Wood Density Estimates Research Articles

Infrared thermography for wood density estimation

Infrared thermography (IRT) is becoming a commonly used technique to non-destructively inspect and evaluate wood structures. Based on the radiation emitted by all objects, this technique enables the remote visualization of the surface temperature without making contact using a thermographic device. The process of transforming radiant energy into temperature depends on many parameters, and interpreting the results is usually complicated. However, some works have analyzed the operation of IRT and expanded its applications, as found in the latest literature.This work analyzes the effect of density on the thermodynamic behavior of timber to be determined by IRT. The cooling of various wood samples has been registered, and a statistical procedure that enables one to quantitatively estimate the density of timber has been designed. This procedure represents a new method to physically characterize this material.

Structure and allometry in tropical forests of Chocó, Colombia

Abstract The very wet forests of Choco on the Pacific region of Colombia are unique in their plant diversity and function but are poorly known in aboveground structure and biomass (AGB). Conservation of these forests from widespread degradation in recent years has resulted in numerous REDD+ projects that require rapid assessment of the carbon stocks and changes. Here, we use a combination of inventory plots and harvested trees to develop allometric models and to quantify the variations of forest structure and carbon density across the Choco region. We used 45 ha of inventory plots sampled in 15 1-ha permanent and 120 temporary 0.25-ha plots set under a stratified design. Likewise, we used 296 and 96 harvested trees and palms, respectively, with diameter ranging between 5 and 158.6 cm to develop the tree and palm allometry for AGB estimation. A model to estimate stem wood density from branch wood density was applied. Coarse woody debris (CWD) was also tallied and the dry mass stored quantified. We evaluated an alternative method to rapidly estimate wood density in tropical forests from sampling small branches. The total mean AGB in the 15 plot clusters was 239.8 ± 73.2 Mg ha−1. The mean percentage of AGB represented by trees, palms, and CWD was 88.6 ± 3.2, 3.1 ± 2.0, and 8.3 ± 3.0, respectively. Overall, our total AGB estimates are similar to those reported for wet lowland tropical forests but lower than those reported in tropical dry-seasonal forests. Our results emphasize on the advantage of employing height-diameter models developed from harvested trees to increase the accuracy of the AGB estimation in tropical forests. The wood density estimation was in good agreement with mean plot estimates made to reliable floristic inventories available for four 1-ha plots. This methodology can be further improved by including a sampling design based on tree diameter classes, which can help overcome the use of “static” databases built on species names. Finally, our findings support the use of the existing regional and pantropical allometries to estimate both tree height and the aboveground biomass in tropical wet forests, but emphasize the need of having more local and regional models to quantify the bias and improve the estimation of forest carbon density across different forest types.

Wood Specific Gravity Variations and Biomass of Central African Tree Species: The Simple Choice of the Outer Wood.

ContextWood specific gravity is a key element in tropical forest ecology. It integrates many aspects of tree mechanical properties and functioning and is an important predictor of tree biomass. Wood specific gravity varies widely among and within species and also within individual trees. Notably, contrasted patterns of radial variation of wood specific gravity have been demonstrated and related to regeneration guilds (light demanding vs. shade-bearing). However, although being repeatedly invoked as a potential source of error when estimating the biomass of trees, both intraspecific and radial variations remain little studied. In this study we characterized detailed pith-to-bark wood specific gravity profiles among contrasted species prominently contributing to the biomass of the forest, i.e., the dominant species, and we quantified the consequences of such variations on the biomass.MethodsRadial profiles of wood density at 8% moisture content were compiled for 14 dominant species in the Democratic Republic of Congo, adapting a unique 3D X-ray scanning technique at very high spatial resolution on core samples. Mean wood density estimates were validated by water displacement measurements. Wood density profiles were converted to wood specific gravity and linear mixed models were used to decompose the radial variance. Potential errors in biomass estimation were assessed by comparing the biomass estimated from the wood specific gravity measured from pith-to-bark profiles, from global repositories, and from partial information (outer wood or inner wood).ResultsWood specific gravity profiles from pith-to-bark presented positive, neutral and negative trends. Positive trends mainly characterized light-demanding species, increasing up to 1.8 g.cm-3 per meter for Piptadeniastrum africanum, and negative trends characterized shade-bearing species, decreasing up to 1 g.cm-3 per meter for Strombosia pustulata. The linear mixed model showed the greater part of wood specific gravity variance was explained by species only (45%) followed by a redundant part between species and regeneration guilds (36%). Despite substantial variation in wood specific gravity profiles among species and regeneration guilds, we found that values from the outer wood were strongly correlated to values from the whole profile, without any significant bias. In addition, we found that wood specific gravity from the DRYAD global repository may strongly differ depending on the species (up to 40% for Dialium pachyphyllum).Main ConclusionTherefore, when estimating forest biomass in specific sites, we recommend the systematic collection of outer wood samples on dominant species. This should prevent the main errors in biomass estimations resulting from wood specific gravity and allow for the collection of new information to explore the intraspecific variation of mechanical properties of trees.

