DESCRIPTION

The ARNOLD dataset provides real-world LiDAR data collected across diverse marine environments—including ports, marinas, and open waters—and features both static and dynamic objects. It includes raw point clouds and annotated targets in four categories: quay, motor boat, sailing boat, and ship. The dataset is designed for the training and testing of perception algorithms focused on detection, classification, and tracking.

Parsing tools are provided to facilitate easy integration into existing workflows. Detailed information about the dataset composition and usage can be found in the original paper and in the README file included with the dataset.

HOW TO CITE
Please do not cite this dataset via Zenodo. Instead, refer to the original journal article in which the dataset is presented and described:

Martelli, M., Faggioni, N., & Ponzini, F. (2025). ARNOLD – Annotated Repository of Navigational Obstacles from LiDAR Data. Autonomous Transportation Research.

ADDITIONAL MATERIAL

For an analysis of the obstacle features of each class and a possible application of class feature extraction for classification purposes using a Random Forest Classifier, please refer to:

Ponzini, F., Zaccone, R., & Martelli, M. (2025). LiDAR target detection and classification for ship situational awareness: A hybrid learning approach. Applied Ocean Research, 158, 104552. DOI: https://doi.org/10.1016/j.apor.2025.104552" target="_blank" rel="noreferrer noopener">https://doi.org/10.1016/j.apor.2025.104552

Abstract

LiDAR (Light Detection and Ranging) data was collected for the Geological and Bioregional Assessment Program by Fugro Australia Land Pty Ltd over two mobilisations. The data was acquired at an average density of 1 point per square metre, processed and compiled as LiDAR Classified Data in LAS 1 km tiles and 1 m grid DEM in ESRI ascii 1 km tiles. The total area of survey is 31,780 km². The data was used to develop a Digital Elevation Model and to determine bathymetry for a two-dimensional hydrodynamic flood inundation model of the Cooper Creek floodplain.

Attribution

Geological and Bioregional Assessment Program

History

LiDAR data was acquired by Fugro Australia Land Pty Ltd over two mobilisations. The first mobilisation was flown from 20 March - 4 April 2019 with an aircraft equipped with a Riegl LMS Q780 LiDAR system. The second mobilisation was flown from 6 - 17 October 2019 with an aircraft equipped with a Riegl LMS Q1560 LiDAR system. Nominal flying height was 2800 mAGL with a swath width of 3200 m.\r Processing steps:\r • Project Planning\r • Aircrew briefing\r • Acquisition as per requirements of dry Season\r • Field data QA for integrity and completeness\r • Downloading raw data\r • QA Raw data\r • GNSS-IMU Coupled Solution\r • Riprocess processing\r • Geodetic Validation Flight Line matching\r • Ground Filter\r • LiDAR data Classification\r • Product Generation LiDAR Derivates\r • QA of data\r • Delivery of Data and reports

This study evaluates the integration of airborne LiDAR (Light Detection and Ranging) data with British Columbia’s Vegetation Resource Inventory (VRI) to identify opportunities for enhancement. LiDAR-derived forest structure metrics were compared against VRI estimates in the Resort Municipality of Whistler, revealing discrepancies and variable correlation strengths. Tree height showed a moderately strong correlation (r ≈ 0.66, p < 0.001) between VRI-projected and LiDAR-measured values. LiDAR heights were on average ~0.86 m lower than VRI estimates, through sharing identical patterns and trends, indicating agreement in vertical structure. In contrast, canopy closure (percent cover) has only a weak correlation (r ≈ 0.27). LiDAR-derived canopy closure values were substantially higher, about 55.8% on average than VRI projections, suggesting that our LiDAR method overestimated due to methodological differences. Conifer vs. broadleaf compositions showed that LiDAR morphological metrics can partly predict species composition: percent conifer in VRI had a moderate negative correlation with LiDAR-derived canopy cover (r ≈ –0.41) and moderate positive correlation with tree shape aspect ratio (r ≈ 0.40) though showing some correlation this is too modest for any conclusive predictive power. Our methodology combined terrain-normalized LiDAR point clouds, individual tree segmentation, and extraction of metrics such as height, crown area, aspect ratio, and crown surface curvature. The results demonstrate that LiDAR can augment VRI by providing more objective measurements, higher spatial resolution, and improved estimates of key forest attributes. These enhancements could benefit applications including forest management, wildfire risk assessment, and carbon accounting.

