Search

Photogrammetric point clouds obtained with unmanned aircraft systems (UAS) have emerged as an alternative source of remotely sensed data for small area forest management inventories (FMI). Nonetheless, it is often overlooked that small area FMI require considerable field data in addition to UAS data, to support the modelling of forest attributes. In this study, we propose a method whereby tree volumes by species are predicted with photogrammetric UAS data and Sentinel-2 images, using models fitted with airborne laser scanning data. The study area is in a managed boreal forest area in Eastern Finland. First, we predicted total volume with UAS point cloud metrics using a prior regression model fitted in another area with ALS data. Tree species proportions were then predicted by nearest neighbor (-NN) imputation based on bi-seasonal Sentinel-2 images without measuring new field plot data. Species-specific volumes were then obtained by multiplying the total volume by species proportions. The relative root mean square error (RMSE) values for total and species-specific volume predictions at the validation plot level (30 m à 30 m) were 9.0%, and 33.4â62.6%, respectively. Our approach appears promising for species-specific small area FMI in Finland and in comparable forest conditions in which suitable field plots are available.kk

The aim of this study was to examine how well stem volume, above-ground biomass and dominant height can be predicted using nationwide airborne laser scanning (ALS) based regression models. The study material consisted of nine practical ALS inventory projects taken from different parts of Finland. We used field sample plots and airborne laser scanning data to create nationwide and regional models for each response variable. The final models had one or two ALS predictors, which were chosen based on the root mean square error (RMSE), and cross-validated. Finally, we tested how much predictions would improve if the nationwide models were calibrated with a small number of regional sample plots. Although forest structures differ among different parts of Finland, the nationwide volume and biomass models performed quite well (leave-inventory-area-out RMSE 22.3% to 33.8%, mean difference [MD] â13.8% to 18.7%) compared with regional models (leave-plot-out RMSE 20.2% to 26.8%). However, the nationwide dominant height model (RMSE 5.4% to 7.7%, MD â2.0% to 2.8%, with the exception of the Tornio region â RMSE 11.4%, MD â9.1%) performed nearly as well as the regional models (RMSE 5.2% to 6.7%). The results show that the nationwide volume and biomass models provided different means than real means at regional level, because forest structure and ALS device have a considerable effect on the predictions. Large MDs appeared especially in northern Finland. Local calibration decreased the MD and RMSE of volume and biomass models. However, the nationwide dominant height model did not benefit much from calibration.

Field measurements conducted on sample plots are a major cost component in airborne laser scanning (ALS) based forest inventories, as field data is needed to obtain reference variables for the statistical models. The ALS data also provides an excellent source of prior information that may be used in the design phase of the field survey to reduce the size of the field data set. In the current study, we acquired two independent modeling data sets: one with ALS-assisted and another with random plot selection. A third data set was used for validation. One canopy height and one canopy density variable were used as a basis for the ALS-assisted selection. Ordinary and partial least squares regressions for stem volume were fitted for four different strata using the two data sets separately. The results show that the ALS-assisted plot selection helped to decrease the root mean square error (RMSE) of the predicted volume. Although the differences in RMSE were relatively small, models based on random plot selection showed larger mean differences from the reference in the independent validation data. Furthermore, a sub-sampling experiment showed that 40 well placed plots should be enough for fairly reliable predictions.