Community assessment of tropical tree biomass: challenges and opportunities for REDD.

BackgroundREDD+ programs rely on accurate forest carbon monitoring. Several REDD+ projects have recently shown that local communities can monitor above ground biomass as well as external professionals, but at lower costs. However, the precision and accuracy of carbon monitoring conducted by local communities have rarely been assessed in the tropics. The aim of this study was to investigate different sources of error in tree biomass measurements conducted by community monitors and determine the effect on biomass estimates. Furthermore, we explored the potential of local ecological knowledge to assess wood density and botanical identification of trees.ResultsCommunity monitors were able to measure tree DBH accurately, but some large errors were found in girth measurements of large and odd-shaped trees. Monitors with experience from the logging industry performed better than monitors without previous experience. Indeed, only experienced monitors were able to discriminate trees with low wood densities. Local ecological knowledge did not allow consistent tree identification across monitors.ConclusionFuture REDD+ programmes may benefit from the systematic training of local monitors in tree DBH measurement, with special attention given to large and odd-shaped trees. A better understanding of traditional classification systems and concepts is required for local tree identifications and wood density estimates to become useful in monitoring of biomass and tree diversity.Electronic supplementary materialThe online version of this article (doi:10.1186/s13021-015-0028-3) contains supplementary material, which is available to authorized users.

Biophysical modelling of intra-ring variations in tracheid features and wood density of Pinus pinaster trees exposed to seasonal droughts.

Process-based models that link seasonally varying environmental signals to morphological features within tree rings are essential tools to predict tree growth response and commercially important wood quality traits under future climate scenarios. This study evaluated model portrayal of radial growth and wood anatomy observations within a mature maritime pine (Pinus pinaster (L.) Aït.) stand exposed to seasonal droughts. Intra-annual variations in tracheid anatomy and wood density were identified through image analysis and X-ray densitometry on stem cores covering the growth period 1999-2010. A cambial growth model was integrated with modelled plant water status and sugar availability from the soil-plant-atmosphere transfer model MuSICA to generate estimates of cell number, cell volume, cell mass and wood density on a weekly time step. The model successfully predicted inter-annual variations in cell number, ring width and maximum wood density. The model was also able to predict the occurrence of special anatomical features such as intra-annual density fluctuations (IADFs) in growth rings. Since cell wall thickness remained surprisingly constant within and between growth rings, variations in wood density were primarily the result of variations in lumen diameter, both in the model and anatomical data. In the model, changes in plant water status were identified as the main driver of the IADFs through a direct effect on cell volume. The anatomy data also revealed that a trade-off existed between hydraulic safety and hydraulic efficiency. Although a simplified description of cambial physiology is presented, this integrated modelling approach shows potential value for identifying universal patterns of tree-ring growth and anatomical features over a broad climatic gradient.

