Geng Xiaoyuan, He Juanxia, Grima Victoria, Jiang Yefang, Tetreau Maggi, Crittenden Stephen, Kiley Simon, VandenBygaart Albert J, Vanrobaeys Jason
Science and Technology Branch, Agriculture and Agri-Food Canada, Ottawa, Canada.
Swan Lake First Nation, Swan Lake Land Governance Office, Swan Lake, R0G 0B9, Manitoba, Canada.
Sci Data. 2025 Jul 10;12(1):1178. doi: 10.1038/s41597-025-05460-4.
With the latest collection of soil point and available co-variables data, 100 m grids of soil type and selected soil attributes of Canada are developed. While model-based statistical validations during the machine learning runs are satisfactory, limited independent statistical validations are also made. The 100 m soil landscape grids can be used for national and regional soil organic carbon stock inventory and carbon sequestration assessment. There are also potential applications for soil health, soil erosion, land suitability and other large area-based modelling and applications.
利用最新收集的土壤点位和可用协变量数据,绘制了加拿大100米网格的土壤类型和选定土壤属性图。虽然机器学习运行期间基于模型的统计验证结果令人满意,但也进行了有限的独立统计验证。100米的土壤景观网格可用于国家和区域土壤有机碳储量清查及碳固存评估。在土壤健康、土壤侵蚀、土地适宜性及其他大面积建模与应用方面也有潜在用途。