Tree-Lists Estimation for Chinese Boreal Forests by Integrating Weibull Diameter Distributions with MODIS-Based Forest Attributes from kNN Imputation

Abstract

Wall-to-wall tree-lists information (lists of species and diameter for every tree) at a regional scale is required for managers to assess forest sustainability and design effective forest management strategies. Currently, the k-nearest neighbors (kNN) method and the Weibull diameter distribution function have been widely used for estimating tree lists. However, the kNN method usually relies on a large number of field inventory plots to impute tree lists, whereas the Weibull function relies on strong correlations between stand attributes and diameter distribution across large regions. In this study, we developed a framework to estimate wall-to-wall tree lists over large areas based on a limited number of forest inventory plots. This framework integrates the ability of extrapolating diameter distribution from Weibull and kNN imputation of wall-to-wall forest stand attributes from Moderate Resolution Imaging Spectroradiometer (MODIS). We estimated tree lists using this framework in Chinese boreal forests (Great Xing’an Mountains) and evaluated the accuracy of this framework. The results showed that the passing rate of the Kolmogorov–Smirnov (KS) test for Weibull diameter distribution by species was from 52% to 88.16%, which means that Weibull distribution could describe the diameter distribution by species well. The imputed stand attributes (diameter at breast height (DBH), height, and age) from the kNN method showed comparable accuracy with the previous studies for all species. There was no significant difference in the tree density between the estimated and observed tree-lists. Results suggest that this framework is well-suited to estimating the tree-lists in a large area. Our results were also ecologically realistic, capturing dominant ecological patterns and processes.

Combining Weibull distribution and k-nearest neighbor imputation method to predict wall-to-wall tree lists for the entire forest region of Northeast China

“…For example, stand projection models (growth and yield models) require detailed stand attributes at the stand scale to set the initial conditions (Scolforo et al 2019), whereas forest ecosystem process research may require forest attributes at the ecotype scale (Dijak et al 2017). In this study, the stand mean DBH predicted using the kNN imputation model over Northeast China at the stand scale had a higher accuracy than that of similar studies conducted in Chinese boreal forests (Zhang et al 2018b). Owing to the relatively homogeneous temperature, humidity, topography, forest type, and species composition of each ecotype, the prediction accuracy of stand mean DBH at the ecotype scale was significantly improved, indicating that the map of stand mean DBH generated in this study is more suitable for application to forest ecosystem models (e.g., LINKAGES) operated at the ecotype scale.…”

“…In step 3, to assess the accuracy of the predicted wall-to-wall tree lists in 2000, the tree lists of the 641 sample plots in 2000 were first obtained as the validation data by transforming the tree lists of 641 sample plots from 2009 to 2013 by subtracting the DBH growth of the 17 species in the past 9-13 years using the age-DBH relationships summarized in previous studies (Xu et al 2020;Zhang et al 2018b). The central points of the 641 sample plots were used to extract the predicted wall-to-wall tree lists at the corresponding pixel locations.…”

“…Although previous studies have applied both parametric diameter distribution modeling (e.g., Weibull function) and nonparametric imputation (e.g., kNN model) to predict tree lists (Packalén and Maltamo 2008;Schütz and Rosset 2020;Shang et al 2017), the integration of these two methods proposed in our study is novel for generating wall-to-wall tree lists over large areas with limited forest inventory data. The Weibull function, especially the two-parameter Weibull function, is extensively used to describe the diameter distributions of single-and mixed-species stands (Hao et al 2022;Liu et al 2014;Schmidt et al 2020;Zhang et al 2018b). Although Liu et al (2014) found that a finite mixture model (FMM) of Weibull functions was more flexible for describing irregular and highly skewed diameter distributions for the whole plot in the mixed-species forest stands of the Greater Khingan Mountains in Northeast China than a single Weibull function to fit the whole plot only, a single Weibull function to fit each species component separately was more accurate than the FMM.…”

“…For example, the scale parameter was usually strongly correlated with the stand attributes (Zhang et al 2018b), consistent with previous studies, and the prediction accuracy of the scale parameter of the 17 tree species in this study was better than the results (R 2 = 0.96-0.99) obtained in the natural stands of the Crimean Juniper in the southern and southwestern (Mediterranean) region of Turkey (Diamantopoulou et al 2015). Furthermore, our studies led to a conclusion similar to that previously reported by Hao et al (2022), Wang et al (2006), andZhang et al (2018b), who found that the shape parameter is often more difficult to accurately predict than the scale parameter. This finding might be due to the remarkable complexity of the shape parameter in the mixture distribution to allow characterization by the stand mean DBH (Wang et al 2006).…”

“…The nonparametric k-nearest neighbor (kNN) imputation model has been widely applied to predict wallto-wall forest stand attributes across large regions as Fu et al Annals of Forest Science (2022) 79:42 it can simultaneously predict a suite of response variables, such as species-specific volume (Breidenbach et al 2010), species-level biomass (Fu et al 2019), basal area (Wilson et al 2012), or stand mean DBH (Zhang et al 2018b). kNN imputation is also a reliable strategy for the attribute prediction of complex forest stands as it can maintain both the spatial and attribute covariance structures of the stand attributes (Temesgen et al 2003).…”