Revisiting a universal airborne light detection and ranging approach for tropical forest carbon mapping: scaling-up from tree to stand to landscape

Airborne laser scanning provides continuous coverage mapping of forest canopy height and thereby is a powerful tool to scale-up above-ground biomass (AGB) estimates from stand to landscape. A critical first step is the selection of the plot variables which can be related to light detection and ranging (LiDAR) statistics. A universal approach was previously proposed which combines local and regional estimates of basal area (BA) and wood density with LiDAR-derived canopy height to map carbon at a regional scale (Asner et al. in Oecologia 168:1147-1160, 2012). Here we explore the contribution of stem diameter distribution, specific wood density and height-diameter (H-D) allometry to forest stand AGB and propose an alternative model. By applying the new model to a large tropical forest data set we show that an appropriate choice of input variables is essential to minimize prediction error of stand AGB which will propagate at larger scale. Stem number (N) and average stem cross-sectional area should be used instead of BA when scaling from tree to plot. Stand quadratic mean diameter above the census threshold diameter size should be preferred over stand mean diameter as it reduces the prediction error of stand AGB by a factor of ten. Wood density should be weighted by stem volume per species instead of BA. LiDAR-derived statistics should prove useful for estimating local H-D allometries as well as mapping N and the mean quadratic diameter above 10 cm at the landscape level. Prior stratification into forest types is likely to improve both estimation procedures significantly and is considered the foremost current challenge.

Propagating Uncertainty in Plot-based Estimates of Forest Carbon Stock and Carbon Stock Change

Ecosystem science increasingly relies on highly derived metrics to synthesize across large datasets. However, full uncertainty associated with these metrics is seldom quantified. Our objective was to evaluate measurement error and model uncertainty in plot-based estimates of carbon stock and carbon change. We quantified the measurement error associated with live stems, deadwood and plot-level variables in temperate rainforest in New Zealand. We also quantified model uncertainty for height–diameter allometry, stem volume equations and wood-density estimates. We used Monte Carlo simulation to assess the net effects on carbon stock and carbon change estimated using data from 227 plots from throughout New Zealand. Plot-to-plot variation was the greatest source of uncertainty, amounting to 9.1% of mean aboveground carbon stock estimates (201.11 MgC ha−1). Propagation of the measurement error and model uncertainty resulted in a 1% increase in uncertainty (0.1% of mean stock estimate). Carbon change estimates (mean −0.86 MgC ha−1 y−1) were more uncertain, with sampling error equating to 56% of the mean, and when measurement error and model uncertainty were included this uncertainty increased by 35% (22.1% of the mean change estimate). For carbon change, the largest sources of measurement error were missed/double counted stems and fallen coarse woody debris. Overall, our findings show that national-scale plot-based estimates of carbon stock and carbon change in New Zealand are robust to measurement error and model uncertainty. We recommend that calculations of carbon stock and carbon change incorporate both these sources of uncertainty so that management implications and policy decisions can be assessed with the appropriate level of confidence.

Variations of wood basic density with tree age and social classes in the axial direction within Pinus massoniana stems in Southern China

Reliable estimates of wood density (WD) within individual trees could maximize the value of Pinus massoniana for specific end-use. We examined and quantified the axial patterns of WD in trees with different social status in the stands. Wood disks were sampled at the bottom, breast height, and middle of each 1-m sections from 108 stems, harvesting from three social classes in subtropical forests. A mixed-effects model was designed to quantify axial variation. The WD at different height was significantly different from the whole-stem WD (WWD) except the relative height of 0.1. An overall decrease of 133.8 kg m−3 in WD was found from stem base to top. WD was significantly influenced by relative heights, tree age, and social class. WD of each relative height in mature trees was significantly higher than that of younger trees. Tree social class can affect WD development in the axial direction at age classes 2 and 3. Combining the fixed plus random effects, the final model explained 91 % of the observed variation in WD. The WD development patterns in the axial direction vary considerably among tree age, diameter at breast height, and social class. To distinguish and supply timber for specific end-uses, we should use the axial variation in disk WD (DWD) instead of WWD directly. The accurate predictions of WD provided by the model could be used to optimally classify logs into different product classes and maximize economic benefits. We can use DWD at the relative height of 0.1 instead of WWD of a single tree.