Seabed landform features were classified from the New South Wales statewide marine lidar dataset, acquired in 2018 by Fugro Pty Ltd on behalf of the Department of Planning and Environment (data available for download on SEED, see below). Seabed features were extracted from the marine lidar data and classified into seabed landform classes. Classified landform features include reefs, plains, peaks, scarps, depressions and channels. These landforms capture variation in the shape and structure of reef outcrops along the NSW coastal and nearshore environment. \r Features were classified using the Seabed Landforms Classification Toolset developed for ArcGIS by the Coastal and Marine Unit, DPE (Linklater et al. 2023) which are publicly available on SEED (https://datasets.seed.nsw.gov.au/dataset/seabed-landforms-classification-toolset) and GitHub (https://github.com/LinklaterM/Seabed-Landforms-Classification-Toolset/). \r \r The statewide dataset is provided as ArcGIS shapefiles divided into 9 segments along the coast. The data covers 4060 km2, extending from the coastline (0 m AHD) to a maximum of 50 m depth, reaching an average depth of 35 m. Data coverage extends a maximum distance of 9 km offshore, with coverage extending on average 3 km offshore.\r \r This dataset provides an understanding of the extent and distribution of submerged reefs along the NSW coast, which contributes fundamental baseline information for managers, users and custodians of the marine environment. \r \r This dataset was funded by the Marine Estate Management Authority and NSW Climate Change Fund through the Coastal Management Funding Package. \r \r Please cite this dataset as: Linklater, M., Morris, B., Kinsela, M., Ingleton, T. and Hanslow, D. (2022), Exploring patterns of reef distribution along the southeast Australian coast using marine lidar data. Manuscript in preparation. \r \r NSW statewide marine lidar data – available for download on SEED: https://datasets.seed.nsw.gov.au/dataset/marine-lidar-topo-bathy-2018\r \r Linklater, M., Morris, B.D. and Hanslow, D.J. (2023), Classification of seabed landforms on continental and island shelves. Frontiers in Marine Science, 10, https://www.frontiersin.org/articles/10.3389/fmars.2023.1258556/full.\r \r Linklater, M., Ingleton, T. C., Kinsela, M. A., Morris, B. D., Allen, K. M., Sutherland, M. D., & Hanslow, D. J. 2019. Techniques for classifying seabed morphology and composition on a subtropical-temperate continental shelf. Geosciences, 9(3), 141.

Standing dead trees (known as snags) are historically difficult to map and model using airborne laser scanning (ALS), or lidar. Specific snag characteristics are important for wildlife; for instance, a larger snag with a broken top can serve as a nesting platform for raptors. The objective of this study was to evaluate whether characteristics such as top intactness could be inferred from discrete-return ALS data. We collected structural information for 198 snags in closed-canopy conifer forest plots in Idaho. We selected 13 lidar metrics within 5 m diameter point clouds to serve as predictor variables in random forest (RF) models to classify snags into four groups by size (small [<40 cm diameter] or large [≥40 cm diameter]) and intactness (intact or broken top) across multiple iterations. We conducted these models first with all snags combined, and then ran the same models with only small or large snags. Overall accuracies were highest in RF models with large snags only (77%), but ka...

R script for reading hdf5 files

This teaching data subset contains1. a subset of spatial data (gis layers for the California Madera County and NEON SOAP and SJER sites). 2. Some other general spatial boundary layers from natural earth3. NEON lidar data and insitu measurements for SOAP and SJER sites. The data are used in both the Earth Analytics R and python courses. The Lidar data can be used to teach uncertainty given there are ground measurements available. We have recently added an additional vector layer so that cropping raster data can be taught using this data set as well.

Data and code used in Lidar-derived eigenfeatures improve species classification using deep-learning in Southwestern mixed conifer forests. Lidar data is provided for UAV-LS and ALS in lidar_data.zip. Las files are named by site and plot (e.g., 3-2). Field data, shapefiles, and code is provided in data.zip. R code is provided for tree segmentation, tree matching, image creation, and neural network cross validation.

