Estimating single-tree branch biomass of Norway spruce with terrestrial laser scanning using voxel-based and crown dimension features

Abstract

Many remote sensing-based methods estimating forest biomass rely on allometric biomass models for field reference data. Terrestrial laser scanning (TLS) has emerged as a tool for detailed data collection in forestry applications, and the methods have been proposed to derive, e.g. tree position, diameter-at-breast-height, and stem volume from TLS data. In this study, TLS-derived features were related to destructively sampled branch biomass of Norway spruce at the single-tree level, and the results were compared to conventional allometric models with field measured diameter and height. TLS features were derived following two approaches: one voxel-based approach with a detailed analysis of the interaction between individual voxels and each laser beam. The features were derived using voxels of size 0.1, 0.2, and 0.4 m, and the effect of the voxel size was assessed. The voxel-derived features were compared to features derived from crown dimension measurements in the unified TLS point cloud data. TLS-derived variables were used in regression models, and prediction accuracies were assessed through a Monte Carlo cross-validation procedure. The model based on 0.4 m voxel data yielded the best prediction accuracy, with a root mean square error (RMSE) of 32%. The accuracy was found to decrease with an increase in voxel size, i.e. the model based on the 0.1 m voxel yielded the lowest accuracy. The model based on crown measurements had an RMSE of 34%. The accuracies of the predictions from the TLS-based models were found to be higher than from conventional allometric models, but the improvement was relatively small.     

Predicting probability of A-quality lumber of Scots pine (Pinus sylvestris L.) prior to or concurrently with logging operation

Knot properties have a profound influence on the suitability of wood for many wood products leading to significant value differences between different quality grades. It would therefore be rather advantageous to maximise the volume of good quality timber attained from the logs. The objective of this study was to assess how well A-quality lumber of Scots pine derived from log tomography features can be predicted with characteristics measured prior to or concurrently with the logging operation. The study is based on field experiments and X-ray scanning of 204 stems from southern Finland in 2014. We employed mixed logistic regression techniques to model the relationship between the main stem characteristics and probability of A-quality lumber. From the tree characteristics that can be measured or detected from standing trees, the height from the ground level to the lowest dead branch was found to be the best predictor of A-quality lumber. From the characteristics that could, at least in theory, be detected and measured at the moment of harvest, early growth rate and size of tree were found to be the best combination for predicting the probability of A-class quality.     

Modelling the economically viable wood in the crown of European beech trees

Long-term forest development programs in Germany aim on an increase of close-to-nature broadleaf forest stands. This means that the economic importance of European beech is expected to increase. The economic potential of a tree basically consists of the stem as well as the economically viable wood volume in the crown. Due to the high morphological variability of European beech crowns, taper models are often not satisfactory for predicting the economically viable wood volume arising from crowns. Prediction models with a higher precision are recently still lacking. Aim of this study is thus the development of prediction model for the economically viable crown wood volume of European beech trees.We determined the distribution of the wood volume in the crown over the branch diameters using the multistage ‘randomized branch sampling’ method (RBS). The tree-specific wood volume distribution on the branch diameters were used to cluster all sampled trees into 3 groups. Additionally, we developed a method able to distinguish between economically viable and unviable crown branches. Basing on the RBS measurements as well as revenues and processing costs, we modeled the economically viable wood volume from the crown for each tree. To calculate the wood volume under bark, we parameterized a bark thickness function from disk samples of the trees.We showed that the European beech crowns could be clustered into 3 groups differing in their wood volume distribution. The economically viable wood volume in the crown significantly depended on this grouping parameter as well as diameter at breast height (DBH). By contrast, the total amount of wood in the crown only depended on DBH. The differing viable wood volumes in the crowns were thus explained by different wood distributions and not by differing total crown wood volume. To make the results applicable in practice forestry, the modelling results were used to develop a regression formula able to predict the economically viable wood volume in the crown depending on the DBH and the crown type. As the crown type can also be predicted via measurable tree covariates, the regression model of the viable wood volume in the crown can be used as a support tool for the management of European beech stands. Sensitivity analysis quantifies how harvest revenues and costs translate into different viable tree volume.     

