Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
419
datasets available to search
ShareScore release 0.9.0
Dataset results
419 results for “airborn”
AOP01 Correspondence between plant traits and NEON Airborne Observatory Platform (AOP) data at Konza Prairie (2017)
Understanding spatial and temporal variation in plant traits is needed to accurately predict how communities and ecosystems will respond to global change. The National Observatory Ecological Network (NEON) Airborne Observation Platform (AOP) provides hyperspectral images and associated data products at numerous field sites at 1 m spatial resolution, allowing high-resolution trait mapping. However, the reliability of these data depend on establishing rigorous links with in-situ field measurements. We tested the accuracy of NEON’s readily available AOP derived data products – Leaf Area Index, Total biomass, Ecosystem structure (Canopy height model; CHM), and Canopy Nitrogen by comparing them to spatially extensive field measurements from a mesic tallgrass prairie. Correlations with AOP data products exhibited generally weak or no relationships with corresponding field measurements. The weakest relationships were between AOP Canopy Nitrogen and ground-based measures of Nitrogen, as well as the CHM and ground-based canopy height measurements. We also examined how well the full reflectance spectra (380-2500 nm), as opposed to derived products, could predict vegetation traits using partial least-squares regression models. Only one of the eight traits examined, Nitrogen, had an R2 of more than 0.25. For all vegetation traits, R2 ranged from 0.08-0.29 and the root mean square error of prediction ranged from 14-64%. Our results suggest that currently available AOP derived data products are unreliable, at least at this grassland site, and should not be used without extensive ground-based validation. Relationships using the full reflectance spectra may be more promising, although additional assessment of varying spatial scales of field and AOP data, as well as corrections and data pre-processing to improve data quality, are recommended. Finally, grassland sites may be especially challenging for airborne spectroscopy because of their high species diversity within a small area,
AIRBORNE SPECTROMETER MEASUREMENTS FMOM BOREAL SNOW-COVERED LANDSCAPE
<p>The dataset contains 10 meter resolution reflectance data from boreal snow-covered landscape. The purpose of the airborne measurements was to investigate the effect of forest canopy on optical remote sensing signals for snow-covered surfaces. The hyperspectral airborne data was acquired with an AisaDUAL imaging spectrometer on March 18 and on March 21, 2010 in Sodankylä, Finland. The image swath was 240 meters and flight lines were several kilometers long. The original spatial resolution of the data is 80 cm x 80 cm, but it was resampled to pixel size of 10 m x 10 m. All measurements were carried out in non-cloudy conditions (0/8 to 2/8 cloud cover). On 18 March, the tree canopy was snow-free and snow on the ground was several days old, while on 21 March, the tree canopy was snow-covered and snow on ground was fresh. The data contains mosaics of the flight lines for the bands 555 nm, 645 nm, 858.5 nm and 1640 nm for both days 18 March 2010 and 21 March 2010.</p>
AIRBORNE SPECTROMETER MEASUREMENTS FROM BOREAL AND TUNDRA SITE DURING SPRING SNOW MELT
<p>The dataset contains 10 meter resolution reflectance data from boreal and tundra sites during spring snow melt. The purpose of the airborne measurements was to investigate the effect of forest canopy and snow melting on optical remote sensing signals at the very end of melting period. The hyperspectral airborne data was acquired with an AisaDUAL imaging spectrometer on 5 May 2011 in Sodankylä and in Saariselkä, Finland. Saariselkä is a fell region and partly represents open tundra. The image swath was 240 meters and flight lines were several kilometers long. The original spatial resolution of the data is 80 cm x 80 cm, but it was resampled to pixel size of 10 m x 10 m. Snow depth was between 0 cm and 30 cm at the Sodankylä site and between 0 cm and 60 cm at the Saariselkä site implying that the spring melt was clearly more advanced in Sodankylä. Additionally, more snow-free pixels were found at Sodankylä than Saariselkä. During the measurements the sky was cloudless in Sodankylä (cloud cover 0/8) and cloudy (cloud cover 7/8) in Saariselkä. The data contains mosaics of the flight lines for the bands 555 nm, 645 nm, 858.5 nm and 1640 nm for both study sites.</p>
Data from paper: "Large-scale variations in the dynamics of Amazon forest canopy gaps from airborne lidar data and opportunities for tree mortality estimates"
<p>Data from the paper:</p> <p>Dalagnol, R. <em>et al.</em> Large-scale variations in the dynamics of Amazon forest canopy gaps from airborne lidar data and opportunities for tree mortality estimates. <em>Sci Rep</em> <strong>11, </strong>1388 (2021). https://doi.org/10.1038/s41598-020-80809-w</p> <p>Link: https://www.nature.com/articles/s41598-020-80809-w</p> <p> </p> <p>This repository contains:</p> <p>1) Data frame with data from static and dynamic gaps used in Figure 2 (Dalagnol_2020_Data_Multitemporal_gaps.csv). Each row is the aggregated measurement at 5-km resolution. The site component referes to the five site studied with multitemporal data. Site order from 1 to 5 is DUC, TAP, FN1, BON and TAL.</p> <p>2) Data frame with data from static gaps and environmental factors used in Table 1, Figure 3, 4, 5 (Dalagnol_2020_Data_Singledate_gaps_Modeling.csv). Each row is the aggregated measurement of one site observed by airborne lidar data.</p> <p>3) Raster file at 5-km resolution with dynamic gap fraction estimates presented in Figure 5 (dynamic_gap_fraction_amazon.tif).</p> <p> </p> <p>If you need anything else, please contact the corresponding author: Ricardo Dalagnol (ricds@hotmail.com).</p>
Extracted trails from airborne laser scanning in the Oostvaardersplassen nature reserve