This data set contains terrestrial LIDAR survey (TLS) point cloud data collected at Grand Mesa, Colorado as part of the 2017 SnowEx campaign. Data were collected in the fall (September and October) and winter (February) seasons. Each point contains X, Y, and Z coordinates (Easting, Northing, and Elevation), along with ancillary information, such as intensity (i) and color (R,G,B), where available. This is unprocessed data which has not been classified by land use (e.g. bare earth, low vegetation, trees).

Description: The work was carried out in the oil palm plantations within the Stability of Altered Forest Ecosystem (SAFE) Project, located within lowland dipterocarp forest regions of East Sabah in Malaysian Borneo. Airborne LiDAR data were acquired on 5 November 2014 using a Leica LiDAR50-II flown at 1850 m altitude on a Dornier 228-201 travelling at 135 knots. The LiDAR sensor emitted pulses at 83.1 Hz with a field of view of 12.0°, and a footprint of about 40 cm diameter. The average pulse density was 7.3/m2. The Leica LiDAR50-II sensor records full waveform LiDAR, but for the purposes of this study the data were discretised, with up to four returns recorded per pulse. The LiDAR data were pre-processed by NERC's Data Analysis Node and delivered in standard LAS format. All further processing was undertaken using LAStools (Rapidlasso GmbH, LAStools). Points were classified as ground and non-ground, and a digital elevation model (DEM) was fitted to the ground returns, producing a raster of 1 m resolution. The DEM elevations were subtracted from elevations of all non-ground returns to produce a normalised point cloud, and a canopy height model (CHM) was constructed from this on a 0.5 m raster by averaging the first returns. Finally, holes in the raster were filled by averaging neighbouring cells. Project: This dataset was collected as part of the following SAFE research project: Influences of disturbance and environmental variation on biomass change in Malaysian Borneo Funding: These data were collected as part of research funded by:

NERC (Standard grant, JKM/MBS.1000-2/2 JLD.3 (128)) This dataset is released under the CC-BY 4.0 licence, requiring that you cite the dataset in any outputs, but has the additional condition that you acknowledge the contribution of these funders in any outputs.

Permits: These data were collected under permit from the following authorities:

Sabah Biodiversity Centre (Research licence JKM/MBS.1000-2/2 JLD.3 (128))

XML metadata: GEMINI compliant metadata for this dataset is available here Files: This consists of 1 file: LiDAR_Aboveground_Carbon.xlsx LiDAR_Aboveground_Carbon.xlsx This file contains dataset metadata and 1 data tables:

LiDAR aboveground carbon in Oil palm plantations (described in worksheet LiDAR_ Aboveground_Carbon) Description: The output of a LiDAR50-II sensor records full waveform LiDAR, but for the purposes of this study the data were discretised, with up to four returns recorded per pulse. The LiDAR data was pre-processed by NERC's Data Analysis Node and delivered in standard LAS format. All further processing was undertaken using LAStools (Rapidlasso GmbH, LAStools). Points were classified as ground and non-ground, and a digital elevation model (DEM) was fitted to the ground returns, producing a raster of 1 m resolution. The DEM elevations were subtracted from elevations of all non-ground returns to produce a normalised point cloud, and a canopy height model (CHM) was constructed from this on a 0.5 m raster by averaging the first returns. Finally, holes in the raster were filled by averaging neighbouring cells. Number of fields: 35 Number of data rows: 27 Fields:

