LiDAR individual tree detection (ITD) is a promising tool for measuring forests at a scale that is meaningful ecologically and useful for forest managers. However, most ITD research evaluates methods over small homogeneous areas, while many forest managers work over large, complex landscapes. We investigated how ITD results varied across diverse structural conditions in California’s Sierra Nevada mixed-conifer forests and what this taught us about when and how to apply ITD. Our results suggest that it is advantageous to use ITD when it improves analysis interpretability, when measuring horizontal patterns, or when field data are unavailable. In the latter case, it is best to focus on measures dominated by large trees, like basal area and biomass. Thinking of ITD results as “tree-approximate objects” including one dominant tree and up to a few subordinate tree respects LiDAR’s strengths and limitations; we illustrate how this concept keeps analyses consistent across varying structural conditions.

Low-cost methods to measure forest structure are needed to consistently and repeatedly inventory forest conditions over large areas. In this study we investigate low-cost pushbroom Digital Aerial Photography (DAP) to aid in the estimation of forest volume over large areas in Washington State (USA). We also examine the effects of plot location precision (low versus high) and Digital Terrain Model (DTM) resolution (1 m versus 10 m) on estimation performance. Estimation with DAP and post-stratification with high-precision plot locations and a 1 m DTM was 4 times as efficient (precision per number of plots) as estimation without remote sensing and 3 times as efficient when using low-precision plot locations and a 10 m DTM. These findings can contribute significantly to efforts to consistently estimate and map forest yield across entire states (or equivalent) or even nations. The broad-scale, high-resolution, and high-precision information provided by pushbroom DAP facilitates used by a wide variety of user types such a towns and cities, small private timber owners, fire prevention groups, Non-Governmental Organizations (NGOs), counties, and state and federal organizations.

Relationships between discrete-return light detection and ranging (LiDAR) data and radiata pine leaf area index (LAI), stem volume, above ground carbon, and carbon sequestration were developed using 10 plots with directly measured biomass and leaf area data, and 36 plots with modelled carbon data. The plots included a range of genetic types established on north- and south-facing aspects. Modelled carbon was highly correlated with directly measured crown, stem, and above ground biomass data, with r = 0.92, 0.97 and 0.98, respectively. LiDAR canopy percentile height (P30) and cover, based on all returns above 0.5 m, explained 81, 88, and 93% of the variation in directly measured crown, stem, and above ground live carbon and 75, 89 and 88% of the modelled carbon, respectively. LAI (all surfaces) ranged between 8.8–19.1 in the 10 plots measured at age 9 years. The difference in canopy percentile heights (P95–P30) and cover based on first returns explained 80% of the variation in total LAI. Periodic mean annual increments in stem volume, above ground live carbon, and total carbon between ages 9 and 13 years were significantly related to (P95–P30), with regression models explaining 56, 58, and 55%, respectively, of the variation in growth rate per plot. When plot aspect and genetic type were included with (P95–P30), the R² of the regression models for stem volume, above ground live carbon, and total carbon increment increased to 90, 88, and 88%, respectively, which indicates that LiDAR regression equations for estimating stock changes can be substantially improved by incorporating supplementary site and crop data.

The application of lidar data to assist with forest inventory is common around the world. However, the determination of tree species is still somewhat elusive. Lidar data collected using UAS (uncrewed aircraft systems) platforms offer high density point cloud data for areas from a few to several hundred hectares. General point cloud metrics computed using these data captured differences in the crown structure that proved useful for species classification. For our study, we manually adjusted plot and tree locations to align field trees and UAS lidar point data and computed common descriptive metrics using a small cylindrical sample of points designed to capture the top three meters and leader of each tree. These metrics were used to train a random forest classifier to differentiate between two conifer species, Douglas fir and western hemlock, common in the Pacific Northwest region of the United States. Our UAS lidar data had a single swath pulse density of 90 pulses/m² and an aggregate pulse density of 556 pulses/m². We trained classification models using both height and intensity metrics, height metrics alone, intensity metrics alone, and a small subset of five metrics, and achieved overall accuracies of 91.8%, 88.7%, 78.6%, and 91.5%, respectively. Overall, we showed that UAS lidar data captured morphological differences between the upper crowns of our two target species and produced a classification model that could be applied over large areas.

A fundamental question of forestry is that of species composition: which species are present, and which are not. However, traditional forest measurements needed to map species over large areas can be both time consuming and costly. In this study, we combined airborne light detection and ranging (LiDAR) data with extensive field data from the Long-Term Ecosystem Productivity study located near Sappho, Washington, USA to increase the accuracy of our GIS data and to differentiate between red alder (Alnus rubra Bong.) and other dominant tree species. We adjusted plot and tree locations using LiDAR canopy height models (CHMs) by manually matching tree canopies on the CHMs with tree stem maps based on field data. We then used the adjusted tree locations and metrics computed from LiDAR point cloud data to create a classification model to identify and map red alder. The manual matching of field stem maps to CHMs improved tree locations, allowing us to create model training data. These data were used to create a random forest model that discriminated between red alder and conifer species with an accuracy of 96%. Our findings highlight the potential of LiDAR to improve coordinates of individual trees as well as discriminate between selected coniferous and deciduous tree species using LiDAR data collected in leaf-off conditions in Pacific Northwest ecosystems.