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.     

Aboveground biomass quantification and tree-level prediction models for the Brazilian subtropical Atlantic Forest

Forests are the largest biological reservoir of biomass and carbon on the planet. This fact turns them into the main tool to neutralise the CO₂ emitted by human activities. Despite such importance, the uncertainties associated with biomass estimates in forests, especially in (sub)tropical forests, are enormous. Facing this scenario, the objectives of this study were (1) to quantify through destructive sampling the aboveground biomass (AGB) of 105 trees of 47 species occurring in a secondary subtropical evergreen rainforest in Brazil; (2) to investigate the AGB distribution in different tree compartments; and (3) to fit tree-level models to improve biomass estimates for the referred forest type. The results revealed that most of the AGB was stored in the compartments stem and large branches (diameter ≥ 5 cm). There was an increase in the proportion of biomass – in relation to the total tree AGB (kg) – allocated in the large branches as tree diameter at breast height (DBH) increased; this pattern was not observed for the compartments stem, thin branches (diameter < 5 cm), and leaves. The compartments thin branches and leaves represented between 5.4% and 17.0%, and 1.3% and 2.9% of the total tree AGB, respectively. From the 10 fitted biomass models, the linearised power models yielded the smallest errors. The best performance model, which returned a mean bias of 1.7%, may be written as AGB = exp(?8.9807 + 2.1642·ln[DBH] + 0.5072·ln[h] + 0.9999·ln[ρ_bas]); Baskerville’s factor = 1.0175. If there are no (reliable) data on tree total height (h; m), the following model, which embedded the DBH and wood basic specific gravity (ρ_bas; kg m^?3), may be employed: AGB = exp(?9.0086 + 2.4606·ln[DBH] + 1.0895·ln[ρ_bas]); Baskerville’s factor = 1.0206.     

Developing an airborne laser scanning dominant height model from a countrywide scanning survey and national forest inventory data

Abstract

An airborne laser scanning (ALS) dominant height model was developed based on data from a national scanning survey with the aim of developing a digital terrain model (DTM) for Denmark. Data obtained in the ongoing Danish national forest inventory (NFI) were used as reference data. The data comprised a total of 2072 measurements of dominant height on NFI sample plots inventoried in 2006–2007 and their corresponding ALS data. The dominant height model included four variables derived from the ALS point cloud distribution. The variables were related to canopy height, canopy density and species composition on individual plots. The RMSE of the final model was 2.25 m and the model explained 93.9% of the variation (R ²). The model was successful in predicting dominant height across a wide range of forest tree species, stand heights, stand densities, canopy cover and growing conditions. The study demonstrated how low-density ALS data obtained in a survey not specifically aimed at forest applications may be used for obtaining biophysical forest properties such as dominant height, thereby reducing the overall forest inventory costs.     

Effect of tree-level airborne laser-scanning measurement accuracy on the timing and expected value of harvest decisions

The objective was to compare tree-level airborne laser-scanning (ALS) data accuracy with standwise estimation data accuracy as input data for forest planning, using tree- and stand-level simulators. The influence of the input data accuracy was studied with respect to (1) timing of the next thinning or clear-cutting and (2) the relative variation in the predicted income of the next logging expressed as the net present value (NPV). The timing and predicted NPV of thinning and clear-cutting operations were considered separately. The research was based on Monte Carlo simulations carried out with the tree- and stand-level simulators using a simulation and optimisation (SIMO) framework. The simulations used treewise measurements taken on 270 circular plots measured at the Evo Field Station, Finland, as input data. Deviations in the tree data measured were generated according to the mean standard errors found in standwise field estimation and tree-level ALS. The accuracy factors of ALS individual tree detection were based on the EUROSDR/ISPRS Tree Extraction Project. The results show that input data accuracy significantly affects both the timing and relative NPV of loggings. Tree-level ALS produces more accurate simulation results than standwise estimation with the error levels assumed. Diameter-based characteristics are the most important input data in all simulations. Accurate tree height estimates cannot be fully utilised in current simulators.     