<p>Ungulates and other mammalian herbivores can create trails in dense vegetation by trampling and browsing. This can affect vegetation structure and results in the fragmentation of closed, high vegetation, with subsequent impacts on biodiversity. Manually mapping trails in the field or from aerial photographs can be challenging and time consuming, especially in inaccessible or difficult to access habitats such as wetlands and if trails occur beneath the canopy. Airborne laser scanning provides an alternative method because it penetrates vegetation canopies and efficiently obtains highly accurate data in the form of dense 3D point clouds. This repository consists of the extracted trails in wetland area of the Oostvaardersplassen nature reserve in the Netherlands using 3D airborne point cloud data (AHN4) and the manually created 50 plots of ground truth in two regions, i.e. grazed only by red deer and grazed by both red deer and geese. </p>
Code for "New land-use-change emissions indicate a declining CO2 airborne fraction"
<p>Data and programming scripts for reproducing the results from the Nature publication titled:</p> <p>"New land-use-change emissions indicate a declining CO2 airborne fraction".</p> <p>Authors: Margreet J. E. van Marle*, Dave van Wees*, Richard A. Houghton, Robert D. Field, Jan Verbesselt, and Guido R. van der Werf<br> * These authors contributed equally.</p> <p>DOI: https://doi.org/10.1038/s41586-021-04376-4</p> <p> </p> <p>This dataset includes the following (All files are preceded by "Marle_et_al_Nature_AirborneFraction_"):</p> <p>- "Datasheet.xlsx": Excel dataset containing all annual and monthly emissions and CO2 time series used for the analysis, and the resulting airborne fraction time series.</p> <p>- "Script.py":<br> BEFORE RUNNING THE SCRIPT: change the 'wdir' variable to the directory containing the provided script and files.<br> NOTE: This script requires the Python module: 'pymannkendall'<br> Python script used for reproducing the results and figures from the paper. The provided Datasheet.xlsx file and the .zip and .npz files are required for this program. In case all these files are found by the script, it should run within several seconds. Successful execution of the script will save Figures 1-4 from the main text and print the data from Table 1. In case script execution takes longer, please check if the .xlsx, .zip and .npz files are correctly present in the assigned 'wdir' directory. Otherwise the script will start recalculating these files, which might take a while (see notes below).</p> <p>- "MC10000_MK_ts_TRENDabs.zip": .zip file containing all results from the Monte-Carlo simulation for trend estimation for Figure 3 (calculated using Python function 'calc_AF_MonteCarlo()'). This .zip file contains multiple .npz files for different emission scenarios and data treatments. This .zip file is managed by the Python script function 'calc_AF_MonteCarlo_filemanager()', there is no need to unzip the file manually. In case the .zip file is not found by the Python script (e.g. because the .zip file was unpacked manually and deleted), the program will start recalculating and save a new .zip file. This can take several minutes dependent on the computer used. Recalculated results could differ very slightly due to the random factor in the Monte-Carlo approach, even though the 10,000 iterations bring this variation to a minimum.</p> <p>- "MC1000_MK_run50x50_TRENDabs.npz": .npz file containing the Monte-Carlo results used for producing Figure 4 (calculated using Python function 'calc_AF_MonteCarlo_ARR()'). In case the .npz file is not found by the Python script (e.g. because it was deleted or not downloaded), the program will start recalculating and save a new file. This can take around 30 hours(!) dependent on the computer used. Recalculated results could differ slightly due to the random factor in the Monte-Carlo approach.</p> <p>- "tol_colors.py": Additional Python module used in script.py, required for producing the colors used in the Main text figures. Source: https://personal.sron.nl/~pault/</p> <p>- Figure files: Figures 1-4 from the Main text saved as .pdf files. Figure 3 is saved as three independent panels. The Figures are also reproduced by script.py if executed successfully.</p> <p> </p>
Hong Kong Annotated Airborne LiDAR Point Clouds
<p>The annotated point clouds were generated to train the weakly supervised semantic segmentation algorithm Semantic Query Network (SQN) to classify point clouds <sup>[1]</sup>. The dataset covers 16 tiles of airborne LiDAR data in an area of 7.2 km2 in Shatin, Hong Kong, China. 11 tiles were used for training, while 5 tiles were used for validation. There are multiple types of construction in the dataset including high-rise residential buildings, low-rise village houses, and large public buildings. Green spaces are mainly composed of wood areas in open spaces (e.g., in parks and hills) and planted trees in residential gardens and nearby roads. Point clouds are classified in ground, buildings, and trees.</p> <p>The LiDAR data is owned by the Hong Kong government. Please visit the Spatial Data Portal, Survey Division, CEDD (https://sdportal.cedd.gov.hk/#/en/) for more details.</p> <p> </p>
Bathymetry beneath the Amery ice shelf, East Antarctica, revealed by airborne gravity
<p>We estimated the seafloor topography beneath the Amery Ice Shelf, East Antarctica, from airborne gravity anomaly through a nonlinear inversion method called simulated annealing. The estimation results provide a view of the seafloor beneath the Amery Ice Shelf, where direct bathymetric observations are rare. The model, 'gravity_estimated_seafloor_topography_beneath_the_Amery_Ice_Shelf.nc', is in NetCDF format which can be read through MATLAB commands "ncdisp" and "ncread". Contents of the model can be found in "contents.txt". The MATLAB program "nc2mat.m" reads the NetCDF ".nc" format model and saves the variables in the model to a MATLAB ".mat" format file.</p>
Gridded depth and accumulation products from dated airborne radar stratigraphy over West Antarctica during the mid-Holocene, v.1.0.0