A model‐based meta‐analysis for estimating species‐specific wood density and identifying potential sources of variation

Summary Plant functional traits are important determinants of survival and fitness, and wood density (WD) is a key trait linked to mechanical stability, growth rates and drought‐ and shade‐tolerance strategies. Thus, rigorous WD estimates are necessary to identify factors affecting tree performance. We obtained 1766 records of WD from the literature for 141 tree species in the United States. We implemented a hierarchical Bayesian (HB) meta‐analysis that incorporated sample size, variance, covariate (e.g. moisture content and latewood proportion) and methodological information to obtain standardized estimates of WD for 305 U.S. tree species. The HB framework allowed ‘borrowing of strength’ between species such that WD estimates for data‐poor species were informed by data‐rich species via taxonomic or phylogenetic relationships. After accounting for important covariates and sampling effects, evaluation of the residual variation revealed the potential importance of environmental factors and evolutionary history. Differential variation in WD between species within genera and between genera within orders suggested that WD is relatively conserved in some genera and orders, but not in others. WD also varied between studies (or sites) indicating the potential influence of edaphic, topographic, or population factors on intraspecific variation in WD. Synthesis. Our hierarchical Bayesian approach overcomes many of the limitations of traditional meta‐analyses, and the incorporation of phylogenetic or taxonomic information facilitates estimates of trait values for data‐poor species. We provide relatively well‐constrained WD estimates for 305 tree species, which may be useful for tree growth and forest models, and the uncertainties associated with the estimates may inform future sampling campaigns.

Predicting Douglas-fir wood density by artificial neural networks (ANN) based on progeny testing information

Abstract A heuristic wood density prediction model has been developed by means of artificial neural networks (ANNs). Four populations of 32-year-old coastal Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco var. menziesi) trees representing 20 full-sib families growing on comparable sites were in focus of this study. Tree height, diameter, volume, wood density, and acoustic velocity data from 632 trees were considered for the calculations. Two different ANN platforms were developed employing different classes and architectures, namely, the multilayer feed-forward (MLFF) and modular (MOD) models. After establishing the optimal configuration of the model, a MLFF network and a MOD neural network (with the obtained optimal structure) were developed and tested without cross-validation by employing a typical training and testing set methodology. To this purpose, the data set was divided in 480 trees for training and 152 trees for validation. A significant relationship between actual and predicted wood density was obtained with R2 values of 0.50 and 0.52 for the two networks, respectively, demonstrating their predictive potential for wood density estimation. A classic multiple regression analysis produced substantially lower predictive power with an R2 of 0.23. The application of ANNs as a viable predictive tool in determining wood density using growth and acoustic velocity data without additional intrusive sampling and laboratory work was demonstrated. An additional work including other species is required for these approaches.

A multidimensional statistical model for wood data analysis, with density estimated from CT scanning data as an example

The trunk of a tree can be seen as a spatiotemporal sampling domain from the statistical perspective, where space is represented by direction horizontally and height vertically, and time through annual growth rings. In this framework, wood properties such as density can be the object of data collection for given estimation and testing purposes. We present a multidimensional statistical model, the tensor normal distribution, in which the variation (variance) of and dependency (covariance) between wood property measurements made for different years at various locations in a tree trunk can be inferred. Its application requires a smaller number of replicates (trees) than the traditional vector normal distribution because variances and covariances for directions and growth rings, for example, must be the same at all heights, up to a multiplicative constant. This assumption on the variance–covariance structure is called “separability”, and we explain how to test it. An illustration with wood density estimates obtained from computed tomography scanning data for 11 white spruce ( Picea glauca (Moench) Voss) trees is presented. This example is completed by assessing differences in mean wood density according to location in the trunk, with analysis-of-variance F-tests adjusted for the estimated variances and covariances obtained by fitting the model.