Additive biomass equations for small diameter trees of temperate mixed deciduous forests

The carbon stored in small diameter trees of temperate forests has been ignored in most studies and there is a lack of biomass equations for this component of forests. We harvested nine main tree species at sapling stage (dbh?相似文献   

Estimation of stem attributes using a combination of terrestrial and airborne laser scanning

Properties of individual trees can be estimated from airborne laser scanning (ALS) data provided that the scanning is dense enough and the positions of field-measured trees are available as training data. However, such detailed manual field measurements are laborious. This paper presents new methods to use terrestrial laser scanning (TLS) for automatic measurements of tree stems and to further link these ground measurements to ALS data analyzed at the single tree level. The methods have been validated in six 80 × 80 m field plots in spruce-dominated forest (lat. 58°N, long. 13°E). In a first step, individual tree stems were automatically detected from TLS data. The root mean square error (RMSE) for DBH was 38.0 mm (13.1 %), and the bias was 1.6 mm (0.5 %). In a second step, trees detected from the TLS data were automatically co-registered and linked with the corresponding trees detected from the ALS data. In a third step, tree level regression models were created for stem attributes derived from the TLS data using independent variables derived from trees detected from the ALS data. Leave-one-out cross-validation for one field plot at a time provided an RMSE for tree level ALS estimates trained with TLS data of 46.0 mm (15.4 %) for DBH, 9.4 dm (3.7 %) for tree height, and 197.4 dm³ (34.0 %) for stem volume, which was nearly as accurate as when data from manual field inventory were used for training.     

Modeling internode length and branch characteristics for Pinus radiata in New Zealand

It is recognized that estimation of internode length and maximum branch size is important for the prediction of clearwood in unpruned timber stands, as well as for evaluating the quality and value of logs in general. A review of existing branch models reveals a diversity of approaches as well as a tendency for results to be species specific. Here, a branch model is developed for Pinus radiata in New Zealand, capable of predicting successive internode lengths, the number of branches on each branch cluster, and the size of each branch up to the green crown (GC) at site index age 20. Inputs to the model include tree height and diameter at breast height (dbh) (both at age 20) and basal area per hectare of the top 100 stems. Further optional inputs are an ocular count of the number of branch clusters up to the green crown, and branch factor (BF)—the size of the biggest branch in the first cluster encountered above 6 m. The vertical distributions of internode length and maximum branch diameter are found to reach maxima around 0.3–0.4 of relative height. Internode length and the number of branches per whorl were found to be independent of tree size, site index and stand density. Stems per hectare is not required as an explicit predictor variable but it appears implicitly through tree dbh. Some model output is given and the results are discussed.     

Reconstruction of Pinus Sylvestris knots using measurable log features in the Swedish Pine Stem Bank

Abstract

The objective of this study was to develop a method for reconstruction of parametrically described whorls and knots from data possible to extract from industrial scanning of logs, using X-ray scanners. The method was conceived using the logs in the Swedish Pine Stem Bank as a foundation, and was based on a few predictor features extracted from these logs; namely whorl volume, distance between whorls and distance between pith and surface. These features were not measured in images but calculated from existing parameterised knots. Simulated test sawing shows that the reconstruction method results in a representative model of the knot structure in the log, when considering the grade distribution of the sawn timber produced by the simulation program. The results of this study could, for instance, be used for improved online quality predictions at sawmills. One step in this direction is to use industrial X-ray data to enlarge the amount of log data available for sawing simulation research. Future work can, therefore, focus on developing a practical application of the results presented here.     