<p>This dataset comprises of codes (written in MATLAB) and gridded files (exported as GeoTIFF) presented in Bodart et al. (2023; The Crysophere; <a href="https://doi.org/10.5194/tc-2022-199">https://doi.org/10.5194/tc-2022-199</a>). A summary of the key findings from this study is provided as follows:</p> <p>"Using a spatially extensive IRH over Pine Island Glacier, Thwaites Glacier, and Institute and Möller Ice Streams (covering a total of 610 000 km2 or 30% of the WAIS), and a local layer approximation model, we infer mid-Holocene accumulation rates over the slow-flowing parts of these catchments for the past ~4700 years. By comparing our results with modern climate reanalysis models (1979 – 2019) and observational syntheses (1651 – 2010), we estimate that accumulation rates over the Amundsen-Weddell-Ross divide were on average 18% higher during the mid-Holocene than modern rates. However, no significant spatial changes in the accumulation pattern were observed."</p> <p>This dataset contains a series of files (5x .m files, 10x .tif files). The numbering of the figures in the description below refers to the order of the figures in the associated paper.</p> <ul> <li><strong> 5x MATLAB files:</strong> <ul> <li><strong>Calculate_accumulation_rates.m</strong>: calculates accumulation rates for the mid-Holocene-to-present, as well as uncertainties associated with the age and model structural uncertainty;</li> <li><strong>Calculate_D_parameter.m</strong>: calculates the D parameter (and associated L_path, L_H and L_b) to assess the feasability of the LLA over our grid;</li> <li><strong>Calculate_longitudinal_strain_rates.m</strong>: calculates the longitudinal strain rates over our grid from modern ice-flow velocities;</li> <li><strong>Calculate_vertical_strain_rates.m</strong>: calculates vertical strain rates for the mid-Holocene-to-present part of the ice column from accumulation estimates;</li> <li><strong>Resample_IRH_data.m</strong>: Re-samples the along-track IRH data into evenly distributed 500-m points for speeding up the gridding and calculations of accumulation rates;<br> </li> </ul> </li> <li><strong>10x GeoTIFF files:</strong> <ul> <li><strong>Holocene_IRH_depth_Fig2a.tif: </strong>Figure 2a;</li> <li><strong>Holocene_accumulation_rates_Fig3a.tif:</strong> Figure 3a;</li> <li><strong>Difference_Holocene_accumulation_RACMO2_Fig3c.tif:</strong> Figure 3c;</li> <li><strong>Relative_difference_Holocene_accumulation_RACMO2_Fig4.tif:</strong> Figure 4;</li> <li><strong>D_parameter_FigS1d.tif:</strong> Figure S1d;</li> <li><strong>Holocene_vertical_strain_rates_FigS2a.tif: </strong>Figure S2a;</li> <li><strong>Longitudinal_strain_rates_FigS2b.tif:</strong> Figure S2b;</li> <li><strong>Holocene_accumulation_lower_uncertainty_FigS4a.tif:</strong> Figure S4a;</li> <li><strong>Holocene_accumulation_upper_uncertainty_FigS4b.tif:</strong> Figure S4b;</li> <li><strong>Holocene_accumulation_relative_uncertainty_FigS4c.tif: </strong>Figure S4c;</li> </ul> </li> </ul> <p>Please also cite the associated paper when using this dataset.</p> <p>Any questions, please direct them to the corresponding author, Julien Bodart (julien.bodart@ed.ac.uk).</p>
Protocol for a systematic review: association between airborne pollen, intermittent allergic rhinitis and blood pressure
<p>Previous epidemiological studies have found an increased risk of cardiovascular morbidity and mortality following days with heightened pollen exposure and suggested that intermittent allergic rhinitis might be associated with blood pressure. Pollen sensitization and subsequent pollen exposure cause local inflammation and cytokine release in individuals with intermittent allergic rhinitis (pollen allergy). Inflammatory mediators can travel throughout the body, hence providing the physiologic basis by which pollen allergy may lead to systemic inflammation, which is known to be a risk factor for cardiovascular events. However, the findings regarding the potential association between intermittent allergic rhinitis, pollen exposure, and cardiovascular health are not fully conclusive. To date, no systematic review has been published on this topic.</p> <p>This systematic review seeks to answer: Are exposure to airborne pollen and intermittent allergic rhinitis associated with blood pressure? Secondary questions include: (1) Are there personal characteristics (sex, age) which modify a potential association between intermittent allergic rhinitis or pollen exposure with blood pressure and/or hypertension? (2) What research gaps exist in our understanding of how intermittent allergic rhinitis, pollen exposure, and cardiovascular health are interrelated?</p> <p>Published herein are:</p> <ul> <li>Protocol for the systematic review, including the search strategy</li> <li>Supplement 1: PROSPERO registration</li> <li>Supplement 2: Data extraction table</li> <li>Supplement 3: Risk of bias assessment strategy</li> <li>Supplement 4: Risk of bias assessment tool</li> </ul>
Global Airborne Observatory: Hawaiian Islands Live Coral Cover in 2019
<p>An airborne mapping approach combining laser-guided imaging spectroscopy and deep learning models was used to quantify the geographic distribution of live corals to 16 m water depth throughout the eight main Hawaiian Islands. Full metadata and methods are provided in:</p> <p>Asner, G.P., N.R. Vaughn, J. Heckler, D.E. Knapp, C. Balzotti, E. Shafron, R.E. Martin, B.J. Neilson, J.M. Gove. 2020. Large-scale mapping of live corals to guide reef conservation. Proceedings of the National Academy of Sciences. doi:10.1073/pnas.2017628117.</p> <p>Asner, G.P., N.R. Vaughn, J. Heckler. 2020. Global Airborne Observatory: Hawaiian Islands Live Coral Cover in 2019 (Version 3.0) [Data set]. Zenodo. http://doi.org/10.5281/zenodo.4292660</p> <p>Use of our data requires that you cite both of these sources together. In addition, we ask that these data be used to make the world a better place.</p> <p>The data are in standard GeoTIFF file format organized by island.</p> <p>These data files will be updated as further improvements are made.</p>