Accurate field measurement of the forest canopy cover is too laborious to be used in extensive forest inventories. A possible alternative to the separate canopy cover measurements is to utilize the correlations between the percent canopy cover and easier-to-measure forest variables, especially the basal area. A fairly new analysis technique, the beta regression, is specially designed for modelling percentages. As an extension to the generalized linear models, the beta regression takes into account the distribution of the model residuals, and uses a logistic link function to ensure logical predictions. In this study, the beta regression method was found to perform well in conifer dominated study area located in central Finland. The same model shape, with basal area, tree height and an additional predictor (Scots pine: site fertility, Norway spruce: percentage of hardwoods) as independent variables, produced good results for both pine and spruce dominated sites. The models had reasonably high pseudo R-squared values (pine: 0.91, spruce: 0.87) and low standard errors (pine: 6.3%, spruce: 5.9%) for the fitting data, and also performed well in a cross validation test. The models were also tested on separate test plots located in a different geographical area, where the prediction errors were slightly larger (pine: 8.8%, spruce: 7.4%). In pine plots, the model fit was further improved by introducing additional predictors such as stand age and density. This improved also the performance of the models in the cross validation test, but weakened the results for the external data set. Our results indicated that the beta regression method offers a noteworthy alternative to separate canopy cover measurements, especially if time is limited and the models can be applied in the same region where the modelling data were collected.

Estimation of forest canopy cover has recently been included in many forest inventory programmes. In this study, after discussing how canopy cover is defined, different ground-based canopy cover estimation techniques are compared to determine which would be the most feasible for a large scale forest inventory. Canopy cover was estimated in 19 Scots pine or Norway spruce dominated plots using the Cajanus tube, line intersect sampling, modified spherical densiometer, digital photographs, and ocular estimation. The comparisons were based on the differences in values acquired with selected techniques and control values acquired with the Cajanus tube. The statistical significance of the differences between the techniques was tested with the nonparametric Kruskall-Wallis analysis of variance and multiple comparisons. The results indicate that different techniques yield considerably different canopy cover estimates. In general, labour intensive techniques (the Cajanus tube, line intersect sampling) provide unbiased and more precise estimates, whereas the estimates provided by fast techniques (digital photographs, ocular estimation) have larger variances and may also be seriously biased.

Very high point density laser scanning data from unmanned aerial systems (UAS) can be used to segment the trunks of individual trees. Such segmentation (e.g., individual trunk segmentation (ITS)) is useful, for example, in the estimation of diameter at breast height (DBH), which in turn is needed for the estimation of other tree and stand attributes, such as stem volume and diameter of the basal area median tree (DGM). In this paper, we assess the estimation of DBH directly from UAS LiDAR data in open and closed canopy conditions that represent the range of operational conditions encountered in boreal forests in Finland. We also compare trunk-based DBH estimates to corresponding estimates from individual tree crown segmentation (ITC) and fuse the results from the trunk- and crown-based estimates. The results showed that trunk segmentation performed slightly better than ITC in open canopy areas, whereas ITC performed better in closed canopy areas. The DBH prediction error was smaller for ITS (3.0 cm) than ITC (3.9 cm) when considering the trees that were recognized by both methods. We also conclude that a hybrid method, where both segmented tree trunks and tree crowns are fused, considerably increases the number of correctly segmented trees but does not decrease the prediction error associated with DGM compared to using either ITC or ITS individually. Key words: forest inventory, laser scanning, drone, segmentation. Les donnees obtenues par balayage par faisceau laser avec densite de points tres elevee par les systemes d'aeronefs sans pilote (UAS) peuvent etre utilisees pour segmenter les troncs des arbres individuels. Une telle segmentation (p. ex., segmentation d'un tronc individuel (STI)) est utile, par exemple, dans l'estimation du diametre a hauteur d'homme (DHH), laquelle est par ailleurs requise pour l'estimation des attributs des autres arbres et du peuplement forestier, par exemple le volume du tronc et la surface terriere moyenne des arbres (DGM). Dans la presente communication, nous evaluons l'estimation du DHH a partir directement des donnees de l'UAS LiDAR dans des conditions de couvert ouvert et ferme qui representent l'eventail des conditions operationnelles rencontrees dans les forets boreales en Finlande. Nous comparons egalement les estimes du DHH fonde sur les troncs aux estimes correspondants de la segmentation de la couronne des arbres individuels (CAI) et nous fusionnons les resultats des estimes fondes sur les troncs et les couronnes. Les resultats ont indique que la segmentation des troncs fonctionnait un peu mieux que la CAI dans les secteurs de couverts ouverts alors que la CAI fonctionnait mieux dans des secteurs de couverts fermes. L'erreur de prevision du DHH etait plus petite pour la STI (3,0 cm) que la CAI (3,9 cm) lorsqu'on tenait compte des arbres qui etaient reconnus par les deux methodes. Nous concluons egalement qu'une methode hybride oU les troncs et les couronnes des arbres segmentes sont fusionnes accroit considerablement le nombre d'arbres correctement segmentes, mais ne diminue par l'erreur de prevision associee a la DGM comparativement a l'utilisation de la CAI ou de la STI individuellement. [Traduit par la Redaction] Mots-cles : inventaire forestier, balayage par faisceau laser, drone, segmentation., 1. Introduction The rapid development of unmanned aerial systems (UAS) and the miniaturization of optical sensors and light detection and ranging (LiDAR) instruments has been followed by an extensive research [...]

