High conservation value forest in Denmark (DK Forest LiDAR v1.0.0)

Assmann et al. (in prep) Temperate forests of high conservation value are successfully identified by satellite and LiDAR data fusion.

Classifications of high conservation value of forests in Denmark using the EcoDes-DK15 dataset and other spatial data.

Disclaimer: This project is currently in peer-review.

Results

• Leaflet web app - best model (map of projections)
• Leaflet web app - all models (map of projections)
• Summary stats (area estimates)

Project overview

• Workflow Overview

Data description

• Forest annotations and training data
• Predictor overview
• Focal (window) predictor selection

Model performance

• Gradient Boosting performance
• Random Forest performance

Data / Outputs

All data outputs are available on Zendo:

• https://zenodo.org/records/10419006

Compressed and cloud optimised rasters & vectors can also be downloaded from AWS:

• Best model: Random Forest Projections BIOWIDE v1.0.0 (40.2 MB, GeoTiff)
• Random Forest Projections Derek’s Stratification v1.0.0 (40.2 MB, GeoTiff)
• Gradient Boosting Projections BIOWIDE v1.0.0 (41.5 MB, GeoTiff)
• Gradient Boosting Projections Derek’s Stratification v1.0.0 (41.5 MB, GeoTiff)
• Disturbance map v0.9.1 (17.8 MB, GeoTiff)
• Training Polygons v0.9.0 (115.7 MB, GeoJson)

Summary report

• Summary report - website snapshot v1.0.1 (3.28 MB, PDF))

Auxiliary

• Guide on how to visualise cloud optimised rasters

---

[last update: 17 October 2024]