Global Airborne Observatory: Hawaiian Islands Reef Rugosity 2019+2020
<p><strong>Summary</strong></p> <p>Coral reef rugosity maps were developed by the Global Airborne Observatory (GAO) team at the Center for Global Discovery and Conservation Science at Arizona State University. The maps show high-resolution seafloor rugosity derived from airborne imaging spectroscopy data collected by the GAO in January 2019 and January 2020.</p> <p><strong>Data Use Requirements</strong></p> <p>Use of these data must acknowledge the source its funders as:</p> <p>“The bathymetry and rugosity data maps were created by the Global Airborne Observatory, Center for Global Discovery and Conservation Science, Arizona State University. The project received financial support from the Lenfest Ocean Program, The Battery Foundation, John D. and Catherine T. MacArthur Foundation, Avatar Alliance Foundation, State of Hawaiʻi Division of Aquatic Resources, State of Hawaiʻi Department of Planning, National Oceanic and Atmospheric Administration.”</p> <p>In addition, provide citations to the following two publications on any materials or presentations utilizing the data or products and results derived from the data:</p> <p>Asner, G.P., N.R. Vaughn, C. Balzotti, P.G. Brodrick, and J. Heckler. 2020. High-resolution reef bathymetry and coral habitat complexity from airborne imaging spectroscopy. <em>Remote Sensing</em> 12:310 (doi:10.3390/rs12020310)</p> <p>Asner, G.P., N.R. Vaughn, S.A. Foo, J. Heckler, and R.E. Martin. 2021. Drivers of reef habitat complexity throughout the Main Hawaiian Islands. <em>Frontiers in Marine Science </em>8:631842. (doi: 10.3389/fmars.2021.631842)</p> <p><strong>Map Properties</strong></p> <p>There are two types of map products available as part of this collection: fine rugosity, and coarse rugosity. Except for Hawaii Island, there are three separate map files for each of the Main Hawaiian Islands (Maui, Kahoolawe, Lanai, Molokai, Oahu, Kauai and Niihau). Hawaii Island was large enough that it needed to be split into quarters for manageability, and each of the three maps are available for all quarters (12 maps total). Coordinates for all maps refer to the UTM Coordinate System, Zone 4 North using datum WGS-84, with the exception of those for Hawaii Island which refer to Zone 5 North.</p> <p>The rugosity maps in meters at a 2-meter spatial resolution up to approximately 22 meters in depth, where data quality allowed. To minimize the effect of water properties, these maps are built using a blend of data from both the 2019 and 2020 collection periods. Fine-scale rugosity seeks high frequency changes on the seafloor arising from coral colonies, rocks, and other bottom features that generate local habitat variability, and coarse-scale rugosity is more responsive to variations in larger terrain features resulting from geologic, reef-scale accretion, and subsidence processes. Fine-scale rugosity maps are produced at 2-meter horizontal resolution, but each pixel represents the conditions of a 6-meter square window centered at the given pixel. Similarly, coarse-scale rugosity maps are produced at 6-meter resolution, but each pixel represents the conditions of a 54-meter square window centered at the given pixel. Rugosity values in the maps are unitless and range from 0.0 (low rugosity) to 100.0 (high rugosity).</p> <p><strong>Methods</strong></p> <p>We computed island-wide maps of rugosity at two resolution using a standard planar image rugosity metric on the GAO blended bathymetry maps (Asner, Gregory. P., Vaughn, Nicholas, & Heckler, Joseph. (2020). Global Airborne Observatory: Hawaiian Islands Live Coral Cover in 2019. Zenodo. <a href="http://doi.org/10.5281/zenodo.4292660">http://doi.org/10.5281/zenodo.4292660</a>). Prior to running the algorithm, missing data of less than two pixels in width were filled using an inverse-distance weighted average of the three nearest neighboring pixels. Fine-scale rugosity was computed using a 3 x 3 pixel (6 x 6 meters) moving window on the original 2-meter resolution bathymetric maps. Coarse-scale rugosity was computed by first down-sampling the 2-meter depth maps to 6-meter resolution using a mean filter. The rugosity metric was then computed using a 9 x 9 pixel (54.0 x 54.0 m) moving window on the 6-meter depth maps. The distribution of raw rugosity algorithm output values is extremely skewed and difficult to interpret. Thus, the rugosity maps contain rugosity values that are transformed in such a way that they have an approximate uniform [0,1] distribution. This both reduces the influence of noisy depth pixels and gives a more meaningful scale upon which to interpret the maps.</p>
Global Airborne Observatory: Hawaiian Islands Bathymetry 2019+2020
<p><strong>Summary</strong></p> <p>Bathymetry maps were developed by the Global Airborne Observatory (GAO) team at the Center for Global Discovery and Conservation Science at Arizona State University. The maps show high-resolution benthic depth, derived from airborne imaging spectroscopy data collected by the GAO in January 2019 and January 2020.</p> <p><strong>Data Use Requirements</strong></p> <p>Use of these data must acknowledge the source its funders as:</p> <p>“The bathymetry and rugosity data maps were created by the Global Airborne Observatory, Center for Global Discovery and Conservation Science, Arizona State University. The project received financial support from the Lenfest Ocean Program, The Battery Foundation, John D. and Catherine T. MacArthur Foundation, Avatar Alliance Foundation, State of Hawaiʻi Division of Aquatic Resources, State of Hawaiʻi Department of Planning, National Oceanic and Atmospheric Administration.”</p> <p>In addition, provide citations to the following two publications on any materials or presentations utilizing the data or products and results derived from the data:</p> <p>Asner, G.P., N.R. Vaughn, C. Balzotti, P.G. Brodrick, and J. Heckler. 2020. High-resolution reef bathymetry and coral habitat complexity from airborne imaging spectroscopy. <em>Remote Sensing</em> 12:310 (doi:10.3390/rs12020310)</p> <p>Asner, G.P., N.R. Vaughn, S.A. Foo, J. Heckler, and R.E. Martin. 