Most forest growth studies using airborne laser scanning (ALS) consider how the changes in forest attributes are observed in repeated ALS data acquisitions, but the prediction of future forest growth from ALS data is still a rarely discussed topic. This study examined the prediction of the periodic annual increment (PAI) of the width of tree rings over a period of 10 years. The requirement for this approach is that ALS data are acquired at the beginning of the growth period. This is followed by field measurements of growth by drilling after a given growth period. The PAI was modelled in terms of ALS metrics by using the principle of the area-based approach. The metrics related to intensity were particularly significant as predictors, whereas the effective leaf area index was not. The root-mean-square error (RMSE) of the predictions was slightly over 21%. Additional field information (soil type, management operations) improved the RMSE by 2.7 percentage units. Key words: area-based approach, diameter increment, forest growth, leaf area index, LiDAR. La plupart des etudes sur l'accroissement forestier utilisant le balayage par laser aeroporte (BLA) tiennent compte de la facon dont les changements des attributs forestiers sont observes dans les acquisitions de donnees repetees du BLA, mais la prevision d'un accroissement forestier futur a partir des donnees du BLA est un theme qui fait encore rarement l'objet de discussions. La presente etude a examine la prevision de l'augmentation annuelle periodique (AAP) de la largeur des anneaux de croissance des arbres sur une periode de 10 ans. La necessite d'adopter cette approche est que les donnees du BLA sont acquises au debut de la periode de croissance. Ceci est suivi de mesures de la croissance sur le terrain en faisant un forage apres une periode donnee de croissance. L'AAP a ete modelee en fonction des mesures du BLA en utilisant le principe de l'approche locale. Les mesures reliees a l'intensite etaient particulierement importantes comme variable explicative alors que l'indice foliaire reel ne l'etait pas. L'EMQ des previsions etait legerement au-dessus de 21%. Des renseignements additionnels sur le terrain (type de sol, operations de gestion) ont ameliore l'EMQ de 2,7 unites de pourcentage. [Traduit par la Redaction] Mots-cles : approche locale, augmentation du diametre, accroissement forestier, indice foliaire, LiDAR., Introduction Data acquired through airborne laser scanning (ALS) are widely used in forestry, with the forest inventories being one of the main applications of this ALS data. The main interest [...]