Year: Year the oil palm trees were planted (Field type: Numeric) Plot: Plot number based on the SAFE project framework. Each plot is 25 metres x 25 metres size or 0.0625 hectares (Field type: Location) meanH: Average tree height per plot (Field type: Numeric) TreeN_plot: Number of trees per plot (Field type: Numeric) TreeN_ha: Number of trees per hectare obtained by upscaling the number of trees within each 25m x 25m (0.0625 ha) plot to 1 ha (Field type: Numeric) ACD_plot: Sum of the aboveground carbon density per plot (Field type: Numeric) ACD_ha: Sum of the aboveground carbon density per hectare obtained by upscaling the number of aboveground carbon density within each 25m x 25m (0.0625 ha) plot to 1 ha (Field type: Numeric) CC1: Average canopy cover per hectare: the proportion of area occupied by crowns at 1 metre height (Field type: Numeric) CC2: Average canopy cover per hectare: the proportion of area occupied by crowns at 2 metres height (Field type: Numeric) CC3: Average canopy cover per hectare: the proportion of area occupied by crowns at 3 metres height (Field type: Numeric) CC4: Average canopy cover per hectare: the proportion of area occupied by crowns at 4 metres height (Field type: Numeric) CC5: Average canopy cover per hectare: the proportion of area occupied by crowns at 5 metres height (Field type: Numeric) CC6: Average canopy cover per hectare: the proportion of area occupied by crowns at 6 metres height (Field type: Numeric) CC7: Average canopy cover per hectare: the proportion of area occupied by crowns at 7 metres height (Field type: Numeric) CC8: Average canopy cover per hectare: the proportion of area occupied by crowns at 8 metres height (Field type: Numeric) CC9: Average canopy cover per hectare: the proportion of area occupied by crowns at 9 metres height (Field type: Numeric) CC10: Average canopy cover per hectare: the proportion of area occupied by crowns at 10 metres height (Field type: Numeric) CC11: Average canopy cover per hectare: the proportion of area occupied by crowns at 11 metres height (Field type: Numeric) CC12: Average canopy cover per hectare: the proportion of area occupied by crowns at 12 metres height (Field type: Numeric) CC13: Average canopy cover per hectare: the proportion of area occupied by crowns at 13 metres height (Field type: Numeric) CC14: Average canopy cover per hectare: the proportion of area occupied by crowns at 14 metres height (Field type: Numeric) CC15: Average canopy cover per hectare: the proportion of area occupied by crowns at 15 metres height (Field type: Numeric) CC16: Average canopy cover per hectare: the proportion of area occupied by crowns at 16 metres height (Field type: Numeric) CC17: Average canopy cover per hectare: the proportion of area occupied by crowns at 17 metres height (Field type: Numeric) CC18: Average canopy cover per hectare: the proportion of area occupied by crowns at 18 metres height (Field type: Numeric) CC19: Average canopy cover per hectare: the proportion of area occupied by crowns at 19 metres height (Field type: Numeric) CC20: Average canopy cover per hectare: the proportion of area occupied by crowns at 20 metres height (Field type: Numeric) CC21: Average canopy cover per hectare: the proportion of area occupied by crowns at 21 metres height (Field type: Numeric) CC22: Average canopy cover per hectare: the proportion of area occupied by crowns at 22 metres height (Field type: Numeric) CC23: Average canopy cover per hectare: the proportion of area occupied by crowns at 23 metres height (Field type: Numeric) TCH: Top of canopy height: mean height of Canopy Height Model (CHM) pixels per hectare. (Field type: Numeric) TreeN_itc: Number of segmented trees per hectare obtained by using the itcSegment function implemented in R (Field type: Numeric) meanH_itc: Average tree height per hectare obtained by using the itcSegment function inmplement in R (Field type: Numeric) meanHc_itc: Corrected average tree height per hectare obtained by using the itcSegment function inmplement in R (Field type: Numeric) ACDc_itc: Sum of the aboveground carbon density per hectare obtained by using the itcSegment function inmplement in R (Field type: Numeric) Date range: 2014-11-05 to 2014-11-05 Latitudinal extent: 4.5000 to 5.0700 Longitudinal extent: 116.7500 to 117.8200

