An Examination of Diameter Density Prediction with k-NN and Airborne Lidar

Strunk, Jacob L.; Gould, Peter; Packalén, Petteri; Poudel, Krishna P.; Andersen, Hans‐Erik; Temesgen, Hailemariam

doi:10.3390/f8110444

Cited by 13 publications

(17 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With respect to the performance of MSN and RF imputation for diameter distributions, it was found intuitively from Figure 5 that most of RF imputed distributions had higher accuracies than MSN among different response configurations, except that MSN performed slightly better than RF when the response configuration was SET1-2 for all plots, and SET4-5 and SET13-14 for Eucalypt. Similar to Strunk et al [35] and Shang et al [59], RF was fairly insensitive to the choice of response configurations and was more robust than MSN. As a robust method against overfitting, the superiority of RF for predicting forest structure attributes had been reported by previous studies [56,57,86,106].…”

Section: The Selection Of Number Of Neighbors Response Configurationsupporting

confidence: 60%

“…Another reason explaining these results may be the fact that RF has an iterative procedure and the average outcome does not describe the worst possible case [57], which may lead to a stable and robust predictive ability of RF for diameter distributions. However, a limitation of RF was that calculating distances is more time-consuming than MSN [35]. As shown in Figure 5, generally, we also found that there was no significant improvement in separate MSN imputations by tree species compared to the imputed diameter distributions for all plots, although there was a significant improvement in RF imputation (Figure 5b).…”

Section: The Selection Of Number Of Neighbors Response Configurationmentioning

confidence: 71%

“…In recent years, due to strong correlation with LiDAR-derived forest structure attributes (e.g., DBH, basal area, volume, stem density, etc.) [33], airborne LiDAR has shown potential in predicting diameter distributions using a variety of strategies [24,27,28,[34][35][36][37][38][39]. One such strategy, using parametric distributions, is based on the assumption that the diameter distribution's probability density can be characterized by a theoretical probability density function (PDF) [22,35], e.g., Weibull, beta, gamma and Log-normal [27].…”

Section: Introductionmentioning

confidence: 99%

“…A non-parametric strategy with an effective ability to predict multimodal diameter distributions is a k-Nearest Neighbor (KNN) imputation model [22,33,35]. KNN is a reliable strategy in more irregular forest stand structures, because KNN does not rely on any underlying probability distribution [49][50][51] and can accurately describe the local variability of diameter distributions [37,52].…”

Section: Introductionmentioning

confidence: 99%

“…Penner et al [58] found the potential of this imputation method to predict tree diameter distributions in a northeastern Ontario boreal forest. Strunk et al [35] also evaluated RF imputation strategies for prediction of tree diameter distributions with tree species using LiDAR data in a boreal forest. The utility of RF imputation also been reported by Shang et al [59].Using LiDAR data for predicting diameter distributions are usually based on UWM approaches that are difficult to implement for complex forest stands, whereas studies related to using multimodal FMM or KNN approach to predict multimodal diameter distributions are few.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Prediction of Diameter Distributions with Multimodal Models Using LiDAR Data in Subtropical Planted Forests

Zhang

Cao

Mulverhill

et al. 2019

Forests

View full text Add to dashboard Cite

Tree diameter distributions are essential for the calculation of stem volume and biomass, as well as simulation of growth and yield and to understand timber assortments. Accurate and reliable prediction of tree diameter distributions is critical for optimizing forest structure compositions, scheduling silvicultural operations and promoting sustainable management. In this study, we investigated the potential of airborne Light Detection and Ranging (LiDAR) data for predicting tree diameter distributions using a bimodal finite mixture model (FMM) and a multimodal k-nearest neighbor (KNN) model (compared to the unimodal Weibull model (UWM)) over a subtropical planted forest in southern China. To do so, we first evaluated the capability of various LiDAR predictions (i.e., the bimodality coefficient (BC) and Lorenz-based indicators) to stratify forest structural types into unimodal and multimodal stands. Once the best LiDAR prediction for the differentiation was determined, the parameters of UWM (in non-specific and species-specific models) and FMM (in structure-specific models) were estimated by LiDAR-derived metrics and the tree diameter distributions of stands were generated by the estimated LiDAR parameters. When KNN was applied for constructing diameter distributions, optimal KNN strategies, including number of neighbors k, response configurations and imputation methods (i.e., Most Similar Neighbor (MSN) and Random Forest (RF)) for different species were heuristically determined. Finally, the predictive performance of estimated LiDAR the parameters of UWM, FMM and KNN for predicting diameter distributions were assessed. The results showed that LiDAR-predicted Lorenz-based indicators performed best for differentiation. Parameters of UWM and FMM were predicted well and the species-specific models had higher accuracies than the non-specific models. Overall, RF imputation from KNN with an optimal response set (i.e., DBH) were was stable than MSN imputation when k = 5 neighbors. In addition, the inclusion of bimodal FMM for differentiated all plots generally produced a more accurate result (Mean eR = 40.85, Mean eP = 0.20) than multimodal KNN (Mean eR = 52.19, Mean eP = 0.26), whereas the UWM produced the lowest performance (Mean eR = 52.31, Mean eP = 0.26). This study demonstrated the benefits of multimodal models with LiDAR for estimating diameter distributions for supporting forest inventory and sustainable forest management in subtropical planted forests.