Semi- and nonparametric models are popular in the area-based approach (ABA) using airborne laser scanning. It is unclear, however, how many predictors and training plots are needed to provide accurate predictions without overfitting. This work aims to explore these limits for various approaches: ordinary least squares regression (OLS), generalized additive models (GAM), least absolute shrinkage and selection operator (LASSO), random forest (RF), support vector machine (SVM), and Gaussian process regression (GPR). We modeled timber volume ([m.sup.3]*[ha.sup.-1]) for four boreal sites using ABA with 2-39 predictors and 20-500 training plots. OLS, GAM, LASSO, and SVM overfitted as the number of predictors approached the number of training plots. They required [greater than or equal to] 15 plots per predictor to provide accurate predictions (RMSE [less than or equal to] 30%). GAM required [greater than or equal to] 250 plots regardless of the number of predictors. The number of predictors only mildly affected RF and GPR, but they required [greater than or equal to] 200 and [greater than or equal to] 250 training plots, respectively. RF did not overfit in any circumstances, whereas GPR overfit even with 500 training plots. Overall, using up to 39 predictors did not generally result in overfit, and for most model types, it resulted in better accuracy for sufficiently large datasets ([greater than or equal to] 250 plots). Key words: LiDAR, machine learning, remote sensing, area-based approach, sampling size. Les modeles semi et non parametriques sont populaires avec l'approche territoriale (AT) qui utilise le balayage laser aeroporte. On ignore cependant combien de predicteurs et de parcelles d'entrainement sont necessaires pour generer des predictions justes sans surapprentissage. Cet article vise a explorer ces limites pour differentes approches : regression des moindres carres ordinaire (MCO), modeles additifs generaux (MAG), operateur de selection et reduction par moindres valeurs absolues (LASSO); foret aleatoire (FA), machine a vecteurs de support (MVS), et regression de processus gaussien (RPG). Nous avonsmodelise le volume de bois ([m.sup.3]*[ha.sup.-1]) dans quatre stations boreales a l'aide de l'AT avec 2 a 39 predicteurs et 20 a 500 parcelles d'entrainement. Les approches MCO, MAG, LASSO et MVS surajustaient lorsque le nombre de predicteurs approchait le nombre de parcelles d'entrainement. Elles exigeaient au moins 15 parcelles par predicteur pour generer des predictions justes (EMQ [less than or equal to] 30%). L'approche MAG necessitait au moins 250 parcelles peu importe le nombre de predicteurs. Le nombre de predicteurs influencait seulement legerement les approches FA et RPG mais elles exigeaient respectivement au moins 200 et 250 parcelles. L'approche FA ne surajustait pas quelles que soient les circonstances, tandis que l'approche RPG surajustait meme avec 500 parcelles d'entrainement. Dans l'ensemble, l'utilisation de 39 predicteurs n'engendrait generalement pas de surapprentissage et, avec la plupart des types de modeles, cela se traduisait par une precision accrue pour des jeux de donnees suffisamment importants (au moins 250 parcelles). [Traduit par la Redaction] Mots-cles : LiDAR, apprentissage machine, teledetection, approche territoriale, taille de l'echantillon., 1. Introduction The development of prediction models is an essential step when performing forest assessments through airborne laser scanning (ALS). In the area-based approach (ABA) to forest inventory with ALS [...]

Trestima® is a computer vision-based smartphone application that utilises relascope theory to obtain estimates of forest attributes from smartphone photographs. The aim of this study was to investigate the accuracy of Trestima es - timation and evaluate whether it is sufficiently accurate for operational use in forestry. Our data consisted of 37 forest stands, encompassing 73.5 ha in southeastern Finland, where Trestima estimates were obtained by forestry professionals during their work. The results were compared with harvester data obtained from clear-cut stands. The number of photographs taken per stand ranged between 1-29 (average: 7.3; standard deviation: 5.0). The total amount of industrial roundwood harvested from the stands was 21,531 m3 and the average harvest removal per hectare was 282 m3 . The accuracy of Trestima estimation was relatively good when ≥ 10 photographs per stand were taken. In this case, the root mean square error percent (RMSE%) value associated with roundwood volume was 17.7%. When the number of photographs per stand was < 10, the accuracy of Trestima was much weaker (RMSE% 22.7- 55.3%). On average, Trestima underestimated harvested volumes in Scots pine (Pinus sylvestris L.) stands (Bias% 11.4-89.2), although the bias was smaller (Bias% –12.7-12.4) with Norway spruce (Picea abies [L.] Karst.) stands. The Trestima smartphone application is a possible option for traditional field measurements in operational forestry, provided that its usage instructions are strictly followed, which is not always the case in practice. [ABSTRACT FROM AUTHOR]

