Sub-dataset: WRB soil types probabilities (part 2)

Disclaimer

This is the first release of pan-EU predictions of soil health indicators (the Soil Health Data Cube). Use for testing purposes only. A publication describing methods used has been submitted to PeerJ and is in review. Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or European Commision. Neither the European Union nor the granting authority can be held responsible for them. The data is provided "as is". AI4SoilHealth project consortium and its suppliers and licensors hereby disclaim all warranties of any kind, express or implied, including, without limitation, the warranties of merchantability, fitness for a particular purpose and non-infringement. Neither AI4SoilHealth project Consortium nor its suppliers and licensors, makes any warranty that the Website will be error free or that access thereto will be continuous or uninterrupted. You understand that you download from, or otherwise obtain content or services through, the Website at your own discretion and risk.

Description

This dataset covers pan-European areas, including Ukraine, the UK, and Turkey. This data cube could be used for applications such as soil property mapping and comprehensive soil health assessment across Europe. The dataset spans four depth ranges and multiple time periods, providing information for studies on soil organic carbon stock and dynamics.

This dataset is part of the "Soil type (World Reference Base) map of Europe based on Ensemble Machine Learning and multiscale EO data" dataset. Check the related identifiers section below to access other parts of the dataset.

This data set includes:

• Soil types classification and relative entropy:
  This data includes hard classes maps (185 soil type classes) produced by ensemble model and relative entropy (Kullback-Leibler divergence) maps in added information (bit) over a dummy distribution (scaled 1000x).
• Soil types probabilities (part 1):
  This data includes 92 averaged probabilities (0-1) maps for classes from abruptic.acrisols to gleyic.arenosols. The probabilites were scaled 100x (0-100).
• Soil types probabilities (part 2):
  This data includes 93 averaged probabilities maps for classes from gleyic.cambisols to vitric.andosols. The probabilites were scaled 100x (0-100).

Related identifiers

• WRB soil types classification and relative entropy
• WRB soil types probabilities (part 1)
• WRB soil types probabilities (part 2)

Data Details

• Time period: long term.
• Type of data: Soil types classification and model probabilities.
• How the data was collected or derived: The data was derived using ensemble ML models.
• Statistical methods used: Relative entropy (Kullback-Leibler divergence)
• Limitations or exclusions in the data: The dataset does not include data for Svalbard.
• Coordinate reference system: EPSG:3035
• Bounding box (Xmin, Ymin, Xmax, Ymax): (900,000, 899,000, 7,401,000, 5,501,000)
• Spatial resolution: 30m
• Image size: 216,700P x 153,400L
• File format: Cloud Optimized Geotiff (COG) format.

Support

If you discover a bug, artifact, or inconsistency, or if you have a question please raise a GitHub issue: GitLab Issues (tbc)

Name convention

To ensure consistency and ease of use across and within the projects, we follow the standard Ai4SoilHealth and Open-Earth-Monitor file-naming convention. The convention works with 10 fields that describe important properties of the data. In this way users can search files, prepare data analysis etc, without needing to open files. For example, in soil.types_ai4sh.ensemble_c_30m_s_20220101_20221231_epsg.3035_v20240917.tif, the fields are:

1. generic variable name: soil.types = soil types
2. variable procedure combination: ai4sh.ensemble.abruptic.acrisols = AI4SH project, ensemble model, abrupitc acrisols soil type.
3. Position in the probability distribution/variable type: m = mean | c = class | p = probability
4. Spatial support: 30m
5. Depth reference: s = surface
6. Time reference begin time: 20220101 = 2022-01-01
7. Time reference end time: 20221231 = 2022-12-31
8. Bounding box: eu = pan-Europe
9. EPSG code: epsg.3035
10. Version code: v20240917 = version from 2024-09-17