show abstract

Section: The Selection Of Number Of Neighbors Response Configurationsupporting

confidence: 60%

Section: The Selection Of Number Of Neighbors Response Configurationmentioning

confidence: 71%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Prediction of Diameter Distributions with Multimodal Models Using LiDAR Data in Subtropical Planted Forests

Zhang

Cao

Mulverhill

et al. 2019

Forests

View full text Add to dashboard Cite

show abstract

Comparing nearest neighbor configurations in the prediction of species-specific diameter distributions

Räty

Packalén

Maltamo

2018

Annals of Forest Science

Self Cite

View full text Add to dashboard Cite

& Key message We examine how the configurations in nearest neighbor imputation affect the performance of predicted species-specific diameter distributions. The simultaneous nearest neighbor imputation for all tree species and separate imputation by tree species are evaluated with total volume calibration as a prediction method for diameter distributions. & Context This study considers the predictions of species-specific diameter distributions in Finnish boreal forests by means of airborne laser scanning (ALS) data and aerial images. & Aims The aim was to investigate different configurations in non-parametric nearest neighbor (NN) imputation and to determine how changes in configurations affect prediction error rates for timber assortment volumes and the error indices of the diameter distributions. & Methods Non-parametric NN imputation was used as a modeling method and was applied in two different ways: (1) diameter distributions were predicted at the same time for all tree species by simultaneous NN imputation, and (2) diameter distributions were predicted for one tree species at a time by separate NN imputation. Calibration to a regression-based total volume prediction was applied in both cases. & Results The results indicated that significant changes in the volume prediction error rates for timber assortment and for error indices can be achieved by the selection of responses, calibration to total volume, and separate NN imputation by tree species. & Conclusion Overall, the selection of response variables in NN imputation and calibration to total volume improved the predicted diameter distribution error rates. The most successful prediction performance of diameter distribution was achieved by separate NN imputation by tree species.

show abstract

Comparison of two parameter recovery methods for the transformation of Pinus sylvestris yield tables into a diameter distribution model

Mauro

García‐Abril

Ayuga-Téllez

et al. 2021

Annals of Forest Science

View full text Add to dashboard Cite

Key messageWe successfully transformed Pinus sylvestris yield tables into diameter distribution models. The best results were obtained with the parameter recovery method based on both mean and quadratic mean diameter, which explained 70% of the variability of frequencies by diameter classes and provided better results in the analysis of errors. On the other hand, the method based on stand density, dominant diameter and quadratic mean diameter explained less variability of frequencies by diameter classes (64.4%).• Context Old datasets used to develop yield table models can be recovered to transform those yield tables into diameter distribution models that provide a more detailed description of size variability and forest structure.• Aims We used archived measurements collected to develop yield table models for Pinus sylvestris L in central Spain, to transform those yield tables into a diameter distribution model by using parameter recovery methods.• Methods We compared two different parameter recovery methods, one based on both mean and quadratic mean diameter and another one based on dominant diameter, stand density and quadratic mean diameter and used a set of 104 even aged plots to analyze the performance of the said methods for the transformation of Pinus sylvestris L yield tables in central Spain into a diameter distribution model. • ResultsThe parameter recovery method based on both mean and quadratic mean diameter explained 70% of the variability of frequencies by diameter classes and provided better results than the method based on stand density, dominant diameter and quadratic mean diameter that explained 64.4% of the variability of frequencies by diameter classes. However, more important than the method itself were the errors that propagated from the models predicting the different variables used in the parameter recovery. • ConclusionBased on the results from the analysis of errors by diameter classes, the method using both mean and quadratic mean diameter outperformed the method using dominant diameter, stand density and quadratic mean diameter and is the best option to transform P. sylvestris yield tables into diameter distribution models.

show abstract

An Examination of Diameter Density Prediction with k-NN and Airborne Lidar

Cited by 13 publications

References 25 publications

Prediction of Diameter Distributions with Multimodal Models Using LiDAR Data in Subtropical Planted Forests

Prediction of Diameter Distributions with Multimodal Models Using LiDAR Data in Subtropical Planted Forests

Comparing nearest neighbor configurations in the prediction of species-specific diameter distributions

Comparison of two parameter recovery methods for the transformation of Pinus sylvestris yield tables into a diameter distribution model

Contact Info

Product

Resources

About