In this paper, we examine the transferability of airborne laser scanning (ALS) based models for individual-tree detection (ITD) from one ALS inventory area (A1) to two other areas (A2 and A3). All areas were located in eastern Finland less than 100 km from each other and were scanned using different ALS devices and parameters. The tree attributes of interest were diameter at breast height (Dbh), height (H), crown base height (Cbh), stem volume (V), and theoretical sawlog volume (Vlog) of Scots pine (Pinus sylvestris L.) with Dbh [greater than or equal to] 16 cm. All trees were first segmented from the canopy height models, and various ALS metrics were derived for each segment. Then only the segments covering correctly detected pines were chosen for further inspection. The tree attributes were predicted using the k-nearest neighbor (k-NN) imputation. The results showed that the relative root mean square error (RMSE%) values increased for each attribute after the transfers. The RMSE% values were, for A1, A2, and A3, respectively: Dbh, 13.5%, 14.8%, and 18.1%; H, 3.2%, 5.9%, and 6.2%; Cbh, 13.3%, 15.3%, and 18.3%; V, 29.3%, 35.4%, and 39.1%; and Vlog, 38.2%, 54.4% and 51.8%. The observed values indicate that it may be possible to employ ALS-based tree- level k-NN models over different inventory areas without excessive reduction in accuracy, assuming that the tree species are known to be similar. Key words: airborne laser scanning, individual-tree detection, k-nearest neighbor, k-NN, transferability. Dans cet article, nous examinons la transferabilite des modeles fondes sur le balayage laser aeroporte (BLA) pour la detection d'arbres individuels (DAI) d'une zone inventoriee par BLA (Z1) a deux autres zones (Z2 et Z3). Toutes les zones etaient situees dans l'est de la Finlande, a moins de 100 km les unes des autres, et ont ete numerisees a l'aide de differents dispositifs et parametres de BLA. Les attributs pertinents des arbres etaient le diametre a hauteur de poitrine (DHP), la hauteur (H), la hauteur de la base de la cime (HBC), le volume de la tige (V) et le volume theorique des billes de sciage (VSci) dans le tronc de pins sylvestres (Pinus sylvestris L.) avec un DHP [greater than or equal to] 16 cm. Tous les arbres ont d'abord ete delimites a partir des modeles de hauteur du couvert forestier et diverses metriques du BLA ont ete derivees pour chaque cime identifiee. Puis, seules les cimes correctement detectees comme des pins ont ete choisies pour la suite des evaluations. Les attributs de l'arbre ont ete predits a l'aide de l'imputation des k-plus proches voisins (k-PPV). Les resultats ont montre que les erreurs quadratiques moyennes augmentaient pour chaque attribut apres les transferts avec respectivement pour Z1, Z2 et Z3 : DHP, 13,5 /14,8 /18,1 %; H, 3,2 / 5,9 / 6,2 %; HBC, 13,3 /15,3 / 18,3 %; V, 29,3 / 35,4 / 39,1 %; et VSci, 38,2 / 54,4 / 51,8 %. Les valeurs observees indiquent qu'il est possible d'utiliser des modeles k-PPV, fondes sur un BLA a l'echelle de l'arbre, sur differentes zones d'inventaire sans reduction excessive de la precision, a condition que l'espece d'arbre soit reconnue comme similaire. [Traduit par la Redaction] Mots-cles : balayage laser aeroporte, detection d'arbres individuels, k-plus proches voisins, k-PPV, transferabilite., Introduction Airborne laser scanning (ALS) has proven to be very useful in forest inventories (e.g., Vauhkonen et al. 2014a). The two main approaches for deriving forest information from ALS data [...]