Assessment of standing wood and fiber quality using ground and airborne laser scanning: A review

Accurate information on the wood-quality characteristics of standing timber and logs is needed to optimize the forest production value chain and to assess the potential of forest resources to meet other services. Physical and chemical characteristics of wood vary with both tree and site characteristics. At the tree scale, crown development, stem shape and taper, branch size and branch location, knot size, type and placement, and age all influence wood properties. More broadly, at the stand level, stocking density, moisture, nutrient availability, climate, competition, disturbance, and stand age have also been identified as key determinants of wood quality. Such information is often captured in polygon based forest inventory data. Other terrain-related spatial information, such as elevation, slope and aspect, can improve assessments of site conditions and limitations upon plant growth which impact wood quality. Light Detection And Ranging (LiDAR) is an emerging technology, which directly measures the three-dimensional structure of forest canopies using ground or airborne laser instruments, and can provide highly accurate information on individual-tree and stand-level forest structure. In this paper, we explore the potential of LiDAR and other geospatial information sources to model and predict wood quality based on individual-tree and stand structural metrics. We identify a number of key wood quality attributes (i.e., basic wood density, cell perimeter, cell coarseness, fiber length, and microfibril angle) and demonstrate links between these properties and forest structure and site attributes. Finally, the potential for using LiDAR in combination with other geospatial information sources to predict wood quality in standing timber is discussed.     

The effects of timing of pre-commercial thinning and stand density on stem form and branch characteristics of Pinus sylvestris

The effects of pre-commercial thinning on tree and branch characteristicswere analysed at two experimental sites in northern Sweden,where pre-thinning of Scots pine stands to 600, 1000 and 1800stems per hectares at heights of 1.5, 3, 5 and 7 m had beencombined in a factorial design. A total of 90 trees were cutand branches were measured in whorls sampled at 1-m intervals,starting with the whorl closest to 0.5 m above ground. Effectsof the treatments on tree and branch characteristics were analysedby a fixed-effect analysis of variance model, which was appliedseparately for each experimental site and whorl height. Diameterat breast height (d.b.h.) and living crown to height ratio decreased,while the height/d.b.h. ratio (stem form) increased with increasingstand density, and generally with increased height at thinning.No treatment effect was found on the number of branches perwhorl or branch angle, but branch diameter (BD) was significantlyinfluenced by both stand density after thinning and height atthinning. BD decreased with increases in stand density and decreasedwith increases in height at pre-commercial thinning. Relativebranch size (RBS), defined as the ratio between the sum of thediameters of branches in a whorl and stem perimeter under barkat the location of that whorl, was significantly influencedonly by stand height at the time of treatment. RBS decreasedby 4–6 per cent in whorls located below the current baseof the living crown with each step increase in stand heightat pre-commercial thinning. The results indicate that the numberof stems after pre-commercial thinning has a greater impacton tree size and shape than timing of pre-commercial thinning.Stand density after thinning also has a greater impact on branchsize, in the sense that it will affect branch size further upthe tree than the timing of pre-commercial thinning. However,by delaying the time for pre-commercial thinning it is possibleto reduce the RBS in the lower part of the stem. This mightbe an important measure to improve timber quality later in therotation period, which cannot be done solely by regulating standdensity.     

Influence of Initial Spacing on Growth and Branching Characteristics of Cordia africana Trees Established on Eritrean Highland

The study reports on the influence of initial spacing on growth and branching habit of Cordia africana trees established on Eritrean highland. The initial spacings used in this study were 1.0×1.0 m, 1.5×1.5 m, 2.0×2.0 m, 2.5×2.5 m, 3.0×3.0 m, and 3.5×3.5 m; these are equivalent to 1 m², 2.25 m², 4 m², 6.25 m², 9 m² and 12.25 m² in growing space per tree, respectively. No significant relationship between spacing and tree height was found, but increased spacing increased crown diameter and root-collar-diameter. On the other hand, an increase in spacing reduced wood quality because it increased branch diameter (knot size). Thus, from a wood quality perspective, adopting close spacing in plantations of C. africana could be advantageous.     

