skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


  • NSF Public Access
  • Search Results
  • Revealing forest structural "fingerprints": An integration of LiDAR and deep learning uncovers topographical influences on Central Amazon forests
Title: Revealing forest structural "fingerprints": An integration of LiDAR and deep learning uncovers topographical influences on Central Amazon forests
Amazon forests are characterized by rich structural diversity. However, the influence of factors such as topography, soil attributes, and external disturbances on structural variability is not always well characterized, and traditional structural metrics may be inadequate to capture this type of complexity. While LiDAR offers expanded structural insights, traditional parameters used in LiDAR analysis, such as mean or maximum canopy height, are not always well directly linked to environmental variables like topography. Emerging approaches merge LiDAR with machine learning to uncover deeper structural complexities. However, work to date may fail to fully utilize the potential of fine-scale LiDAR information. Here we introduce a novel approach, leveraging 2D point cloud images derived from a profiling canopy LiDAR (PCL). The technique targets intricate details within LiDAR point clouds by using deep learning algorithms. With a dataset from the Central Amazon comprising 18 multitemporal transects of 450 m in length, our objective was to detect structural "fingerprints" of varied topographical types along a hillslope, comprising: Riparian, White-sand, and Plateau, and to detect any gradient of structural shifts based on terrain variations here represented by the height above the nearest drainage (HAND). The dataset was trained and tested using a leave-one-group-out approach (LOGO) in which, for each iteration, a complete 450 m multitemporal transect was excluded from training and tested after each iteration. The fast.ai platform and a ResNet-34 architecture, coupled with transfer learning, were used to perform a classification to distinguish between three topographical types. Furthermore, a hybrid model combining a Convolutional Autoencoder, and Partial Least Square (PLS) regression was designed to detect forest structural gradient correlations with HAND variation. Cross-validation achieved a promising high weighted F1 score of 0.83 to classify forests based on the topographical types. Additionally, a combined Convolutional Autoencoder and PLS regression revealed a strong correlation (R2 = 0.76) between actual and predicted HAND. Innovatively combining deep learning with ground-based PCL LiDAR, our study revealed unique Amazon Forest structures connected to topographic variation. Our findings underscore the transformative potential of such integrative approaches for investigating forest dynamics and promise a powerful new tool for understanding climate-related forest structure change.  more » « less
Award ID(s):
1754357 1950080
PAR ID:
10540607
Author(s) / Creator(s):
; ; ; ; ; ; ;
Publisher / Repository:
Elsevier
Date Published:
Journal Name:
Ecological Informatics
Volume:
81
Issue:
C
ISSN:
1574-9541
Page Range / eLocation ID:
102628
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; (, Methods in Ecology and Evolution)
    Record, Sydne (Ed.)
    1. LiDAR data are being increasingly used to provide a detailed characterization of the vertical profile of forests. This characterization enables the generation of new insights on the influence of environmental drivers and anthropogenic disturbances on forest structure as well as on how forest structure influences important ecosystem functions and services. Unfortunately, extracting information from LiDAR data in a way that enables the spatial visualization of forest structure, as well as its temporal changes, is challenging due to the high-dimensionality of these data. 2. We show how the Latent Dirichlet Allocation model applied to LiDAR data (LidarLDA) can be used to identify forest structural types and how the relative abundance of these forest types changes throughout the landscape. The code to fit this model is made available through the open-source R package LidarLDA in github. We illustrate the use of LidarLDA both with simulated data and data from a large-scale fire experiment in the Brazilian Amazon region. 3. Using simulated data, we demonstrate that LidarLDA accurately identifies the number of forest types as well as their spatial distribution and absorptance probabilities. For the empirical data, we found that LidarLDA detects both landscape-level patterns in forest structure as well as the strong interacting effect of fire and forest fragmentation on forest structure based on the experimental fire plots. More specifically, LidarLDA reveals that proximity to forest edge exacerbates the impact of fires, and that burned forests remain structurally different from unburned areas for at least seven years, even when burned only once. Importantly, LidarLDA generates insights on the 3D structure of forest that cannot be obtained using more standard approaches that just focus on top-of-the-canopy information (e.g., canopy height models based on LiDAR data). 4. By enabling the mapping of forest structure and its temporal changes, we believe that LidarLDA will be of broad utility to the ecological research community. 
    more » « less
  2. ; ; ; ; ; (, INPE, Brazil)
    Canopy heights and vertical profiles were analyzed for 12 airborne lidar tracks acquired over forests of the mid-Juruá region, Brazil. Canopy height models were classified at 1m resolution as floodplain, terrace, hillslope, or interfluvial flat; floodplains were further separated according to Horton-Strahler (HS) stream order. RH95 canopy heights, and vertical profiles at 1m intervals, were aggregated to 30m scale and compared with Copernicus DEM heights, using a DEM transform, the Relative Terrain Height (RTH). Median canopy height ranged from 15.4 m for the Juruá floodplain to 25.5 m for hillslopes; maximum canopy heights varied from 37.4 m to 60.0 m. A strong correlation between RTH and median canopy height (r = 0.75) was found for the Juruá floodplain tracks. Vertical profiles of Juruá floodplain tracks showed that the height above ground of maximum returns increased monotonically with RTH height. Our results clearly show the influence of floodplain topography on forest canopy structure. 
