Search

Accurate structural information about forests, including canopy heights and diameters, is crucial for quantifying tree volume, biomass, and carbon stocks, enabling effective forest ecosystem management, particularly in response to changing environmental conditions. Since late 2018, NASA's Global Ecosystem Dynamics Investigation (GEDI) mission has monitored global canopy structure using a satellite Light Detection and Ranging (LiDAR) instrument. While GEDI has collected billions of LiDAR shots across a near-global range (between 51.6°N and >51.6°S), their spatial distribution remains dispersed, posing challenges for achieving complete forest coverage. This study proposes and evaluates an approach that generates high-resolution canopy height maps by integrating GEDI data with Sentinel-1, Sentinel-2, and topographical ancillary data through three machine learning (ML) algorithms: random forests (RF), gradient tree boost (GB), and classification and regression trees (CART). To achieve this, the secondary aims included the following: (1) to assess the performance of three ML algorithms, RF, GB, and CART, in predicting canopy heights, (2) to evaluate the performance of our canopy height maps using reference canopy height from canopy height models (CHMs), and (3) to compare our canopy height maps with other two existing canopy height maps. RF and GB were the top-performing algorithms, achieving the best 13.32% and 16% root mean squared error for broadleaf and coniferous forests, respectively. Validation of the proposed approach revealed that the 100th and 98th percentile, followed by the average of the 75th, 90th, 95th, and 100th percentiles (AVG), were the most accurate GEDI metrics for predicting real canopy heights. Comparisons between predicted and reference CHMs demonstrated accurate predictions for coniferous stands (R-squared = 0.45, RMSE = 29.16%). [ABSTRACT FROM AUTHOR]

Forest ecosystems strongly contribute to the mitigation of climate change impacts through the carbon stored in forests and through harvested wood products, such as sawed wood and furniture, which are obtained from many types of timber assortments. Timber assortments are defined as log sections of specific dimensions (log length and maximum/minimum end diameters), gathered from felled trunks, that have both specific commercial timber utilisation and economic value. However, it is challenging to discriminate and assess timber assortment types, especially within a forest stand before the forest has been harvested. Accurate estimations of timber assortments are a fundamental prerequisite in supporting forest holdings and assisting practitioners in the optimisation of harvesting activities and promoting forest wood chains, in addition to forest policy and planning. Based on the georeferenced points cloud tool, light detection and ranging (LiDAR) is a powerful technology for rapidly and accurately depicting forest structure, even if the use of LiDAR for timber assortments estimation is lacking and poorly explored. This systematic literature review aimed to highlight the state-of-the-art applications of the LiDAR systems (spaceborne; airborne, including unmanned aerial UASs; and terrestrial) to quantify and classify different timber assortment types. A total of 304 peer-reviewed papers were examined. The results highlight a constant increment of published articles using LiDAR systems for forest-related aspects in the period between 2000 and 2021. The most recurring investigation topics in LiDAR studies were forest inventory and forest productivity. No studies were found that used spaceborne LiDAR systems for timber assortment assessments, as these were conditioned by the time and sample size (sample size = ~12 m/~25 m of laser footprint and 0.7 m/60 m of space along the track for ICESat-2, GEDI and time = since 2018). Terrestrial LiDAR systems demonstrated a higher performance in successfully characterising the trees belonging to an understory layer. Combining airborne/UAS systems with terrestrial LiDAR systems is a promising approach to obtain detailed data concerning the timber assortments of large forest covers. Overall, our results reveal that the interest of scientists in using machine and deep learning algorithms for LiDAR processes is steadily increasing. [ABSTRACT FROM AUTHOR]