Sindewald et al - Identifying alpine treeline species using high-resolution WorldView-3 multispectral imagery and convolutional neural networks dataset
Description
The dataset contains region of interest (ROI) polygons for six treeline species found in Rocky Mountain National Park, CO. The tree and shrub species include limber pine (Pinus flexilis), willow (Salix glauca, Salix brachycarpa, and hybrids), Engelmann spruce (Picea engelmannii), subalpine fir (Abies lasiocarpa), glandular birch (Betula glandulosa), and quaking aspen (Populus tremuloides). The ROIs correspond to WorldView-3 image data obtained from Maxar, collected on July 21, 2020, with 0% cloud cover and an off-nadir angle of 16.8 degrees. WV-3 data include a panchromatic (black-and-white) band with 31 cm spatial resolution and eight multispectral bands with 1.24 m resolution: coastal blue (400-450 nm), blue (450-510nm), green (510-580 nm), yellow (585-625 nm), red (630-690 nm), red edge (705-745 nm), near-infrared 1 (N-IR1, 770-895 nm), and near-infrared 2 (N-IR2, 860-1040 nm). Details on orthorectification and atmospheric correction can be seen in the attendant paper. The dataset also includes a 10 m digital elevation model (DEM) from the U.S. Geological Survey EROS Data Center, interpolated to 1.24 m resolution using a cubic spline resampling method.
The image data were used to train convolutional neural networks (CNNs) to do pixel-based species classification. The CNN inputs were image chips (patches), each centered on an ROI pixel. The data are therefore provided in the form of these image patches. Additionally, a csv file is provided with mean radiance values for each ROI polygon for each multispectral WV-3 band with species labels.
These data accompany a paper titled "Identifying alpine treeline species using high-resolution WorldView-3 multispectral imagery and convolutional neural networks dataset", under review in the Biogeosciences special issue, "Treeline ecotones under global change: linking spatial patterns to ecological processes". Authors include
Laurel A. Sindewald1, Ryan Lagerquist2, Matthew D. Cross3, Theodore A. Scambos4, Peter J. Anthamatten5, Diana F. Tomback1
1 Department of Integrative Biology, University of Colorado Denver, Denver, 80204, USA
2 Cooperative Institute for Research in the Atmosphere, Colorado State University, Fort Collins, 80521, USA
3 Department of Geography and the Environment, University of Denver, Denver, 80208, USA
4 Earth Science and Observation Center, Cooperative Institute for Research in Environmental Sciences, University of Colorado Boulder, Boulder, 80309, USA
5 Department of Geography and Environmental Sciences, University of Colorado Denver, Denver, 80204, USA
Files
Sindewald_et_al_2025_Identifying_alpine_treeline_species_WV-3_CNNs_data.zip
Files
(3.3 GB)
Name | Size | Download all |
---|---|---|
md5:1c71351254d7dd9d63f40d8f6a651968
|
3.3 GB | Preview Download |
Additional details
Funding
- University of Colorado Denver
- Application of high-resolution satellite imagery to resolve treeline species composition: Case study using limber pine in Rocky Mountain National Park
Dates
- Collected
-
2020-07-21Date of satellite imagery acquisition