This data repository contains a set of multi-temporal data products of ecosystem structure derived from four national ALS surveys of the Netherlands (AHN1–AHN4) (folders: 1_AHN1, 2_AHN2, 3_AHN3, and 4_AHN4). Four sets of 25 LiDAR-derived vegetation metrics representing ecosystem height, cover, and structural variability are provided at 10 m spatial resolution, providing valuable data sources for a wide range of ecological research and field beyond. A preview of all generated LiDAR metrics are also provided (folder: 5_Maps). All 25 LiDAR metrics were calculated using Laserfarm workflow (https://laserfarm.readthedocs.io/en/latest/) (building on the user-extendable features from the “Laserchicken” software: https://laserchicken.readthedocs.io/en/latest/#features). All metrics are calculated with the normalized point cloud. More details on metric calculation are provided on GitHub (Laserchicken: https://github.com/eEcoLiDAR/laserchicken and Laserfarm: https://github.com/eEcoLiDAR/Laserfarm), as well as on the “Laserchicken” documentation page (https://laserchicken.readthedocs.io/en/latest/). We also provided masks to minimize the influence of water surfaces, buildings and roads, powerlines and NA values in the data products (folder: 6_Masks). To supplement the generated data products, we also provided a set of raster layers that contains point/pulse density of each AHN survey and the DTM and DSM raster layers for each AHN dataset (folder: 7_Auxiliary_data). To test the robustness of the LiDAR metrics, we also compared the metrics generated from different pulse densities across different habitat types (folder: 8_Sensitivity_analysis). Two use cases demonstrated the utility of the presented data products: (use case 1) monitoring forest structural change across time using multi-temporal ALS data and (use case 2) comparison of vegetation structural difference within Natura 2000 sites. The used data are also provided (folder: 9_Use_case). Note that all the raster layers are provided at 10 m resolution under the local Dutch coordinate system “RD_new” (EPSG: 28992, NAP:5709). To gain more insights of the pre-classification accuracy of the AHN datasets, we also conducted a preliminary assessment of the effect of terrain filtering on vegetation change detection across AHN datasets (i.e. AHN2–AHN4). The data used in this analysis are made available (folder: 10_Ground_classification).

An overview of all the folders in the repository:

1. AHN1

2. AHN2

3. AHN3

4. AHN4

5. Maps

Those folders contain four sets of 25 LiDAR metrics at 10 m resolution generated from each AHN dataset. The file names and their corresponding LiDAR metrics can be found in Table 1. An additional folder (5_Maps) contains the maps (.pdf format) of all 25 metrics for each AHN dataset.

6. Masks

• ahn3_10m_mask_building_road_water.tif
• ahn4_10m_mask_building_road_water.tif
• ahn4_10m_mask_powerline.tif
• ahn1_10m_NA_mask.tif
• ahn2_10m_NA_mask.tif
• ahn3_10m_NA_mask.tif
• ahn4_10m_NA_mask.tif

It contains two mask layers of water surfaces, buildings and roads for both AHN3 and AHN4 data products based on the Dutch cadaster data (TOP10NL) from 2018 (corresponding to AHN3) and 2021 (corresponding to AHN4) (https://www.kadaster.nl/zakelijk/producten/geo-informatie/topnl). In the masks, water surfaces, buildings and roads were merged into one class with pixel value assigned to 1 and the rest has the pixel value of 0. There is also a powerline mask generated from the AHN4 dataset at 10 m resolution, where pixels containing powerlines were assigned a value of 1 and the rest as NoData. We provide those masks to minimize the inaccuracies of the data products caused by human infrastructures and water surfaces. We also provided a mask for each AHN dataset where NA value occurs — areas with no vegetation points (“unclassified” class in the AHN datasets). Pixels with NA value were assigned with a value of 1 and the rest as 0.

7. Auxiliary data

(1) Point_density

• ahn1_10m_point_density.tif
• ahn2_10m_point_density.tif
• ahn3_10m_point_density.tif
• ahn4_10m_point_density.tif

(2) Pulse_density

• ahn3_10m_pulse_density.tif
• ahn4_10m_pulse_density.tif

(3) Flighttime

• ahn3_10m_flighttime.tif
• ahn4_10m_flighttime.tif

(4) DTM_DSM

• ahn2_10m_dtm.tif
• ahn2_10m_dsm.tif
• ahn3_10m_dtm.tif
• ahn3_10m_dsm.tif
• ahn4_10m_dtm.tif
• ahn4_10m_dsm.tif

It contains four raster layers representing the point density of each AHN dataset, two raster layers for pulse density of the AHN3 and AHN4, two raster layers for flight timestamp of the AHN3 and AHN4, and six DTM and DSM layers for AHN2–AHN4. All raster layers are provide at 10 m resolution.

8. Sensitivity analysis

• Dunes
• Marsh
• Grassland
• Shrubland
• Woodland
• Code
• Figure

It contains the 25 metrics generated from point clouds with the original and down-sampled pulse densities (original pulse density of the AHN4, pulse density of the AHN3, ½ of the pulse density of the AHN3, and ¼ of the pulse density of AHN3) for each habitat type (i.e. dunes, marsh, grassland, shrubland, and woodland). We also provided the code and the figures generated from this analysis.