Effect of Betula pendula Admixture on Tree Growth and Branch Diameter in young Pinus sylvestris Stands in Southern Finland

The effects of a birch admixture on the height and diameter growth and maximum branch diameter in planted Scots pine stands was studied using models constructed with a data set from 13 stands of 9–16 yrs of age and 2–8 m dominant height on average sites on mineral soils in southern Finland. The density and height of the birch varied highly between and within stands. Simulated results indicated that the pines were capable of keeping up in height growth with birches that had originated from seed. Even a very high number of birches (10?000 stems ha^?1) had virtually no effect on the height growth of the pines. The number of birches had a pronounced effect on the diameter growth and the maximum branch diameter in pine. Retention of a temporary birch component in young pine stands seems a feasible way of mitigating the adverse effects of low planting densities on the external quality of pine.     

Density measurements in Pinus sylvestris sawlogs combining X-ray and three-dimensional scanning

Abstract

Wood density is an important quality variable, closely related to the mechanical properties of the wood. Precise wood density measurements in the log sorting would enable density sorting of logs for products such as strength-graded wood and finger-jointed wood. Density sorting of logs would also give more homogeneous drying properties and thus improve the quality of the final products. By compensating the radiographs from an X-ray log scanner for the varying path lengths using outer shape data from a three-dimensional (3D) scanner, it is possible to make precise estimates of both green and dry density. Measurements on simulated industrial data were compared with densities measured in computed tomographic (CT) images for 560 Scots pine (Pinus sylvestris L.) logs. It was found that green sapwood density could be measured with predictability R ²=0.65 and root mean square error (RMSE) of 25 kg m^?3. Green and dry heartwood densities were measured with similar precision: R ²=0.79 and RMSE=32 kg m^?3 for green density and R ²=0.83 and RMSE=32 kg m^?3 for dry density.     

Predicting knottiness of Pinus sylvestris for use in tree bucking procedures

Abstract

The quality and accompanying value of Scots pine (Pinus sylvestris L.) lumber varies markedly in the Nordic countries. Consequently, tree bucking has a great impact on the economic result. The objectives of this study were (1) to analyse which knot characteristics usually reduce pine lumber sawn from the butt-logs to the next quality grade, (2) to determine the most appropriate characteristics that could be used in predicting lumber quality, and (3) to develop models that could be used in practice for predicting the probability of certain quality grades. The study was based on field experiments and test sawing data on 100 Scots pine stems from south-western Finland. Since the results showed that the maximum dead knot is the most crucial knot characteristic when the first cross-cutting point is determined, models were created that predict the height of the first dead knot that lowers grade A to B (B-grade dead knot). Both early growth rate and dead branch height should be measured to predict pine butt-log quality. Early growth rate seems to be appropriate in predicting between-stand variation, while dead branch height is appropriate for predicting within-stand variation.     

An application of mixed-effects model to evaluate the effects of initial spacing on radial variation in wood density in Japanese larch (Larix kaempferi)

The effects of initial tree spacing on wood density at breast height were examined for 22-year-old Japanese larch (Larix kaempferi). The experiment involved the use of three plots with different initial tree spacing densities (300, 500, and 1000 trees/ha). For five trees selected from each plot, the total tree height, diameter at breast height, height to the base of the live crown, and crown diameter were measured. Ring width and wood density for individual growth rings were determined by X-ray densitometry. A mixed-effects model was applied for fitting the radial variation in wood density in relation to initial spacing. Models having various mean and covariance structures were tested in devising an appropriate wood density model. The model, consisting of the mean structure with quadratic age effects and heterogeneous first-order autoregressive covariance, was able to describe the radial variation in wood density. Closer spacing of trees (1000 trees/ha) resulted in a faster increase in wood density from the pith outward than for more widely spaced trees, indicating that initial tree spacing may influence the age of transition from juvenile to mature wood.     