Copyright of Methods in Ecology & Evolution is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The estimation of individual biomass components within tree crowns, such as dead branches (DB), needles (NB), and branch biomass (BB), has received limited attention in the scientific literature despite their significant contribution to forest biomass. This study aimed to assess the potential of multispectral LiDAR data for estimating these biomass components in a multi-layered Abies borissi-regis forest. Destructive (i.e., 13) and non-destructive (i.e., 156) field measurements were collected from Abies borisii-regis trees to develop allometric equations for each crown biomass component and enrich the reference data with the non-destructively sampled trees. A set of machine learning regression algorithms, including random forest (RF), support vector regression (SVR) and Gaussian process (GP), were tested for individual-tree-level DB, NB and BB estimation using LiDAR-derived height and intensity metrics for different spectral channels (i.e., green, NIR and merged) as predictors. The results demonstrated that the RF algorithm achieved the best overall predictive performance for DB (RMSE% = 17.45% and R2 = 0.89), NB (RMSE% = 17.31% and R2 = 0.93) and BB (RMSE% = 24.09% and R2 = 0.85) using the green LiDAR channel. This study showed that the tested algorithms, particularly when utilizing the green channel, accurately estimated the crown biomass components of conifer trees, specifically fir. Overall, LiDAR data can provide accurate estimates of crown biomass in coniferous forests, and further exploration of this method's applicability in diverse forest structures and biomes is warranted. [ABSTRACT FROM AUTHOR]

Large-diameter trees are important for both ecological and economic reasons, but they have become increasingly rare. Thus, there is an interest in easily locating such trees, and for this purpose, the use of airborne laser scanning (ALS) seems suitable. Our objective was to assess the accuracy of area-based ALS estimation in predicting the number of large- diameter Scots pines (Pinus sylvestris L.). A sample of 856 trees with a diameter > 35 cm were measured from 1109 sample plots located in eastern Norway. We fitted negative binomial and zero-inflated negative binomial models for predicting large-diameter tree counts. ALS-derived and external variables were used as predictors when fitting the models. The accuracy was assessed based on the weighted kappa coefficient and cross validation. Our best model was based on three ALS height distribution variables, one horizontal ALS variable, and plot elevation. Its overall accuracy was 65.8% and the weighted kappa was 0.55. Although there was a clear relationship between the response and the proposed predictor variables, fairly large errors in the predicted large-diameter tree counts were common. Key words: large trees, LiDAR, area-based method, negative binomial regression, old growth. Les arbres de grand diametre sont importants, a la fois pour des raisons ecologiques et economiques, mais ils sont devenus de plus en plus rares. Il y a donc un interet a localiser facilement ces arbres et l'utilisation du balayage laser aeroporte (BLA) semble appropriee pour ce faire. Notre objectif etait d'evaluer la precision d'une estimation basee sur la surface obtenue par BLA pour predire le nombre de pins sylvestres (Pinus sylvestris L.) de grand diametre. Un echantillon de 856 arbres, d'un diametre > 35 cm, a ete mesure a partir de 1109 placettes echantillons situees dans l'est de la Norvege. Nous avons ajuste des modeles binomiaux negatifs et des modeles binomiaux negatifs a exces de zeros pour predire le nombre d'arbres de grand diametre. Des variables obtenues par BLA et des variables externes ont ete utilisees comme predicteurs lors de l'ajustement des modeles. La precision a ete evaluee sur la base du coefficient kappa pondere et par validation croisee. Notre meilleur modele etait base sur trois variables de distribution de hauteur obtenues par BLA, une variable horizontale obtenue par BLA et l'altitude de la placette. Sa precision globale etait de 65,8 % et le coefficient kappa pondere etait de 0,55. Bien qu'il y ait une relation claire entre la variable de reponse et les variables predictives proposees, d'assez grandes erreurs dans la prediction du nombre d'arbres de grand diametre etaient courantes. [Traduit par la Redaction] Mots-cles: grands arbres, LiDAR, methode d'estimation basee sur la surface, regression binomiale negative, foret ancienne., 1. Introduction Large trees are rare in many forests and difficult to find. They are important for both economic and ecological reasons. Large trees dominate the structure, dynamics, and function [...]