9. Use_case

(1) Multi-temporal_AHN

• Data
• Usecase_multi-temporal_AHN.R

It contains the input data for the use case data processing (i.e. Data folder), including the shapefile of the area (i.e. shp folder), and extracted pixel value from six selected LiDAR metrics from AHN1–AHN5 (i.e. Metrics folder), and the selected LiDAR metrics of the area (e.g. Hp95 folder), and the R code for data processing (i.e. Usecase_multi-temporal_AHN.R).

(2) Natura2000

• Data
• Natura2000_end2021_HABITATCLASS.csv
• Natura2000_NL_habitat_grouped.csv
• Usecase_Natura2000.R

It contains a folder of the input data used for the use case (i.e. Data folder), including the shapefile (i.e. shp folder) of the Natura 2000 sites in the Netherlands (i.e. Nature2000_NL_RDnew.shp) and the 100 random sample plots from each habitat type (e.g. woodland_points.shp), and the LiDAR metrics from AHN4 used for demonstrating the vegetation structure within each habitat type (i.e. AHN4_metrics folder). The table “Natura2000_end2021_HABITATCLASS.csv” is the original attribute table of Natura 2000 sites, including information related to the description of habitat classes (column “DESCRIPTION”), the code corresponding to the habitat class (column “HABITATCODE”), the code for the specific site (column “SITECODE”), and the percentage of the cover of a specific habitat class in one site (column “PERCENTAGECOVER”). The table “Natura2000_NL_habitat_grouped.csv” contains two subtabs, one (i.e. “Habitatclass”) is the copy of the original attribute table of Natura 2000 sites in the Netherlands, and the other one (i.e. “Habitat_class_summary”) is the grouped habitat type based on the dominant habitat class (i.e. class with the highest percentage cover) in each site. Different colors indicate different habitat types, corresponding to the colors in the first tab (“Habitatclass”) where

Canopy structural complexity metrics calculated from sites across the eastern US from portable canopy lidar (PCL) data using the R package forestr (https://cran.r-project.org/web/packages/forestr/index.html)

Note: To better find the files to download, select "Change View: Tree". This dataset is associated with the paper "TreeLearn: A deep learning method for segmenting individual trees from ground-based LiDAR forest point clouds" published in Ecological Informatics and the ML4RS workshop paper "Towards general deep-learning-based tree instance segmentation models" presented at ICLR 2024. It extends the publicly available segmented tree data that was introduced by Calders et al. [1] and Tockner et al. [2]. These two publications only provide segmented trees. For this dataset, these tree labels were propagated to the original point clouds and the remaining points were automatically classified as either "non-tree points" or "unlabeled". Furthermore, some manual correction of the segmented trees was conducted, especially for the tree bases in Tockner et al. [2]. A more comprehensive description of the dataset is given in the linked publications. We provide the laser scans in the original resolution as well as in a voxelized form where the point cloud has been subsampled to contain only one point within a cube with edge length 0.1m. We provide the forest laser scans in the .laz format and follow the same labeling scheme proposed by Puliti et al. [3]. Specifically, a unique identifier is stored as an additional field named "treeID" in the .laz files. Trees are labeled starting from 1 and all non-tree points have the label 0 in the treeID field. The dataset comes with a classification into the three semantic categories "non-tree-points" (label=2), "unlabeled" (label=3) and "tree-points" (label=4) that is saved in the classification field of the .laz file. The .laz format is compatible with popular point cloud processing tools like CloudCompare and can also be loaded in python using the laspy package. Example code for opening .laz files in python as numpy arrays is provided in the open_files.ipynb notebook. References [1] Calders, K., Origo, N., Burt, A., Disney, M., Nightingale, J., Raumonen, P., ... & Lewis, P. (2018). Realistic forest stand reconstruction from terrestrial LiDAR for radiative transfer modelling. Remote Sensing, 10(6), 933. [2] Tockner, A., Gollob, C., Kraßnitzer, R., Ritter, T., & Nothdurft, A. (2022). Automatic tree crown segmentation using dense forest point clouds from Personal Laser Scanning (PLS). International Journal of Applied Earth Observation and Geoinformation, 114, 103025. [3] Puliti, S., Pearse, G., Surový, P., Wallace, L., Hollaus, M., Wielgosz, M., & Astrup, R. (2023). FOR-instance: a UAV laser scanning benchmark dataset for semantic and instance segmentation of individual trees. arXiv preprint arXiv:2309.01279.