A universal airborne LiDAR approach for tropical forest carbon mapping

Airborne light detection and ranging (LiDAR) is fast turning the corner from demonstration technology to a key tool for assessing carbon stocks in tropical forests. With its ability to penetrate tropical forest canopies and detect three-dimensional forest structure, LiDAR may prove to be a major component of international strategies to measure and account for carbon emissions from and uptake by tropical forests. To date, however, basic ecological information such as height-diameter allometry and stand-level wood density have not been mechanistically incorporated into methods for mapping forest carbon at regional and global scales. A better incorporation of these structural patterns in forests may reduce the considerable time needed to calibrate airborne data with ground-based forest inventory plots, which presently necessitate exhaustive measurements of tree diameters and heights, as well as tree identifications for wood density estimation. Here, we develop a new approach that can facilitate rapid LiDAR calibration with minimal field data. Throughout four tropical regions (Panama, Peru, Madagascar, and Hawaii), we were able to predict aboveground carbon density estimated in field inventory plots using a single universal LiDAR model (r ( 2 )=0.80, RMSE=27.6MgCha(-1)). This model is comparable in predictive power to locally calibrated models, but relies on limited inputs of basal area and wood density information for a given region, rather than on traditional plot inventories. With this approach, we propose to radically decrease the time required to calibrate airborne LiDAR data and thus increase the output of high-resolution carbon maps, supporting tropical forest conservation and climate mitigation policy.

A non-destructive field method for measuring wood density of decaying logs

Decaying dead wood density measurements are a useful indicator for multiple purposes, such as for estimating the amount of carbon in dead wood and making predictions of potential diversity of dead wood inhabiting fungi and insects. Currently, qualitative decay phases are used as wood density estimates in many applications, since measuring the density is laborious. A quantitative measure of density would, however, be preferred over the qualitative one. Penetrometers, which are commonly used for measuring the density of standing trees, might also be applicable to dead wood density measurements. We tested the device for making quick, quantitative measurements of decaying logs. The penetrometer measures the depth into which a pre-loaded spring forces a pin in the wood. We tested pins of 5 and 10 mm diameter together with an original 2.5 mm pin and compared the results with gravimetric density measurements of the sample logs. Our results suggest that the standard pin works for less decayed wood, but for more decomposed wood, the thicker 5 mm pin gave more reliable estimates when the penetration measures were converted to densities with a linear regression function (R = 0.62, F = 82.9, p = 0.000). The range of wood densities successfully measured with the 5 mm pin was from 180 to 510 kg m. With the 10 mm pin, the measuring resolution of denser wood was compromised, while the improvement at the other end of density scale was not large. As a conclusion, the penetrometer seems to be a promising tool for quick density testing of decaying logs in field, but it needs to be modified to use a thicker measuring pin than the standard 2.5 mm pin.2–3

High-resolution proxies for wood density variations in Terminalia superba

Density is a crucial variable in forest and wood science and is evaluated by a multitude of methods. Direct gravimetric methods are mostly destructive and time-consuming. Therefore, faster and semi- to non-destructive indirect methods have been developed. Profiles of wood density variations with a resolution of approx. 50 µm were derived from one-dimensional resistance drillings, two-dimensional neutron scans, and three-dimensional neutron and X-ray scans. All methods were applied on Terminalia superba Engl. & Diels, an African pioneer species which sometimes exhibits a brown heart (limba noir). The use of X-ray tomography combined with a reference material permitted direct estimates of wood density. These X-ray-derived densities overestimated gravimetrically determined densities non-significantly and showed high correlation (linear regression, R(2) = 0·995). When comparing X-ray densities with the attenuation coefficients of neutron scans and the amplitude of drilling resistance, a significant linear relation was found with the neutron attenuation coefficient (R(2) = 0·986) yet a weak relation with drilling resistance (R(2) = 0·243). When density patterns are compared, all three methods are capable of revealing the same trends. Differences are mainly due to the orientation of tree rings and the different characteristics of the indirect methods. High-resolution X-ray computed tomography is a promising technique for research on wood cores and will be explored further on other temperate and tropical species. Further study on limba noir is necessary to reveal the causes of density variations and to determine how resistance drillings can be further refined.

