501.1 kg m~(-3) at 3,6,9,and 12 years,respectively,showing significant increase with tree age(p=0.003).The pattern of BD with tree height showed a reduction with tree height although with higher values in the middle portion of the tree.The CV increases(p=0.014)with tree age and reduces with tree height with values of14,560.32,15,447.3,16,079.11,and 16,932.6 kJ kg~(-1) at 3,6,9,and 12 years,respectively.The percentage VS was 11.02,9.84,12.31,and 14.45 for 3-,6-,9-,and 12-year-old trees,respectively,and it did not vary significantly(p=0.088)with tree height.Basing on these property values,E.grandis wood could be used for scaffolding,light constructions and fuel wood production.Its seasoning needs to be longer with well monitored drying schedules to reduce seasoning defects caused by the high VS.Further studies on strength properties,seasoning schedules,panel products properties and tree-age chemical variations would improve the knowledge about its wood quality and would enhance its efficient utilization.     

Quantitative estimates of tree species selectivity by moose (Alces alces) in a forest landscape

Abstract

An extensive literature is available on browsing preference for certain tree species. However, useful predictive tools for estimating the impact of deer on forests production and biodiversity can still be improved. A step in that direction is not only to rank preference among tree species but also to quantify the relative risk of being browsed. The foraging selectivity of moose was evaluated using three different statistical methods developed to study habitat utilization. The general pattern for the three methods was consistent. From the results, groups of forage species were clustered and a quantitative index of selectivity was calculated for the groups. The selectivity index showed that rowan (Sorbus aucuparia), willow (Salix ssp.) and aspen (Populus tremula) had a 14 times higher probability of being browsed than a group consisting of Scots pine (Pinus sylvestris) and downy birch (Betula pubescens), while juniper (Juniperus communis) and silver birch (Betula pendula) had a 3.5 times higher probability than Scots pine and downy birch. Since the most preferred species were the least abundant, one should be cautious about the generality of the index between areas, as it may indicate that preference depends on plant species composition. The method used can easily be applied in forest management. Information on quantitative selectivity indices may improve the possibility of managing moose in accordance with acceptable browsing damage.     

密度对华北落叶松人工林树高、胸径和物种多样性的影响

以河北省木兰围场国有林场管理局克勒沟林场华北落叶松人工林为研究对象,对标准地内保留密度为400株/hm2、600株/hm2、900株/hm2、1 100株/hm2的华北落叶松在上坡位、中坡位、下坡位的树高和胸径以及物种多样性进行了调查分析,结果表明:在保留密度为600株/hm2时,林木生长最好,树高最大;保留密度小于600株/hm2时,树高随着密度的增大而增大,但大于600株/hm2时,树高会随着密度的增大而呈现降低的趋势;胸径会随着保留密度的增大而减小;对不同密度样地中的植物重要值及物种多样性指数进行了计算,林下草本层植物的数量、多样性指数在保留密度为600株/hm2时达到最大值,之后随着密度的增加均呈递减的趋势。     

Assessing logging residues availability for energy production by using forest management plans data and geographic information system (GIS)

European Journal of Forest Research - The aim of the work was to quantify the mass of logging residues (branches and tops; t yr?1 dry matter, DM) for energy generation starting from...     

Evaluation of tree and stand-level growth models using national forest inventory data

Two models, Carbware (CW) and Growfor (GF), of different resolution and based on different frameworks were evaluated in relation to stand-level forecasts of volume and basal area using Ireland’s National Forest Inventory (NFI) data. CW is a distance-independent single-tree model that is based on diameter increment. GF is a stand-level dynamic empirical model that uses the von Bertalanffy–Richards growth equation in a state-space framework. NFI data were used as input to the models, and each model’s projections were compared to NFI data at the next measurement cycle. The NFI is a permanent sampling system with the objective to assess the composition and extent of the forest estate. A subset of the NFI was used in the study, single-species even-aged plots comprising Sitka spruce and lodgepole pine. The accuracy and performance of the CW and GF models were analysed using residual analysis and standard statistical techniques. Results show that both models require improvement, though the study has raised concerns regarding the suitability of the NFI data for this type of investigation.     

层次分析法在林种树种结构优化中的应用研究

应用多目标决策方法——层次分析法对小陇山林业实验局的林种、树种结构优化的两重性（即社会经济性和环境生态性）进行了分析探讨,并建立了综合效益层次分析模型。     

Modeling individual tree height growth of Norway spruce and Scots pine from national forest inventory data in Norway

We developed individual tree height growth models for Norway spruce (Picea abies (L.) Karst.) and Scots pine (Pinus sylvestris L.) in Norway based on national forest inventory data. Potential height growth is based on existing dominant height growth models and reduced due to competition by functions developed in this study. Three spatially explicit and two spatially non-explicit competition indices were tested. Distance effects and diameter ratio effects were estimated from the data simultaneously with parameters of the potential modifier functions. Large height measurement errors in the national forest inventory data caused large residual variation of the models. However, the effects of competition on height growth were significant and plausible. The potential modifier functions show that height growth of dominant trees is largely unaffected by competition. Only at higher levels of competition, height growth is reduced as a consequence of competition. However, Scots pine also reduced height growth at very low levels of competition. Distance effects in the spatially explicit competition indices indicated that the closest neighbors are most important for height growth. However, for Scots pine also competitors at larger distance affected height growth. The five competition indices tested in this study explained similar proportions of the variation in relative height growth. Given that unbiased predictions can only be expected for the same plot size, we recommend a spatially explicit index, which describes the distance function with a negative exponential, for use in growth simulators.     

A simple theory to link bole surface area,stem density and average tree dimensions in a forest stand

A geometrical model of a forest stand has been analyzed. A forest stand has been modeled as a population of cones which was described by the change of total bole surface area with density \(\hat{S}(N)\) , relation between density and a horizontal dimension (radius r) r(N), and the relation between vertical dimension (generatrix l) and radius l(r). It has been shown that there are close relationships between \(\hat{S}(N)\) , l(r) and r(N). In case of \(\hat{S}(N) = const\) , power exponent of l(r) can be predicted from the power exponent of r(N) and vice versa. A comparison of the model analysis with the data available on Scots pine (Pinus sylvestris L.) stands has been performed. In spite of the model simplicity, its inferences proved to be workable in many cases where the data can be interpreted as a dynamics of an even-aged forest stand. In particular, if the estimation of total bole surface area is constant, the power exponent in the relation of diameter and stand density DBH(SD) can be calculated on the basis of the power exponent in the relation of height and diameter H(DBH) and vice versa. Possible limitations and the meaning of the analysis are discussed.     

华北落叶松人工幼林林分密度与生物量的关系

华北落叶松（larix principis—rupprechtii Mayr．）是燕山山地的主要树种,生态和经济效益显著。以河北木兰林管局境内13a生华北落叶松人工幼林为研究对象,用样方调查的方法对立地条件相同密度分别为1800、2400和3000株／hm^2的3种林分生物量进行研究,结果表明：林分总生物量随密度的增大而增大,低、中、高3种密度林分总生物量分别为21180．6、26724．6和31352．4kg/hm^2;乔木层生物量最大,占总生物量的98．9％,草本层次之,占0．7％,灌木层最少,仅占0．4％;林木平均木生物量随密度的增大略微减少,密度由低到高其平均木生物量依次为11537．5、11077．3和10357．8g/株;各密度林分乔木层各器官生物量均存在W干〉W枝〉W根〉W叶〉W皮的分布规律;3种密度的林分生物量结构特征基本相同,3000株／hm^2的林分结构合理,更能达到丰产的目的。