    more » « less
  3. ; ; ; ; ; ; ; ; (, Remote Sensing in Ecology and Conservation)
    Abstract Recently classified as a unique species by the IUCN, African forest elephants (Loxodonta cyclotis) are critically endangered due to severe poaching. With limited knowledge about their ecological role due to the dense tropical forests they inhabit in central Africa, it is unclear how the Afrotropics are influenced by elephants. Although their role as seed dispersers is well known, they may also drive large‐scale processes that determine forest structure through the creation of elephant trails and browsing the understory, allowing larger, carbon‐dense trees to succeed. Multiple scales of lidar were collected by NASA in Lopé National Park, Gabon from 2015 to 2022. Utilizing two airborne lidar datasets in an African forest elephant stronghold, detailed canopy structural information was used in conjunction with elephant trail data to determine how forest structure varies on and off trails. Forest along elephant trails displayed different structural characteristics than forested areas off trails, with lower canopy height, canopy cover, and different vertical distribution of plant density. Less plant area density was found on trails at 1 m in height, while more vegetation was found at 12 m, compared to off trail locations. Trails in forest areas with previous logging history had lower plant area in the top of the canopy. Forest elephants can be considered as “logging light” ecosystem engineers, affecting canopy structure through browsing and movement. Both airborne lidar scales were able to capture elephant impact along trails, with the high‐resolution discrete return lidar performing higher than waveform lidar. 
    more » « less
  4. ; ; ; ; ; ; ; ; ; ; et al (, Science of Remote Sensing)
    Accurate measurements of terrain elevation are crucial for many ecological applications. In this study, we sought to assess new global three-dimensional Earth observation data acquired by the spaceborne Light Detection and Ranging (LiDAR) missions Ice, Cloud, and Land Elevation Satellite-2 (ICESat-2) and Global Ecosystem Dynamics Investigation (GEDI). For this, we examined the “ATLAS/ICESat-2 L3A Land and Vegetation Height”, version 5 (20 × 14 m and 100 × 14 m segments) and the “GEDI Level 2A Footprint Elevation and Height Metrics”, version 2 (25 m circle). We conducted our analysis across four land cover classes (bare soil, herbaceous, forest, savanna), and six forest types (temperate broad-leaved, temperate needle-leaved, temperate mixed, tropical upland, tropical floodplain, and tropical secondary forest). For assessment of terrain elevation estimates from spaceborne LiDAR data we used high resolution airborne data. Our results indicate that both LiDAR missions provide accurate terrain elevation estimates across different land cover classes and forest types with mean error less than 1 m, except in tropical forests. However, using a GEDI algorithm with a lower signal end threshold (e.g., algorithm 5) can improve the accuracy of terrain elevation estimates for tropical upland forests. Specific environmental parameters (terrain slope, canopy height and canopy cover) and sensor parameters (GEDI degrade flags, terrain estimation algorithm; ICESat-2 number of terrain photons, terrain uncertainty) can be applied to improve the accuracy of ICESat-2 and GEDI-based terrain estimates. Although the goodness-of-fit statistics from the two spaceborne LiDARs are not directly comparable since they possess different footprint sizes (100 × 14 m segment or 20 × 14 m segment vs. 25 m circle), we observed similar trends on the impact of terrain slope, canopy cover and canopy height for both sensors. Terrain slope strongly impacts the accuracy of both ICESat-2 and GEDI terrain elevation estimates for both forested and non-forested areas. In the case of GEDI the impact of slope is, however, partly caused by horizontal geolocation error. Moreover, dense canopies (i.e., canopy cover higher than 90%) affect the accuracy of spaceborne LiDAR terrain estimates, while canopy height does not, when considering samples over flat terrains. Our analysis of the accuracy and precision of current versions of spaceborne LiDAR products for different vegetation types and environmental conditions provides insights on parameter selection and estimated uncertainty to inform users of these key global datasets. 
    more » « less
  5. ; ; ; ; ; ; (, Remote Sensing)
    Structural diversity is a key feature of forest ecosystems that influences ecosystem functions from local to macroscales. The ability to measure structural diversity in forests with varying ecological composition and management history can improve the understanding of linkages between forest structure and ecosystem functioning. Terrestrial LiDAR has often been used to provide a detailed characterization of structural diversity at local scales, but it is largely unknown whether these same structural features are detectable using aerial LiDAR data that are available across larger spatial scales. We used univariate and multivariate analyses to quantify cross-compatibility of structural diversity metrics from terrestrial versus aerial LiDAR in seven National Ecological Observatory Network sites across the eastern USA. We found strong univariate agreement between terrestrial and aerial LiDAR metrics of canopy height, openness, internal heterogeneity, and leaf area, but found marginal agreement between metrics that described heterogeneity of the outermost layer of the canopy. Terrestrial and aerial LiDAR both demonstrated the ability to distinguish forest sites from structural diversity metrics in multivariate space, but terrestrial LiDAR was able to resolve finer-scale detail within sites. Our findings indicated that aerial LiDAR could be of use in quantifying broad-scale variation in structural diversity across macroscales. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email