Trends in allometric models and aboveground biomass of family Rhizophoraceae mangroves in the Coral Triangle ecoregion,Southeast Sulawesi,Indonesia

This study aims to establish allometric models and estimate aboveground biomass (AGB) of mangroves Rhizophoraceae in the Southeast Sulawesi, Indonesia. Allometric models of the AGB of mangroves Rhizophora apiculata, R. mucronata, and Ceriops tagal were established using independent variables consisting stem diameter at 30 cm from the ground (D₃₀), diameter at breast height (DBH), D₃₀²H and DBH²H. The AGB of mangroves was estimated by applying allometric model and tree census. The results showed that the best fitting allometric models of AGB for R. apiculata is based on variable DBH, while DBH²H is the best variable for R. mucronata trees. Conversely, the D₃₀ is the best variable for estimating AGB of C. tagal trees. Thus, there is some variation of independent variables on allometric models for the estimation of AGB for Rhizophoraceae mangroves. The AGB (ton ha^?1) of R. apiculata, R. mucronata, and C. tagal was estimated respective 651.60, 232.11 and 154.56 in the protected area, and respective 137.59, 189.35 and 39.06 ton ha^?1 in the unprotected area. Higher AGB of mangroves growing in the protected area indicated the suitable condition and undisturbed by human activities. The conservation of mangroves is necessary for the sustainability of mangroves and coastal ecosystems in the Coral Triangle ecoregion.     

河北省主要树种单木和林分生长率模型研建

利用第六次至第九次全国森林资源清查河北省2001,2006,2011,2016年4个年度的固定样地调查数据,采用非线性回归估计方法,建立了18个树种组的单木胸径生长率和材积生长率模型,以及12个树种组的林分材积生长率模型。结果表明,单木生长率模型的平均预估误差(MPE)基本都在3%以内,而平均百分标准误差(MPSE)、胸径生长率模型大都在10%以内,材积生长率模型大都在20%左右;林分生长率模型的平均预估误差(MPE)基本都在5%以内,平均百分标准误差(MPSE)大都在25%以内。所建模型可为河北省开展森林资源年度更新提供技术支撑。     

Models for Predicting Dry Matter Content of Norway spruce

Models were developed for predicting the dry matter content of single Norway spruce [Picea abies (L.) Karst.] stems from non-destructive field measurement data such as tree height, diameter and age. The material under study comprised 1612 stem discs originating from 235 trees grown at nine locations in Denmark and southern Sweden. In total, 153 trees (1054 discs) originating from a thinning experiment in southern Denmark were used for establishing the models, which were later validated on independent material. Results of the validation were that the 95% confidence interval for single stem density was ±9 to ±16% when based on average ring width at breast height, but within ±7% when based on wood samples taken at breast height. The density models were combined with standard volume models in order to predict dry matter content. Validation on independent material showed that in general the dry matter content of single stems can be estimated with less than 10% error (RMSE).     

Total tree,merchantable stem and branch volume models for miombo woodlands of Malawi

The objective of this study was to develop general (multispecies) models for prediction of total tree, merchantable stem and branch volume including options with diameter at breast height (dbh) only, and with both dbh and total tree height (ht), as independent variables. The modelling data set was based on destructively sampled trees and comprised 74 trees from 33 tree species, collected from four forest reserves located in different ecological zones of Malawi. The dbh and ht ranges for the data set were 5.3–111.2?cm and 3.0–25.0?m, respectively. A number of alternative model forms were tested and the final model selection was based on root mean square error (RMSE) values calculated using a leave-one-out cross-validation procedure. The model performances and the evaluations of the finally selected models (R? 2 range 0.72 to 0.92; RMSE range 38% to 71%; mean prediction errors range ?1.4% to 1.3%) suggest that all models can be used over a wide range of geographical and ecological conditions in Malawi with an appropriate accuracy in predictions. The appropriateness of the developed models was also supported by the fact that the mean prediction errors of these models were much lower than the mean prediction errors (range ?23.6% to 48.9%) of some previously developed models tested on our data.     

Model performance of tree height-diameter relationships in the central Congo Basin

Key message

Tree heights in the central Congo Basin are overestimated using best-available height-diameter models. These errors are propagated into the estimation of aboveground biomass and canopy height, causing significant bias when used for calibration of remote sensing products in this region.

Context

Tree height-diameter models are important components of estimating aboveground biomass (AGB) and calibrating remote sensing products in tropical forests.

Aims

For a data-poor area of the central Congo Basin, we quantified height-diameter model performance of local, regional and pan-tropical models for their use in estimating AGB and canopy height.

Methods

At three old-growth forest sites, we assessed the bias introduced in height estimation by regional and pan-tropical height-diameter models. We developed an optimal local model with site-level randomizations accounted for by using a mixed-effects modeling approach. We quantified the error propagation of modeled heights for estimating AGB and canopy height.