An improved technique for non-destructive measurement of the stem volume of standing wood

An improved technique, cheaper and less time-consuming, to measure standing wood volume by using an electronic theodolite was tested, by which greater information from the forest could be acquired accurately and non-destructively. This was achieved by recording the diameter at breast height and ground-level diameter of a tree as well as the included angle between the electronic theodolite and the left and right tangents of the stem at any point. The standing wood volume then was computed precisely by section. In addition, the factors that influence the precision of the method (observable distance and number of segments) were also analysed. In the study, 175 Larix gmelinii (Rupr.) Kuzen. trees and 190 Populus tomentosa Carrière trees chosen randomly for sampling were measured with the electronic theodolite, and then were cut down for measurement of the average cross-section volume. Based on the data acquired from 100 sample trees, a standard volume table was compiled, and then the data for the remaining random 10 sample trees in each group were selected for a comparison test. The results indicated that the optimal distance for indirect observation should be as high as the sample tree, and the optimal visual distinguished section was about 2?m. The correlation coefficient between the value measured non-destructively and the value of the felled trees of L. gmelinii was 0.97, with an average relative error of 1.62%. With regard to P. tomentosa, the correlation coefficient between the two values obtained by the two methods was 0.905 with an average relative error of 8.40%. It was concluded that the standard volume model based on the non-destructive measurement technique meets the requirements for precision in forest surveys. The precision of the standard volume model for L. gmelinii (a coniferous tree) was superior to that of the model for P. tomentosa (a broad-leaved tree). The electronic theodolite method provides an alternative technique for measuring trees without destructive sampling and is widely applicable for forest surveys.     

Branch characteristics of widely spaced Douglas-fir in south-western Germany: Comparisons of modelling approaches and geographic regions

Models of Douglas-fir branch and whorl characteristics were developed from contrasting spacing experiments in southwest Germany. The dataset was based on 100 young (20–30 years old), unpruned and partially pruned trees from a 100, 200, and 1200 stems ha⁻¹ spacing experiment on Douglas-fir that was replicated 3 times across the region. The material was used to predict (1) the number of branches whorl⁻¹, (2) branch angle, (3) status (living/dead) of the branches within the living crown, (4) maximum branch diameter whorl⁻¹, and (5) relative diameter of branches within a whorl. For each of these models (except branch status), both a linear and nonlinear, generalised hierarchical mixed effects equation was developed. The comparison of the linear and nonlinear approaches showed that both had a relatively similar level of bias, but the nonlinear equations generally performed better (reduction in mean absolute error of 1.1–69.5%). Overall, individual branch and tree properties were sufficient to give logical and precise predictions of the branch characteristics for the models across the range of sampled stand densities. In addition, the models showed a similar behaviour compared to models on Douglas-fir crown structure from the Pacific Northwest, USA. This suggests that the allometric relationship between tree size and branch characteristics for a given species may be relatively consistent across regions, even ones with highly contrasting growing conditions like in this study. The models performed well across a range of stand conditions and now will be further integrated into an individual tree growth and yield simulations system.     

材性改良研究：Ⅰ.X射线木材密度测定

利用Ｘ射线衍射（Ｄ－ｍａｘ３Ｂ型）改装成直接扫描式Ｘ射线木材密度计，可快速、高效测定木材密度组成，即木材平均密度，平均早材密度，平均晚材密度，最大木材密度，最小木材密度，木材密度梯度和木材密度的变异幅度，为材性改良提供了新手段，本文用Ｘ射线木材密度计测定１７个树种木材密度。     

A Windthrow-risk Estimation for Coniferous Trees

A method of windthrow-risk estimation for young coniferous treesis proposed in this paper. This estimation is deduced from stembending theory. Fourteen pairs, each containing one dominant and one suppressedtree, were measured m a 15-year-old Scots pine thicket. Totallength of needled branches and stem diameter in the middle ofinter-whorl sections were measured for each whorl of each tree. The data were used to evaluate and compare growth strategy riskfor dominant and overtopped trees in the thicket. The strategieswere found to be qualitatively different. The suppressed treegrowth strategy was dependent on that of the dominant tree.Certain indices of ‘shelter effect’ for suppressedtree were also found.