Estimating the wood density of species for carbon stock assessments

Summary1. Studies of forest carbon stocks and fluxes rely on estimates of specific wood density to convert tree diameter and height measurements made within permanent plots into carbon stock estimates. However, measurements of wood density are often available for only a subset of species. As there is strong phylogenetic trait conservatism of wood density, missing data are usually estimated by averaging the wood densities of other species within the same genus or family, using whatever data are available locally.2. The Global Wood Density (GWD) database (http://hdl.handle.net/10255/dryad.235) provides wood densities for 8412 species from around the world, providing an opportunity to utilize data from further afield when faced with missing values in a study area. We investigated whether the GWD provides better estimates than local data sets when conventional averaging methods (AM) are used. Secondly, we develop Hierarchical Bayesian Models (HBM) that incorporate phylogenetic covariance to estimate missing wood densities.3. Using AM, we found that correlations between observations and estimates were higher when the GWD was used in place of local data sets, mostly because of larger sample size. Missing wood densities were more accurately estimated from the global data set than from local data sets, indicating that the GWD should be used as a common standard when calculating carbon stocks.4. Estimates based on including phylogenetic dependency in HBMs were also closely correlated to observations, but were no better than those obtained from the simpler AM. Estimations based on HBM could become more useful when phylogenetic trees resolved to the species level are available. Until such improvements are made, we conclude that building more data into the reference data set, rather than improving the method itself, is the most productive way to refine estimates of unknown wood densities.

Decay dynamic of coarse and fine woody debris of a beech (Fagus sylvatica L.) forest in Central Germany

Decay rates of woody debris were estimated and used to model the decay of various diameters of branches and stems in a beech stand in Central Germany. In addition, use of wood density, volume and mass loss to quantitatively describe the degree of decay was tested. The mass loss during decay could be described by a simple exponential function. Under the presented climatic conditions, beech coarse woody debris (CWD) with a diameter >10 cm decays completely in about 35 years. In the first 8 years of decay the mass loss is determined by the decrease in wood density, and subsequently by the loss in volume. Estimation of wood density allows the first three of the four classes of decay to be distinguished, while trees in the last two decay classes could be distinguished using wood volume. Beech fine woody debris with a diameter between 1 and 10 cm decays within about 18 years. The litter fraction of <1 cm is part of the humus layer after 4 years. If there are goals for the amounts, types and dimensions of woody debris to be provided for conservation of biological biodiversity and other ecological functions in managed beech forests, this study offer indications for how long existing woody debris can meet its functions and how frequent new input of CWD is required.

Measuring wood density by means of X-ray computer tomography

• Wood density is a characteristic of major interest. Usually, it is used as an indicator of wood quality; however, in the context of global change, it is increasingly used for biomass and carbon storage estimations. X-ray computer tomography is a method which enables quick estimates of wood density after applying a calibration procedure. • A review of the literature is presented in this article. Most of the previous studies have been performed in the 80’s or at the beginning of the 90’s. • In this study, the relationship between wood density and Hounsfield numbers was investigated using a recent medical scanner. A linear relationship was fitted using a calibration data set which consisted in tropical wood samples representing a large range of densities ranging between 133 and 1319 kg m−3, and then validated using an independent data set (mainly temperate tree species). The fitted relationships were very strong (R2 > 0.999), whichever the tested scanner settings, with slight but significant effects of the current voltage and reconstruction filters. The RMSE values computed from the validation data set ranged between 5.4 and 7.7 kg m−3 for densities ranging between 364 and 821 kg m−3. • In conclusion, this method of calibration enables the use of a medical scanner to obtain maps of wood density, in a fast and non destructive way, and with a very good accuracy. Very interesting perspectives are opened regarding biomass distribution within trees.

Predicting timber properties from tree measurements at felling: Evaluation of the RetroSTEM model and TreeViz software for Norway spruce

The RetroSTEM model was developed for predicting three-dimensional stem and wood quality characteristics in conifer stems from simple measurements at felling. The model has been previously parameterised and tested for Scots pine. This study adapts the model to Norway spruce and evaluates its prediction power regarding stem and wood quality simulations from tree base to tree top, with the objective of identifying any systematic errors in the estimation. This information will be used for future model development. The selected quality factors – stem taper, branch size, ring width and wood density distributions – are critical for visual and strength-based timber grading and for timber yield estimations. The wood density simulations provided through visual tool, TreeViz, were embedded in the RetroSTEM. The model testing indicated that the height growth model embedded in RetroSTEM should be revised regarding growth at early ages. The predicted early height growth was too fast, which led to bias in the number of annual rings at the middle and lower parts of the stem and in stem diameter at the middle stem. However, the average bias in stem diameter was relatively small (−0.3 cm) at tree level. Similarly the average bias of branch maximum diameters was low (0.2 cm) at tree level, but RetroSTEM underestimated branch diameters in trees with long crowns. Overestimation in the number of annual rings also led to underestimation of average ring width. The TreeViz estimates of wood density seem realistic. At this stage the model is applicable to estimating stem quality distributions at stand level, particularly in dense stands younger than 100 years of age.