Results

Regional and pan-tropical height-diameter models produced a significant overestimation in tree height, propagating into significant overestimations of AGB and Lorey’s height. The pan-tropical model accounting for climatic drivers performed better than the regional models. We present a local height-diameter model which produced nonsignificant errors for AGB and canopy height estimations at our study area.

Conclusion

The application of general models at our study area introduced bias in tree height estimations and the derived stand-level variables. Improved delimitation of regions in tropical Africa with similar forest structure is needed to produce models fit for calibrating remote sensing products.

     

Estimation of biomass change in montane forests in Norway along a 1200 km latitudinal gradient using airborne laser scanning: a comparison of direct and indirect prediction of change under a model-based inferential approach

The current study compared two approaches for estimating change of aboveground biomass (AGB) in montane forests in Norway using field- and remotely sensed data from airborne laser scanning (ALS) from two points in time (four-year interval). The first was an indirect method that involved modeling and prediction of AGB at two points in time using ALS metrics as predictors, estimating the change from differences between AGB predictions. The second was a direct method, where change was modeled and predicted directly using differences between corresponding ALS metrics derived at the two measurement occasions as predictors, and the estimate was based on the predicted differences. Both methods were applied over a 1500?km long and 250?m wide transect from south to north in Norway comprising 250?m² grid cells. The results showed that the indirect method was more precise than the direct method. The indirect method estimated 0.65?Mg?ha^?1 change in AGB over the observation period, with a corresponding 95% confidence interval of ±0.27?Mg?ha^?1. The corresponding figures for the direct method was 0.54 and ±0.51?Mg?ha^?1. The direct method has been recommended previously. We conclude that the indirect method is both more precise and versatile.     

Allometric models for estimating aboveground biomass of shade trees and coffee bushes grown together

Allometric models for dominant shade tree species and coffee plants (Coffea arabica) were developed for coffee agroforestry systems in Matagalpa, Nicaragua. The studied shade tree species were Cordia alliodora, Juglans olanchana, Inga tonduzzi and I. punctata. The models predict aboveground biomass based on diameter at breast height (for trees), and the stem diameter at a height of 15 cm and plant height (for coffee plants). In addition, the specific gravity of the studied species was determined.The total aboveground biomass of the shade trees varied between 3.5 and 386 kg per tree, and between 0.005 and 2.8 kg per plant for coffee. The aboveground biomass components (foliage, branch, and stem) are closely related with diameter at breast height (r > 0.75). The best-fit models for aboveground biomass of the shade trees were logarithmic, with adjusted R ² between 0.71 and 0.97. In coffee plants, a high correlation was found (r = 0.84) with the stem diameter at 15 cm height, and the best-fit model was logarithmic, as well. The mean specific gravity was 0.52 (± 0.11) for trees and 0.82 (± 0.06) for coffee plants.     

Assessing 3D point clouds from aerial photographs for species-specific forest inventories

In this study we assessed the potential of using photogrammetric data for species-specific forest inventories. The method is based on a combination of Dirichlet and ordinary linear regression models. This approach was used to predict species proportions, main tree species, total, and species-specific volume. Structural and spectral variables were used as predictors. The models were validated using 63 independent validation stands. The results from airborne laser scanning (ALS) data combined with spectral data and photogrammetric data obtained using aerial imagery with different forward overlaps of 80% and 60% were compared. The best photogrammetry-based models predicted species proportions with a relative root mean square error (RMSE) of 21.4%, classified dominant species with 79% accuracy, predicted total volume with relative RMSE of 13.4%, and predicted species-specific volume with relative RMSE of 36.6%, 46.5%, and 84.9% for spruce, pine, and deciduous species, respectively. The results were similar for the three point cloud datasets obtained from aerial imagery and ALS and the accuracies of the predictions were comparable to methods used in operational FMI. The study highlights the effectiveness of forest inventories carried out using photogrammetric data, which – differently from ALS, can include species-specific information without relying on multiple data sources.     

