Tropical tree size-frequency distributions from airborne lidar

Rethinking the complexity and uncertainty of spatial networks applied to forest ecology

Characterizing tree spatial patterns and interactions are helpful to reveal underlying processes assembling forest communities. Spatial networks, despite their complexity, are powerful to examine spatial interactions at an individual level using well-defined patterns. However, complex forestation networks introduce uncertainties. Validation methods are needed to assess whether network-based metrics can identify different processes. Here, we constructed three types of networks, which reflect various aspects of tree competition. Based on five spatial null models and 199 Monte-Carlo simulations, we were able to select network-based metrics that exhibited well performance in distinguishing different processes. This technique was then applied to a tropical forest dataset in Costa Rica. We found that the average node degree and the clustering coefficient are good metrics like the paired correlation function. In addition, the network approach can identify fine-scale spatial variations of tree competition and its underlying causes. Our analyzes also indicate that a bit of caution is needed when defining the network structure as well as designing network-based metrics. We suggested that validation techniques using corresponding spatial null models are critically important to reduce the negative effects caused by uncertainties of the network.

Estimating Species-Specific Stem Size Distributions of Uneven-Aged Mixed Deciduous Forests Using ALS Data and Neural Networks

Sustainable forest management requires accurate fine-scale description of wood resources. Stem size distribution (SSD) by species is used by foresters worldwide as a representative overview of forest structure and species composition suitable for informing management decisions at shorter and longer terms. In mixed uneven-aged deciduous forests, tree data required for SSD estimation are most often collected in the field through traditional forest management inventories (FMIs), but these are time-consuming and costly with respect to the sampled area. Combining FMIs with remote sensing methods such as airborne laser scanning (ALS), which has high potential for predicting forest structure and composition, and is becoming increasingly accessible and affordable, could provide cheaper and faster SSD data across large areas. In this study, we developed a method for estimating species-specific SSDs by combining FMIs and dual-wavelength ALS data using neural networks (NNs). The proposed method was tested and validated using 178 FMI plots within 22,000 ha of a mixed uneven-aged deciduous forest in Belgium. The forest canopy was segmented, and metrics were derived from the ALS point cloud. A NN with a custom architecture was set up to simultaneously predict the three components required to compute species-specific SSDs (species, circumference, and number of stems) at segment level. Species-specific SSDs were thereafter estimated at stand level by aggregating the estimates for the segments. A robustness test was set up using fully independent plots to thoroughly assess the method precision at stand-level on a larger area. The global Reynolds index for the species-specific SSDs was 21.2 for the training dataset and 54.0 for the independent dataset. The proposed method does not require allometric models, prior knowledge of the structure, or the predefinition of variables; it is versatile and thus potentially adaptable to other forest types having different structures and compositions.

Methodology of Calculating the Number of Trees Based on ALS Data for Forestry Applications for the Area of Samławki Forest District

Airborne Laser Scanning (ALS) is a technology often used to study forest areas. The main area of application of ALS in forests is collecting data to determine the height of individual trees and entire stands, tree density and stand biomass. The content of the ALS data is also classified, i.e., registered objects are identified, including the species affiliation of individual trees. Important information for forest districts includes other parameters related to the structure and share of stands and the number of trees in the forest district. The main goal of this study was to propose the new ALS data processing methodology for detecting single trees in the Samławki Forest District. The idea of the proposed methodology is to indicate a free and accessible solution for any user (at least in Poland). This new ALS data processing methodology contributes to research on the use of ALS data in forest districts to maintain up-to-date and accurate stand statistics. This methodology was based on free data from the geoportal.gov.pl portal and free software, which allowed to minimize the costs of preparing data for the needs of forestry activities. In cooperation with the Samławki Forest District, the proposed methodology was used to detect the number and heights of trees for two forest addresses 13b and 30a, and then to calculate the volume of stands. As a result, the volume of the analyzed stands was calculated, obtaining values differing from the nominal ones included in the FMP (Forest Management Plan) by about 25% and 5%, respectively, for larch and oak.

Tree Crowns Cause Border Effects in Area-Based Biomass Estimations from Remote Sensing

The estimation of forest biomass by remote sensing is constrained by different uncertainties. An important source of uncertainty is the border effect, as tree crowns are not constrained by plot borders. Lidar remote sensing systems record the canopy height within a certain area, while the ground-truth is commonly the aboveground biomass of inventory trees geolocated at their stem positions. Hence, tree crowns reaching out of or into the observed area are contributing to the uncertainty in canopy-height–based biomass estimation. In this study, forest inventory data and simulations of a tropical rainforest’s canopy were used to quantify the amount of incoming and outgoing canopy volume and surface at different plot sizes (10, 20, 50, and 100 m). This was performed with a bottom-up approach entirely based on forest inventory data and allometric relationships, from which idealized lidar canopy heights were simulated by representing the forest canopy as a 3D voxel space. In this voxel space, the position of each voxel is known, and it is also known to which tree each voxel belongs and where the stem of this tree is located. This knowledge was used to analyze the role of incoming and outgoing crowns. The contribution of the border effects to the biomass estimation uncertainty was quantified for the case of small-footprint lidar (a simulated canopy height model, CHM) and large-footprint lidar (simulated waveforms with footprint sizes of 23 and 65 m, corresponding to the GEDI and ICESat GLAS sensors). A strong effect of spatial scale was found: e.g., for 20-m plots, on average, 16% of the CHM surface belonged to trees located outside of the plots, while for 100-m plots this incoming CHM fraction was only 3%. The border effects accounted for 40% of the biomass estimation uncertainty at the 20-m scale, but had no contribution at the 100-m scale. For GEDI- and GLAS-based biomass estimates, the contributions of border effects were 23% and 6%, respectively. This study presents a novel approach for disentangling the sources of uncertainty in the remote sensing of forest structures using virtual canopy modeling.