2021. Drivers of reef habitat complexity throughout the Main Hawaiian Islands. <em>Frontiers in Marine Science </em>8:631842. (doi: 10.3389/fmars.2021.631842)</p> <p><strong>Map Properties</strong></p> <p>There are multiple map products available as part of this collection. Except for Hawaii Island, there are three separate map files for each of the Main Hawaiian Islands (Maui, Kahoolawe, Lanai, Molokai, Oahu, Kauai and Niihau). Hawaii Island was large enough that it needed to be split into quarters for manageability, and each of the three maps are available for all quarters. Coordinates for all maps refer to the UTM Coordinate System, Zone 4 North using datum WGS-84, with the exception of those for Hawaii Island which refer to Zone 5 North.</p> <p>The blended bathymetry maps give modeled depth as a floating-point values in meters at a 2-meter spatial resolution up to approximately 22 meters in depth, where data quality allowed. To minimize the effect of water properties, these maps are built using a blend of data from both the 2019 and 2020 collection periods.</p> <p><strong>Methods</strong></p> <p>GAO spectrometer data for Hawaii were collected in 1.3 km wide flight line strips and the flights were planned such that individual flight lines overlap each other by about 50%, giving at least two passes of coverage per year of collection. Thus, we have two or more passes of data over most of the Hawaiian coastlines. Details of data collection protocols can be found in <strong><em>Asner et al. (2020) and Asner et al. (2021)</em></strong>. To build the blended bathymetry maps, we identified areas with sufficient sunlight and low surface glint for each flight line, and then applied GAO-created a neural network model to derive estimated depth of each pixel in such areas.</p>
Airborne Radar Quality Control with Machine Learning
<p>This repository contains the radar data collected by ELDORA required to train and test the random forest model discussed in "Airborne radar quality control with machine learning" by Alexander DesRosiers and Michael M. Bell at the Colorado State University Department of Atmospheric Science. The model used in the manuscript is also contained in a '.pkl' file. Upon publication, a link to the paper will be provided here. Finer points of the methodology were discussed in the manuscript and the python script (make_radarQC_rf_model.py) is commented to guide users through the process of creating the model.</p>
Data repository of multi-temporal high-resolution data products of ecosystem structure derived from country-wide airborne laser scanning surveys of the Netherlands
<p><span lang="EN-GB">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:<strong> 1_AHN1, 2_AHN2, 3_AHN3, and 4_AHN4</strong>). 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: <strong>5_Maps</strong>). All 25 LiDAR metrics were calculated using Laserfarm workflow (<a href="https://laserfarm.readthedocs.io/en/latest/">https://laserfarm.readthedocs.io/en/latest/</a>) (building on the user-extendable features from the “Laserchicken” software: <a href="https://laserchicken.readthedocs.io/en/latest/#features">https://laserchicken.readthedocs.io/en/latest/#features</a>). All metrics are calculated with the normalized point cloud. More details on metric calculation are provided on GitHub (Laserchicken: <a href="https://github.com/eEcoLiDAR/laserchicken">https://github.com/eEcoLiDAR/laserchicken</a> and Laserfarm: <a href="https://github.com/eEcoLiDAR/Laserfarm">https://github.com/eEcoLiDAR/Laserfarm</a>), as well as on the “Laserchicken” documentation page (<a href="https://laserchicken.readthedocs.io/en/latest/">https://laserchicken.readthedocs.io/en/latest/</a>). We also provided masks to minimize the influence of water surfaces, buildings and roads, powerlines and NA values in the data products (folder: <strong>6_Masks</strong>). 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: <strong>7_Auxiliary_data</strong>). To test the robustness of the LiDAR metrics, we also compared the metrics generated from different pulse densities across different habitat types (folder: <strong>8_Sensitivity_analysis</strong>). 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: <strong>9_Use_case</strong>). 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: <strong>10_Ground_classification</strong>). </span></p> <p><span lang="EN-GB">An overview of all the folders in the repository:</span></p> <p><strong><span lang="EN-GB">1. </span></strong><strong><span lang="EN-GB">AHN1</span></strong></p> <p><strong><span lang="EN-GB">2. </span></strong><strong><span lang="EN-GB">AHN2</span></strong></p> <p><strong><span lang="EN-GB">3. </span></strong><strong><span lang="EN-GB">AHN3</span></strong></p> <p><strong><span lang="EN-GB">4. </span></strong><strong><span lang="EN-GB">AHN4</span></strong></p> <p><strong><span lang="EN-GB">5. </span></strong><strong><span lang="EN-GB">Maps</span></strong></p> <p><span lang="EN-GB">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.</span></p> <p><strong><span lang="EN-GB">6. Masks</span></strong></p> <ul> <li><span lang="EN-GB">ahn3_10m_mask_building_road_water.tif</span></li> <li><span lang="EN-GB">ahn4_10m_mask_building_road_water.tif</span></li> <li><span lang="EN-GB">ahn4_10m_mask_powerline.tif</span></li> <li><span lang="NL">ahn1_10m_NA_mask.tif</span></li> <li><span lang="NL">ahn2_10m_NA_mask.tif</span></li> <li><span lang="NL">ahn3_10m_NA_mask.tif</span></li> <li><span lang="NL">a</span><span lang="NL">hn4_10m_NA_mask.tif</span></li> </ul> <p><span lang="NL"> </span></p> <p><span lang="EN-GB">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) (<a href="https://www.kadaster.nl/zakelijk/producten/geo-informatie/topnl">https://www.kadaster.nl/zakelijk/producten/geo-informatie/topnl</a>). 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.