南岭小坑木荷群落地上生物量

森林生物量是评价森林生态系统生产力、研究森林生态系统结构与功能的重要指标,也是深入了解森林生态系统变化规律的重要途径和评估森林碳收支的重要参数(吴仲民等,1998)。亚热带常绿阔叶林是我国面积最大的森林类型,它在世界森林植     

Aboveground biomass increment and stand dynamics in tropical evergreen broadleaved forest

Forest ecosystems can modify the atmospheric CO₂ through biomass accumulation mostly in tree stems with diameter at breast height (DBH) ≥ 10 cm. Aboveground biomass increment (ΔAGB), and changes in stand AGB, no. stems and basal area (BA) were calculated from mortality, recruitment, and growth data of tree stems in tropical evergreen broadleaved forest, Central Highland Vietnam. Data were derived from ten 1-ha permanent plots established in 2004, where all stems with DBH ≥ 10 cm were tagged, identified to species, and measured for DBH in 2004 and 2012. In an 8-year duration, the increment was 53 ± 10 stems ha^–1, 7.8 ± 0.3 m² ha^–1 for BA and 86.0 ± 4.6 Mg ha^–1 for AGB. The stem mortality rate was 0.9% year^–1 and the stem recruitment rate was 2.2% year^–1. Annual ΔAGB was 10.8 Mg ha^–1 year^–1, equaling to 5.4 Mg C ha^–1 year^–1. Of which, tree stems of 35–80 cm DBH classes accounted for 65%. The results indicated that the forest is in stage of carbon sequestration. Any disturbances causing death of 35–80 cm DBH tree stems will much reduce carbon sequestration capacity and it will take a long time for AGB to return to pre-disturbance stage.     

Comparing biophysical forest characteristics estimated from photogrammetric matching of aerial images and airborne laser scanning data

Recent development in aerial digital cameras and software facilitate the photogrammetric point cloud as a new data source in forest management planning. A total of 151 field training plots were distributed systematically within three predefined strata in a 852.6 ha study area located in the boreal forest in southeastern Norway. Stratum-specific regression models were fitted for six studied biophysical forest characteristics. The explanatory variables were various canopy height and canopy density metrics derived by means of photogrammetric matching of aerial images and small-footprint laser scanning. The ground sampling distance was 17 cm for the images and the airborne laser scanning (ALS) pulse density was 7.4 points m^–2. Resampled images were assessed to mimic acquisitions at higher flying altitudes. The digital terrain model derived from the ALS data was used to represent the ground surface. The results were evaluated using 63 independent test stands. When estimating height in young forest and mature forest on poor sites, the root mean square error (RMSE) values were slightly better using data from image matching compared to ALS. However, for all other combinations of biophysical forest characteristics and strata, better results were obtained using ALS data. In general, the best results were found using the highest image resolution.     

Modeling height–diameter relationship for <Emphasis Type="Italic">Populus euphratica</Emphasis> in the Tarim riparian forest ecosystem,Northwest China

Modeling height–diameter relationships is an important component in estimating and predicting forest development under different forest management scenarios. In this paper, ten widely used candidate height–diameter models were fitted to tree height and diameter at breast height(DBH)data for Populus euphratica Oliv. within a 100 ha permanent plots at Arghan Village in the lower reaches of the Tarim River, Xinjiang Uyghur Autonomous Region of China. Data from 4781 trees were used and split randomly into two sets:75 % of the data were used to estimate model parameters(model calibration), and the remaining data(25 %) were reserved for model validation. All model performances were evaluated and compared by means of multiple model performance criteria such as asymptotic t-statistics of model parameters, standardized residuals against predicted height,root mean square error(RMSE), Akaike’s informationcriterion(AIC), mean prediction error(ME) and mean absolute error(MAE). The estimated parameter a for model(6) was not statistically significant at a level of a = 0.05. RMSE and AIC test result for all models showed that exponential models(1),(2),(3) and(4) performed significantly better than others. All ten models had very small MEs and MAEs. Nearly all models underestimated tree heights except for model(6). Comparing the MEs and MAEs of models, model(1) produced smaller MEs(0.0059) and MAEs(1.3754) than other models. To assess the predictive performance of models, we also calculated MEs by dividing the model validation data set into 10-cm DBH classes. This suggested that all models were likely to create higher mean prediction errors for tree DBH classes[20 cm. However, no clear trend was found among models.Model(6) generated significantly smaller mean prediction errors across all tree DBH classes. Considering all the aforementioned criteria, model(1): TH ? 1:3 t a= e1 t b?eàc?DBHT and model(6): TH ? 1:3 t DBH2= ea t b?DBH t c ? DBH2T are recommended as suitable models for describing the height–diameter relationship of P. euphratica. The limitations of other models showing poor performance in predicting tree height are discussed. We provide explanations for these shortcomings.     