Estimating Forest Structure from UAV-Mounted LiDAR Point Cloud Using Machine Learning

Monitoring the structure of forest stands is of high importance for forest managers to help them in maintaining ecosystem services. For that purpose, Unmanned Aerial Vehicles (UAVs) open new prospects, especially in combination with Light Detection and Ranging (LiDAR) technology. Indeed, the shorter distance from the Earth’s surface significantly increases the point density beneath the canopy, thus offering new possibilities for the extraction of the underlying semantics. For example, tree stems can now be captured with sufficient detail, which is a gateway to accurately locating trees and directly retrieving metrics—e.g., the Diameter at Breast Height (DBH). Current practices usually require numerous site-specific parameters, which may preclude their use when applied beyond their initial application context. To overcome this shortcoming, the machine learning Hierarchical Density-Based Spatial Clustering of Application of Noise (HDBSCAN) clustering algorithm was further improved and implemented to segment tree stems. Afterwards, Principal Component Analysis (PCA) was applied to extract tree stem orientation for subsequent DBH estimation. This workflow was then validated using LiDAR point clouds collected in a temperate deciduous closed-canopy forest stand during the leaf-on and leaf-off seasons, along with multiple scanning angle ranges. The results show that the proposed methodology can correctly detect up to 82% of tree stems (with a precision of 98%) during the leaf-off season and have a Maximum Scanning Angle Range (MSAR) of 75 degrees, without having to set up any site-specific parameters for the segmentation procedure. In the future, our method could then minimize the omission and commission errors when initially detecting trees, along with assisting further tree metrics retrieval. Finally, this research shows that, under the study conditions, the point density within an approximately 1.3-meter height above the ground remains low within closed-canopy forest stands even during the leaf-off season, thus restricting the accurate estimation of the DBH. As a result, autonomous UAVs that can both fly above and under the canopy provide a clear opportunity to achieve this purpose.

Deriving Tree Size Distributions of Tropical Forests from Lidar

Remote sensing is an important tool to monitor forests to rapidly detect changes due to global change and other threats. Here, we present a novel methodology to infer the tree size distribution from light detection and ranging (lidar) measurements. Our approach is based on a theoretical leaf–tree matrix derived from allometric relations of trees. Using the leaf–tree matrix, we compute the tree size distribution that fit to the observed leaf area density profile via lidar. To validate our approach, we analyzed the stem diameter distribution of a tropical forest in Panama and compared lidar-derived data with data from forest inventories at different spatial scales (0.04 ha to 50 ha). Our estimates had a high accuracy at scales above 1 ha (1 ha: root mean square error (RMSE) 67.6 trees ha−1/normalized RMSE 18.8%/R² 0.76; 50 ha: 22.8 trees ha−1/6.2%/0.89). Estimates for smaller scales (1-ha to 0.04-ha) were reliably for forests with low height, dense canopy or low tree height heterogeneity. Estimates for the basal area were accurate at the 1-ha scale (RMSE 4.7 tree ha−1, bias 0.8 m² ha−1) but less accurate at smaller scales. Our methodology, further tested at additional sites, provides a useful approach to determine the tree size distribution of forests by integrating information on tree allometries.

Impacts of Degradation on Water, Energy, and Carbon Cycling of the Amazon Tropical Forests

Selective logging, fragmentation, and understory fires directly degrade forest structure and composition. However, studies addressing the effects of forest degradation on carbon, water, and energy cycles are scarce. Here, we integrate field observations and high-resolution remote sensing from airborne lidar to provide realistic initial conditions to the Ecosystem Demography Model (ED-2.2) and investigate how disturbances from forest degradation affect gross primary production (GPP), evapotranspiration (ET), and sensible heat flux (H). We used forest structural information retrieved from airborne lidar samples (13,500 ha) and calibrated with 817 inventory plots (0.25 ha) across precipitation and degradation gradients in the eastern Amazon as initial conditions to ED-2.2 model. Our results show that the magnitude and seasonality of fluxes were modulated by changes in forest structure caused by degradation. During the dry season and under typical conditions, severely degraded forests (biomass loss ≥66%) experienced water stress with declines in ET (up to 34%) and GPP (up to 35%) and increases of H (up to 43%) and daily mean ground temperatures (up to 6.5°C) relative to intact forests. In contrast, the relative impact of forest degradation on energy, water, and carbon cycles markedly diminishes under extreme, multiyear droughts, as a consequence of severe stress experienced by intact forests. Our results highlight that the water and energy cycles in the Amazon are driven by not only climate and deforestation but also the past disturbance and changes of forest structure from degradation, suggesting a much broader influence of human land use activities on the tropical ecosystems.