</span></p> <p><strong><span lang="EN-GB">7. Auxiliary data</span></strong></p> <p><span lang="EN-GB">(1) Point_density</span></p> <ul> <li><span lang="EN-GB">ahn1_10m_point_density.tif</span></li> <li><span lang="EN-GB">ahn2_10m_point_density.tif</span></li> <li><span lang="EN-GB">ahn3_10m_point_density.tif</span></li> <li><span lang="EN-GB">ahn4_10m_point_density.tif</span></li> </ul> <p><span lang="EN-GB">(2) Pulse_density</span></p> <ul> <li><span lang="EN-GB">ahn3_10m_pulse_density.tif</span></li> <li><span lang="EN-GB">ahn4_10m_pulse_density.tif</span></li> </ul> <p><span lang="EN-GB">(3) Flighttime</span></p> <ul> <li><span lang="EN-GB">ahn3_10m_flighttime.tif</span></li> <li><span lang="EN-GB">ahn4_10m_flighttime.tif</span></li> </ul> <p><span lang="EN-GB">(4) DTM_DSM</span></p> <ul> <li><span lang="EN-GB">ahn2_10m_dtm.tif</span></li> <li><span lang="EN-GB">ahn2_10m_dsm.tif</span></li> <li><span lang="EN-GB">ahn3_10m_dtm.tif</span></li> <li><span lang="EN-GB">ahn3_10m_dsm.tif</span></li> <li><span lang="EN-GB">ahn4_10m_dtm.tif</span></li> <li><span lang="EN-GB">ahn4_10m_dsm.tif</span></li> </ul> <p><span lang="EN-GB">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</span><span lang="EN-GB">–</span><span lang="EN-GB">AHN4. All raster layers are provide at 10 m resolution.</span></p> <p><strong><span lang="EN-GB">8. Sensitivity analysis</span></strong></p> <ul> <li><span lang="EN-GB">Dunes</span></li> <li><span lang="EN-GB">Marsh</span></li> <li><span lang="EN-GB">Grassland</span></li> <li><span lang="EN-GB">Shrubland</span></li> <li><span lang="EN-GB">Woodland</span></li> <li><span lang="EN-GB">Code</span></li> <li><span lang="EN-GB">Figure</span></li> </ul> <p><span lang="EN-GB">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.</span></p> <p><strong><span lang="EN-GB">9. Use_case</span></strong></p> <p><span lang="EN-GB">(1) Multi-temporal_AHN</span></p> <ul> <li><span lang="EN-GB">Data</span></li> <li><span lang="EN-GB">Usecase_multi-temporal_AHN.R</span></li> </ul> <p><span lang="EN-GB">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). </span></p> <p><span lang="EN-GB">(2) Natura2000</span></p> <ul> <li><span lang="EN-GB">Data</span></li> <li><span lang="EN-GB">Natura2000_end2021_HABITATCLASS.csv</span></li> <li><span lang="EN-GB">Natura2000_NL_habitat_grouped.csv</span></li> <li><span lang="EN-GB">Usecase_Natura2000.R</span></li> </ul> <p><span lang="EN-GB">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 the dominant habitat class was highlighted for each site. </span></p> <p><strong><span lang="EN-GB">10. Ground classification</span></strong></p> <ul> <li><span lang="EN-GB">Raw_point_cloud</span></li> <li><span lang="EN-GB">Computed_metrics </span></li> <li><span lang="EN-GB">Plottings_and_code</span></li> <li><span lang="EN-GB">ArcGIS_project</span></li> </ul> <p><span lang="EN-GB">It contains four subfolders: (1) The original point cloud for each sample area (AHN2–AHN4) (subfolder: Raw_point_cloud); (2) The 25 LiDAR metrics computed from the original point clouds with pre-classification of AHN and from the new terrain filtering method across AHN2–AHN4 (subfolder: Computed_metrics); (3) Generated violin plots for the comparison of vegetation change detection and the python code employed (subfolder: Plottings_and_code); (4) an ArcGIS project which the shapefiles of the study area and sample plots are provided (subfolder: ArcGIS_project).</span></p> <p><strong><span lang="EN-GB">Code availability</span></strong></p> <p><span lang="EN-GB">Jupyter Notebooks for processing AHN datasets: </span></p> <p><span lang="EN-GB"><a href="https://github.com/ShiYifang/AHN">https://github.com/ShiYifang/AHN</a></span></p> <p><span lang="EN-GB">Laserfarm workflow repository: </span></p> <p><span lang="EN-GB"><a href="https://github.com/eEcoLiDAR/Laserfarm">https://github.com/eEcoLiDAR/Laserfarm</a></span></p> <p><span lang="EN-GB">Laserchicken software repository: </span></p> <p><span lang="EN-GB"><a href="https://github.com/eEcoLiDAR/laserchicken">https://github.com/eEcoLiDAR/laserchicken</a></span></p> <p><span lang="EN-GB">Code for downloading AHN dataset: <a href="https://github.com/ShiYifang/AHN/tree/main/AHN_downloading">https://github.com/ShiYifang/AHN/tree/main/AHN_downloading</a></span></p> <p><span lang="EN-GB">Code for generating masks for AHN datasets: <a href="https://github.com/ShiYifang/AHN/tree/main/AHN_masks">https://github.com/ShiYifang/AHN/tree/main/AHN_masks</a></span></p> <p><span lang="EN-GB">Code for demonstration of ecological use cases: <a href="https://github.com/ShiYifang/AHN/tree/main/Use_case">https://github.com/ShiYifang/AHN/tree/main/Use_case</a></span></p> <p> </p>
Dataset for: A revised and expanded deep radiostratigraphy of the Greenland Ice Sheet from airborne radar sounding surveys between 1993–2019
<p>Version 5 changes the variable names in the NetCDF file so that are more readable by xarray.</p> <p>The NetCDF v4 file (Greenland_radiostratigraphy_v2.nc) contains both the gridded depths of synthetic isochrones (i.e., at ages of interest that were not necessarily directly observed) and the age at regular normalized depths (10–80%) in the ice sheet.</p> <p>For each campaign that was traced, there is an HDF5-compliant .mat MATLAB file (Greenland_radiostratigraphy_v2_X_Y, where X is the year the campaign was flown and Y is the aircraft that was used) that contains the original traced and dated reflections for each traced segment in that campaign. If using Python instead of MATLAB, the mat73 package and its loadmat function can be used to load these files. An example Jupyter notebook is included that illustrates how to access the full contents of .mat file using Python and this package. There is also a zipped (.zip) archive for each campaign that contains GeoPackage (.gpkg) files for each traced segment of that campaign. Because of format restrictions, these GeoPackages only include the traced reflections' depths, with the ages in the metadata. </p> <p> </p>
Band Ratio Mosaics from Airborne Hyperspectral Data at Aramo, Spain