This data publication contains five raster datasets detailing the land cover and forest structure of the United States Virgin Islands and British Virgin Islands. There are two spatial land cover datasets, provided in multiple formats, that represent land-cover and woody vegetation formations using multi-part Landsat ETM+ scenes centered around the year 2000. Also included is a digital elevation model (DEM) that was created using discrete lidar data collected in 2004. The DEM was created by filtering ground (minimum elevation) and non-ground (mean and maximum elevation) returns to create a bare earth DEM and forest height estimates for the islands of St. John and St. Thomas, United States Virgin Islands. Two additional raster datasets include models that displays the spatial representation of lidar derived (1) dominant co-dominant tree height estimates and (2) above ground biomass estimates for the islands of St. John and St. Thomas, United States Virgin Islands.1. The landcover datasets represent a year 2000 classification for the United States and British Virgin Islands as part of a Caribbean wide effort to develop detailed landuse/landcover products for the Caribbean region.

1. The lidar DEM was created to provide the USDA Forest Service, International Institute of Tropical Forestry with spatial representation of lidar derived surface elevation estimates for the islands of St. John and St. Thomas, United States Virgin Islands. St. John and St. Thomas were selected as the lidar study area based on the availability of lidar data coverage.

3 & 4. The models were created to provide the USDA Forest Service, International Institute of Tropical Forestry with spatial representation of lidar derived dominant co-dominant tree height estimates and above ground biomass estimates for the islands of St. John and St. Thomas, United States Virgin Islands.

St. John and St. Thomas were selected as the lidar study area based on the availability of lidar data coverage. The island of St. John (18°22'N, 64°40'W) and the island of St. Thomas (18°21'N, 64°55') are approximately 5,000 and 7,200 hectares in area, respectively, and consist of mountainous topography with elevations ranging from sea level to 387 meters on St. John and 471 meters on St. Thomas. The woody vegetation on both islands is similar to other islands in the Virgin Islands and includes both late and early stage successional forests.These datasets were completed as part of the fullfillment of a Master's of Science degree by Todd Kennaway within the Department of Forest Rangeland and Watershed Stewardship (FRWS) under the College of Natural Resources at Colorado State University.

For more information about these data see Kennaway et al. (2008).

Training dataset in hdf5 file format including the class labels. Scripts are provided for reading in the dataset in R, Python, and Matlab.

Testing dataset in hdf5 file format without the class labels. Scripts are provided for reading in the dataset in R, Python, and Matlab.

As urban forest provides ecological, social, and economic values to the residents, forest inventory can monitor forest health. Based on the land classification map, the campus planning team pays attention to tree health in the public green space of the University of British Columbia (UBC) campus in Vancouver, Canada. Working together, the forest inventory and land classification map are the priorities of urban planning and forest health in UBC. In order to solve the knowledge gap of no current inventory and land classification map on campus, this study aimed to update the UBC tree inventory and land classification map. R algorithms extracted individual trees’ parameters and metrics like tree height and crown area using Light Detection and Ranging (LiDAR) data 2018 by the City of Vancouver. The author applied random forest classification to determine the tree species (coniferous/deciduous) with the metrics. Four major land cover types were classified by the supervised classification scheme using the UBC orthophoto 2020. The results show that there are 14165 trees (crown diameter more than 4 m) on campus, and the height estimation by the LiDAR method had an overall accuracy of 80% comparing to the field data. The campus’s total vegetation cover was 44% that is higher than the cities in Great Vancouver. The land classification map shows that most of the vegetation cover is on the southern campus. Considering the campus’s topography, coniferous trees on the southwest campus provided potential ecological implications of water retention and preventing soil erosion. The study provided the basis for future studies of trees, vegetation, and UBC Vancouver Campus land planning.