Generalized height–diameter models—an application for major tree species in complex stands of interior British Columbia

Using permanent sample-plot data, selected tree height and diameter functions were evaluated for their predictive abilities for major tree species in complex (multiple age, size and species cohort) stands of interior British Columbia (BC), Canada. Two sets of models were evaluated. The first set included five models for estimating height as a function of individual tree diameter, the second set also included five models for estimating height as a function of individual tree diameter and other stand-level attributes. The inclusion of the BAL index (which simultaneously indicates the relative position of a tree and stand density) into the base height–diameter models increased the accuracy of prediction for all species. On average, by including stand level attributes, root mean square values were reduced by 30.0 cm. Based on the residual plots and fit statistics, these models can be recommended for estimating tree heights for major tree species in complex stands of interior BC. The model coefficients are documented for future use.     

Development of equations for predicting Puerto Rican subtropical dry forest biomass and volume

Carbon accounting, forest health monitoring and sustainable management of the subtropical dry forests of Puerto Rico and other Caribbean Islands require an accurate assessment of forest aboveground biomass (AGB) and stem volume. One means of improving assessment accuracy is the development of predictive equations derived from locally collected data. Forest inventory and analysis (FIA) measured tree diameter and height, and then destructively sampled 30 trees from 6 species at an upland deciduous dry forest site near Ponce, Puerto Rico. This data was used to develop best parsimonious equations fit with ordinary least squares procedures and additive models fit with nonlinear seemingly unrelated regressions that estimate subtropical dry forest leaf, woody, and total AGB for Bucida buceras and mixed dry forest species. We also fit equations for estimating inside and outside bark total and merchantable stem volume using both diameter at breast height (d.b.h.) and total height, and diameter at breast height alone for B. buceras and Bursera simaruba. Model fits for total and woody biomass were generally good, while leaf biomass showed more variation, possibly due to seasonal leaf loss at the time of sampling. While the distribution of total AGB into components appeared to remain relatively constant across diameter classes, AGB variability increased and B. simaruba and B. buceras allocated more carbon into branch biomass than the other species. When comparing our observed and predicted values to other published dry forest AGB equations, the equation developed in Mexico and recommended for areas with rainfall >900 mm/year gave estimates substantially lower than our observed values, while equations developed using dry forest data from forest in Australia, India and Mexico were lower than our observed values for trees with d.b.h. <25 cm and slightly higher for trees with d.b.h. >30 cm. Although our ability to accurately estimate merchantable stem volume and live tree AGB for subtropical dry forests in Puerto Rico and other Caribbean islands has been improved, much work remains to be done to sample a wider range of species and tree sizes.     

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.     

Effect of species grouping and site variables on aboveground biomass models for lowland tropical forests of the Indo-Malay region

Key message

This study assessed the effect of ecological variables on tree allometry and provides more accurate aboveground biomass (AGB) models through the involvement of large samples representing major islands, biogeographical zones and various succession and degradation levels of natural lowland forests in the Indo-Malay region. The only additional variable that significantly and largely contributed to explaining AGB variation is grouping based on wood-density classes.

Context

There is a need for an AGB equation at tree level for the lowland tropical forests of the Indo-Malay region. In this respect, the influence of geographical, climatic and ecological gradients needs to be assessed.

Aims

The overall aim of this research is to provide a regional-scale analysis of allometric models for tree AGB of lowland tropical forests in the Indo-Malay region.

Methods

A dataset of 1300 harvested trees (5 cm ≤ trunk diameter ≤ 172 cm) was collected from a wide range of succession and degradation levels of natural lowland forests through direct measurement and an intensive literature search of principally grey publications. We performed ANCOVA to assess possible irregular datasets from the 43 study sites. After ANCOVA, a 1201-tree dataset was selected for the development of allometric equations. We tested whether the variables related to climate, geographical region and species grouping affected tree allometry in the lowland forest of the Indo-Malay region.

Results

Climatic and major taxon-based variables were not significant in explaining AGB variations. Biogeographical zone was a significant variable explaining AGB variation, but it made only a minor contribution on the accuracy of AGB models. The biogeographical effect on AGB variation is more indirect than its effect on species and stand characteristics. In contrast, the integration of wood-density classes improved the models significantly.

Conclusion

Our AGB models outperformed existing local models and will be useful for improving the accuracy on the estimation of greenhouse gas emissions from deforestation and forest degradation in tropical forests. However, more samples of large trees are required to improve our understanding of biomass distribution across various forest types and along geographical and elevation gradients.