<p> </p> <table> <tbody> <tr> <td> <h2>Metadata information</h2> </td> <td> </td> </tr> <tr> <td><strong>Full Title</strong></td> <td>Band Ratio Mosaics from Airborne Hyperspectral Data at Aramo, Spain</td> </tr> <tr> <td><strong>Abstract</strong></td> <td> <p>This dataset comprises results from the S34I Project, derived from processing airborne hyperspectral data acquired at the Aramo pilot site in Spain. Spectral Mapping Services (SMAPS Oy) conducted the airborne data acquisition in May 2024 using the Specim AisaFENIX sensor (covering VNIR-SWIR spectral ranges) over 17 flight lines. SMAPS performed geometric correction, radiometric calibration to reflectance, and atmospheric correction of the data. Subsequent processing steps included spectral smoothing with a Savitzky-Golay filter, cloud masking, bad pixel corrections, and hull correction (continuum removal).</p> <p>Manual processing and interpretation of hyperspectral data is a challenging, time-consuming, and subjective task, necessitating automated or semi-automated approaches. Therefore, we present a semi-automated workflow for large-scale interpretation of hyperspectral data, based on a combination of state-of-the-art methodologies. This dataset results from the calculation of a series of band ratios applied to the images and their subsequent mosaicking into a TIFF file. The mosaics are delivered as georeferenced TIFF files that cover approximately 97 km² with a spatial resolution of 1.2 m per pixel. The NoData value is set to -9999, representing areas of cloud removal or missing flight lines. The projected coordinate system is UTM Zone 30 Northern Hemisphere WGS 1984, EPSG 4326.</p> <p>Hyperspectral band ratios involve applying mathematical operations (such as division, subtraction, addition, or multiplication) among the reflectance values of different spectral bands. This technique enhances subtle variations in how materials absorb and reflect light across the electromagnetic spectrum. These variations are caused by electronic transitions, vibrations of chemical bonds (including -OH, Si-O, Al-O, and others), and lattice vibrations within the material's crystal structure.</p> <p>By creating these mathematical combinations, specific absorption features are emphasized, generating unique spectral fingerprints for different materials. However, these fingerprints alone cannot definitively identify a mineral, as different minerals may share similar absorption features due to common chemical bonds or crystal structures. Spectral geologists use band ratios as a tool to highlight potential areas of interest, but they must integrate this information with other geological knowledge and analyses to accurately interpret the mineralogy of an area. </p> <p>This dataset includes nine spectral band ratios. The mathematical formulas used to calculate each ratio are provided below:</p> <p> </p> <p>BR1 target Carbonate / Chlorite / Epidote</p> <p>BR1 = ((C7 + C9) / (C8))</p> <p>C7= Mean of bands between 2246.6 and 2257.55 nm</p> <p>C8= Mean of bands between 2339 and 2345 nm</p> <p>C9= Mean of bands between 2400 and 2410 nm</p> <p> </p> <p>BR2 target Chlorite</p> <p>BR2 = ((Cl1 + Cl2) / (Cl2))</p> <p>Cl1 = Mean of bands between 2191.93 and 2197.4 nm</p> <p>Cl2 = Mean of bands between 2246.63 and 2257.55 nm</p> <p> </p> <p>BR3 target Clay</p> <p>BR3 = ((C1 + C2) / (C2))</p> <p>C1 = Mean of bands between 1590.32 and 1612.56 nm</p> <p>C2 = Mean of bands between 2191.93 and 2208.35 nm</p> <p> </p> <p>BR4 target Dolomite</p> <p>BR4 = ((C6 + C8) / (C7))</p> <p>C6= Mean of bands between 2186 and 2191 nm</p> <p>C7= Mean of bands between 2246.6 and 2257.55 nm</p> <p>C8= Mean of bands between 2339 and 2345 nm</p> <p> </p> <p>BR5 target Fe2</p> <p>BR5 = ((Fe2n + Fe2d) / (Fe2d))</p> <p>Fe2n = Mean of bands between 721.85 and 742.48 nm</p> <p> </p> <p>BR6 target Fe3</p> <p>BR6 = ((Fe3n - Fe3d) / (Fe3n + Fe3d))</p> <p>Fe3n = Mean of bands between 776.87 to 811.26 nm</p> <p>Fe3d = = Mean of 3 bands around 610 nm</p> <p> </p> <p>BR7 target = Kaolinite / clays</p> <p>BR7 = ((K1 + K2) / (K3 + K4))</p> <p>K1 = Mean of bands between 2082.27 and 2104.23 nm</p> <p>K2 = Mean of bands between 2104.23 and 2115.2 nm</p> <p>K3 = Mean of bands between 2159.07 and 2164.55 nm</p> <p>K4 = Mean of bands between 2202.88 and 2208.35 nm</p> <p> </p> <p>BR8 target Kaolinite2 / clays</p> <p>BR8 = ((K1_2 + K2_2) / (K2_2))</p> <p>K1_2 = Mean of bands between 2197.4 and 2219.29 nm</p> <p>K2_2 = Mean of bands between 2159.07 and 2170.03 nm</p> <p> </p> <p>BR9 target NDVI (Normalized Difference Vegetation Index)</p> <p>BR9 = ((NIR - Red) / (NIR + Red))</p> <p>NIR= Mean of bands between 776.87 and 811.26 nm</p> <p>Red = Mean of bands between 666.87 to 680.6 nm</p> </td> </tr> <tr> <td>Keywords</td> <td>Earth Observation, Remote Sensing, Hyperspestral Imaging, Automated Processing, Hyperspectral Data Processing, Mineral Exploration, Critical Raw Materials</td> </tr> <tr> <td>Pilot area</td> <td>Aramo</td> </tr> <tr> <td>Language</td> <td> <p>English</p> </td> </tr> <tr> <td>URL Zenodo</td> <td>https://zenodo.org/uploads/14193286</td> </tr> <tr> <td><strong>Temporal reference</strong></td> <td> </td> </tr> <tr> <td>Acquisition date (dd.mm.yyyy)</td> <td>01.05.2024</td> </tr> <tr> <td>Upload date (dd.mm.yyyy)</td> <td>20.11.2024</td> </tr> <tr> <td><strong>Quality and validity</strong></td> <td> </td> </tr> <tr> <td>Fromat</td> <td>GeoTiff</td> </tr> <tr> <td>Spatial resolution</td> <td>1.2m </td> </tr> <tr> <td>Positional accuracy</td> <td>0.5m </td> </tr> <tr> <td>Coordinate system</td> <td>EPGS 4326</td> </tr> <tr> <td><strong>Access and use constrains</strong></td> <td> </td> </tr> <tr> <td>Use limitation</td> <td>None</td> </tr> <tr> <td>Access constraint</td> <td>None</td> </tr> <tr> <td>Public/Private</td> <td>Public</td> </tr> <tr> <td><strong>Responsible organisation</strong></td> <td> </td> </tr> <tr> <td>Responsible Party</td> <td>Beak Consultants GmbH</td> </tr> <tr> <td>Responsible Contact</td> <td>Roberto De La Rosa</td> </tr> <tr> <td><strong>Metadata on metadata</strong></td> <td> </td> </tr> <tr> <td>Contact</td> <td>Roberto.delarosa@beak.de</td> </tr> <tr> <td>Metadata language</td> <td>English</td> </tr> </tbody> </table>
Large-scale 3D building and tree datasets constructed from airborne LiDAR point clouds in Glasgow, UK
