EUNIS Habitat Maps: Enhancing Thematic and Spatial Resolution for Europe through Machine Learning

The EUNIS habitat classification is essential for categorising European habitats and supporting European policy on nature conservation and to implement the Nature Restoration Law. As such, to meet the growing demand for detailed and accurate habitat information, we provide spatial predictions for 260+ EUNIS habitat types at EUNIS level 3, together with validation and uncertainty analyses. More specifically, using ensemble machine learning models together with high-resolution satellite imagery and other climatic, terrain and soil variables, we produced an European habitat map at a 100-m resolution indicating the most likely EUNIS habitat at level 3 for every location across Europe. Predictions were validated for three independent countries, namely for France, the Netherlands and Austria. We also provide information on uncertainty and the most probable habitats at level 3 within each EUNIS level 1 formation. Products can be further refined with accurate and local land cover data. This product is thus likely to be particularly useful for restoration but also conservation purposes. Figure: Wall-to-wall map of EUNIS habitats at level 3 - (color coded at level 2 for visibility)EUNIS生境分类（EUNIS habitat classification）是欧洲生境分类的核心依据，用于对欧洲各类生境进行归类，支撑欧盟自然保护政策落地，并助力《自然恢复法》（Nature Restoration Law）的实施。为满足当前对生境详细精准信息日益增长的需求，本数据集针对EUNIS三级（EUNIS level 3）的260余种生境类型提供空间预测结果，同时附带验证与不确定性分析内容。 具体而言，本研究结合集成机器学习（ensemble machine learning）模型、高分辨率卫星影像（high-resolution satellite imagery），以及气候、地形与土壤等多源环境变量，生成了分辨率为100米的全欧生境分布图，该图可标注欧洲全境每个点位最匹配的EUNIS三级生境。模型预测结果已通过法国、荷兰与奥地利三个独立国家的独立验证数据进行校验。此外，本数据集还提供了EUNIS一级群系（EUNIS level 1 formation）内各点位的不确定性信息与三级生境的最可能分类结果。若辅以精准的本地土地覆被数据（land cover data），该产品的精度可进一步优化。本数据集不仅适用于自然保护工作，在生境恢复领域同样具有极高的应用价值。 图：EUNIS三级生境全覆盖分布图（为提升可读性，按二级分类进行色彩编码）