Near Infrared Calibration Models for the Estimation of Wood Density in Pinus Taeda Using Repeated Sample Measurements

Near infrared (NIR) diffuse reflectance was used for the estimation of air-dry density and basic density in wood radial strip samples obtained at breast height (1.4 m) from 60 Pinus taeda trees established in three progeny tests in the south-eastern United States. NIR calibration models were fitted using raw spectra and pre-processed spectra with second derivative, multiplicative scatter correction and orthogonal signal correction. Successful calibrations were obtained for both wood properties using data collected in consecutive 10 mm sections from the samples. Data pre-processing did not result in model improvements compared to the models fitted using raw data. The effects of using repeated measures were evaluated by incorporating serial correlation into the partial least squares regression algorithm. The empirical autocorrelation of the normalised residuals showed that serial dependence among residuals was successfully removed by using an autoregressive correlation structure of second order. However, because the initial dependence among observations was not strong, the predictions were similar using the modified algorithm to those obtained with the traditional approach. These results indicate that the use of repeated measurements does not represent a serious problem for the development of NIR calibration models for the prediction of wood properties using radial samples measured in 10 mm sections and that the specification of the correlation structure may not be required when the models are used only for predictive purposes.

Wood density in forests of Brazil's ‘arc of deforestation’: Implications for biomass and flux of carbon from land-use change in Amazonia

Wood density is an important variable in estimates of forest biomass and greenhouse-gas emissions from land-use change. The mean wood density used in estimates of forest biomass in the Brazilian Amazon has heretofore been based on samples from outside the “arc of deforestation”, where most of the carbon flux from land-use change takes place. This paper presents new wood density estimates for the southern and southwest Brazilian Amazon (SSWA) portions of the arc of deforestation, using locally collected species weighted by their volume in large local inventories. Mean wood density was computed for the entire bole, including the bark, and taking into account radial and longitudinal variation. A total of 403 trees were sampled at 6 sites. In the southern Brazilian Amazon (SBA), 225 trees (119 species or morpho-species) were sampled at 4 sites. In eastern Acre state 178 trees (128 species or morpho-species) were sampled at breast height in 2 forest types. Mean basic density in the SBA sites was 0.593 ± 0.113 (mean ± 1 S.D.; n = 225; range 0.265–0.825). For the trees sampled in Acre the mean wood density at breast height was 0.540 ± 0.149 ( n = 87) in open bamboo-dominated forest and 0.619 ± 0.149 ( n = 91) in dense bamboo-free forest. Mean wood density in the SBA sites was significantly higher than in the bamboo dominated forest but not the dense forest at the Acre site. From commercial wood inventories by the RadamBrasil Project in the SSWA portion of the arc of deforestation, the wood volume and wood density of each species or genus were used to estimate average wood density of all wood volume in each vegetation unit. These units were defined by the intersection of mapped forest types and states. The area of each unit was then used to compute a mean wood density of 0.583 g cm −3 for all wood volume in the SSWA. This is 13.6% lower than the value applied to this region in previous estimates of mean wood density. When combined with the new estimates for the SSWA, this gave an average wood density of 0.642 g cm −3 for all the wood volume in the entire Brazilian Amazon, which is 7% less than a prior estimate of 0.69 g cm −3. These results suggest that current estimates of carbon emissions from land-use change in the Brazilian Amazon are too high. The impact on biomass estimates and carbon emissions is substantial because the downward adjustment is greater in forest types undergoing the most deforestation. For 1990, with 13.8 × 10 3 km 2 of deforestation, emissions for the Brazilian Amazon would be reduced by 23.4–24.4 × 10 6 Mg CO 2-equivalent C/year (for high- and low-trace gas scenarios), or 9.4–9.5% of the gross emission and 10.7% of the net committed emission, both excluding soils.