<p>This is the updated version of building 3D model data. The revision includes appending attributes to the lod1 and lod2 shapefile and creating cityjson file for each 3D building model. All 3D building models are available in mesh (.obj), multipath shapefile, and cityjson (.json) now.</p> <p><strong>IMPORTANT NOTE: We suggest using the building footprint, lod1, and lod2 data of this version (Version v4).</strong></p> <p>Urban Big Data Centre of the University of Glasgow generates 3D city models via the airborne LiDAR point clouds acquired between 2020-2021 on behalf of Glasgow City Council. It is a large-scale 3D city model containing 3D information on terrain, trees, and buildings in Glasgow City. This dataset comprises terrain, tree canopy, and building products derived from high-density airborne LiDAR point clouds. </p> <p>The terrain products include Digital Terrain Model (DTM), Digital Surface Model (DSM), and normalized Digital Surface Model (nDSM) in 0.5 m spatial resolution. The DTM and DSM rasters were provided by the vendor and nDSM rasters were obtained by subtracting DTM from DSM. Terrain products are provided in 5 km by 5 km GeoTIF format raster.</p> <p>The tree canopy products are composed of canopy height models (CHM) and tree top locations. Classified tree point clouds were applied with pit-free algorithm to generate CHM in 0.5 m grid raster in GeoTIF format [1]-[2]. Treetop locations were identified by using Local Maximum Filter based on CHM and are recorded as points in Shapefile format. The tree canopy products are provided in 5 km by 5 km tiles.</p> <p>Building 3D model products include footprint polygons with building height attributes and 3D mesh of building models in LoD1 and LoD2 levels. A series of processes such as converting building point clouds to building height models (BHM), converting BHM to polygons, and polygon regularization were conducted to obtain the building footprint polygons. Building height attributes were calculated from BHM for each footprint. The building footprint data are provided in Shapefile format. LoD1 models were generated based on the footprint and average height of the building. LoD2 models were constructed based on footprint and building point cloud with City3D tool[3]. LoD1 and LoD2 models are provided in OBJ and shapefile format. Building 3D model products are provided in 5 km by 5 km tiles. The RMSE of Euclidean distances between each point in the point cloud to the reconstructed model was calculated to evaluate the LoD2 model construction. A table of RMSE and a note for a few problematic models are provided.</p>
Bathymetry beneath the Amery ice shelf, East Antarctica, revealed by airborne gravity
<p>We estimated the seafloor topography beneath the Amery Ice Shelf, East Antarctica, from airborne gravity anomaly through a nonlinear inversion method called simulated annealing. The estimation results provide a view of the seafloor beneath the Amery Ice Shelf, where direct bathymetric observations are rare. The model, 'gravity_estimated_seafloor_topography_beneath_the_Amery_Ice_Shelf.nc', is in NetCDF format which can be read through MATLAB commands "ncdisp" and "ncread". Contents of the model can be found in "contents.txt". The MATLAB program "nc2mat.m" reads the NetCDF ".nc" format model and saves the variables in the model to a MATLAB ".mat" format file.</p>
Dataset from paper "Canopy palm cover across the Brazilian Amazon forests mapped with airborne LiDAR data and deep learning"
<p><strong>Data and code from the paper:</strong></p> <p>Dalagnol, R., Wagner, F. H., Emilio, T., Streher, A. S., Galvão, L. S., Ometto, J. P. H. B., & Aragão, L. E. O. C. (2022). Canopy palm cover across the Brazilian Amazon forests mapped with airborne LiDAR data and deep learning. Remote Sensing in Ecology and Conservation, 1–14. https://doi.org/10.1002/rse2.264</p> <p><strong>Link:</strong> <a href="https://doi.org/10.1002/rse2.264">https://doi.org/10.1002/rse2.264</a></p> <p> </p> <p><strong>This repository contains:</strong></p> <p><strong>1) model_train.R:</strong> This is the code to run the U-Net model in R language.</p> <p><strong>2) input.rar:</strong> Dataset of lidar canopy height model (CHM) images and masks (labels) patches of canopy palms obtained from four sites in the Brazilian Amazon. The images/masks have 128 x 128 pixels, where each pixel represents 0.5 m in the terrain. The dataset contains 2,269 images and masks, with close to 7,000 palms manually labelled.</p> <p><strong>3) unet_weights_best.h5:</strong> These are the best weights for the U-Net architecture achieved in the paper.</p> <p><strong>4) palm_stats.RData:</strong> Data frame with the lat/lon coordinates and palm metrics extracted for the 610 lidar sites in the Brazilian Amazon. (i) n_total is the number of palms, (ii) n_ha is the density of palms per hectare, (iii) crown_ metrics are based on the area of palm segments (in square meters), (iv) cover_total is the total area occupied by palms in the forest canopy (in square meters), (v) cover_rel is the relative cover of palms in the forest canopy (in percentage), (vi) height_ metrics are based on the height of palm segments (in meters), (vii) palm_height_dif_mean is the mean difference between palm height and local canopy height, and (viii) palm_height_dif_pvalue is the p-value assessing the statistical difference between the palm and canopy heights where 0 means no difference and -1/+1 means a negative/positive difference.</p> <p> </p> <p>If you need anything else, please contact the corresponding author: Ricardo Dalagnol (ricds@hotmail.com).</p> <p> </p> <p><strong>If you use these data, please cite the paper:</strong></p> <p>Dalagnol, R., Wagner, F. H., Emilio, T., Streher, A. S., Galvão, L. S., Ometto, J. P. H. B., & Aragão, L. E. O. C. (2022). Canopy palm cover across the Brazilian Amazon forests mapped with airborne LiDAR data and deep learning. Remote Sensing in Ecology and Conservation, 1–14. https://doi.org/10.1002/rse2.264</p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.