Processed LiDAR data and environmental covariates from 2015 and 2019 LiDAR scans in the Vicinity of Snodgrass Mountain (Western Colorado, USA), in a geographic subset used in primary analysis for the research paper. This package contains LiDAR-derived canopy height maps for 2015 and 2019, crown polygons derived from the height maps using a segmentation algorithm, and environmental covariates supporting the model of forest growth. Source datasets include August 2015 and August 2019 discrete-return LiDAR point clouds collected by Quantum Geospatial for terrain mapping purposes on behalf of the Colorado Hazard Mapping Program and the Colorado Water Conservation Board. Both datasets adhere to the USGS QL2 quality standard. The point cloud data were processed using the R package lidR to generate a canopy height model representing maximum vegetation height above the ground surface, using a pit-free algorithm. This dataset was compiled to assess how spatial patterns of tree growth in montane and subalpine forests are influenced by water and energy availability. Understanding these growth patterns can provide insight into forest dynamics in the Southern Rocky Mountains under changing climatic conditions. This dataset contains .tif, .csv, and .txt files. This dataset additionally includes a file-level metadata (flmd.csv) file that lists each file contained in the dataset with associated metadata; and a data dictionary (dd.csv) file that contains column/row headers used throughout the files along with a definition, units, and data type.

These data were collected during 2016 and 207 using portable canopy lidar and a Decagon handheld ceptometer. CSC data include measures of canopy rugosity, top rugosity, porosity, vegetation area index, mean leaf height, mean outer canopy height, max canopy height derived using the forestr R package (version 1.0.0). Light data include LAI and fPAR. Sites included are Arnot Forest, University of Virginia, Rice Rivers Center, and NEON tower plots at Harvard Forest, Mountain Lake Biological Station, Smithsonian Conservation Biology Institute, Smithsonian Environmental Center, University of Notre Dame Environmental Research Center, Talladega National Forest, Ordway-Swisher Biological Station

Remotely sensed topographic (elevation) and bathymetric (depth) information were acquired for the NSW coast (Point Danger to Cape Howe) and southern Queensland (Palm Beach to Point Danger) using Airborne LiDAR Bathymetry (ALB - a combination of Light Detection And Ranging (LiDAR) and Laser Airborne Depth Sounding (LADS) sensors) during July – December 2018. Data were acquired by Fugro Pty Ltd on behalf of NSW Office of Environment and Heritage using a Riegl VQ-820-G ALB (LiDAR) and Fugro LADS High-Definition sensors aboard sub-contracted Corporate Air Cessna C441 (VH-VEH). Funding was provided through the NSW Coastal Reforms package. The objective of the project was to provide high-resolution data better than 3-5 m spaced soundings (0.5 m spot spacing terrestrial; 3.4 m spot spacing marine) from the mean high-water mark to ~200m inland, and from the shore, seaward (LADS - bathymetry) to the point of laser extinction (~20-40m water depth depending on in-water conditions). Positioning data were collected on the ellipsoid ITRF 2014 GRS80 in UTM Z56 and post-processed using local base stations (CORSnet NSW) to provide a Post Processed Kinematic GNSS solution for final aircraft trajectory before being applied to all data. The final data Geotif products are provided on the * Geosciences Australia ELVIS website .They are combined gridded terrestrial (elevation) and subtidal marine (bathymetry) data at 5 x 5 m (horizontal resolution) Geotifs exported using ESRI ArcMap from rasters (weighted average of clean soundings) in GDA 2020 (horizontal datum) to Australian Height Datum (vertical datum) and vertical precision to International Hydrographic Order (IHO) 1B. Data covers an area of 6862 km2 provided in 48 sub-datasets the extents of which are generally defined in their alongshore extent by the boundaries of NSW Secondary Sediment Compartments (Geosciences Australia). Other data outputs will include raw and classified LAS format files, aerial imagery and raw seabed reflectance data to be made available shortly on the ELVIS website. Data packages containing Arc Grids (topo-bathy, contours), XYZ, KMZ, tif, pdf maps and Fledermaus SD files will be made publicly available via the AODN (* Australian Ocean Data network ).\r Metadata, data quality statements and a geographical data coverage ArcGIS shapefile are available via SEED . The data are intended to inform coastal and marine management and should not be used for navigation without additional processing.