Spatially explicit distribution of hedgerows across German agricultural landscapes

This dataset contains a spatially explicit estimation of hedgerows across agricultural areas in Germany at 3 m resolution in cloud optimized geotif format. The dataset has been developed as part of the project KlimaFern, Remote Sensing for Improved Climate reporting. Hedgerows here are defined as linear woody features of 3-20 m wide with agriculture on at least one side, akin to the Land Use/Cover Area Survey (LUCAS) criteria. The hedgerow predictions are based on multitemporal PlanetScope imagery for the year 2022 and a convolutional neural network with a UNet backbone. No postprocessing was applied. Details on the calibration and accuracies can be found in Muro et al., (2025) https://doi.org/10.1016/j.rse.2025.114870. Known limitations of the product are: hedgerows narrower than 3 m in canopy and hedgerows recently coppiced (false negatives) hedgerows close (<~10m) and parallel to other hedgerows or forest patches (false negatives) hedgerows slightly above the 20 m width definition may or may not be included dense tree lines without hedgerows underneath (false positives) Users can request the data before the embargo date. Just press "Request access" below. Contacts: Javier Muro: [e-mail] Stefan Erasmi: [e-mail]本数据集以3米分辨率的云优化GeoTIFF（Cloud Optimized GeoTIFF）格式，提供了德国农业区域内树篱的空间显性估算结果。本数据集作为"KlimaFern：助力精准气候报告的遥感技术"项目的一部分开发完成。 此处的树篱被定义为宽度3至20米的线性木本植被，且至少一侧毗邻农业用地，该定义与土地利用/覆盖面积调查（Land Use/Cover Area Survey, LUCAS）的标准一致。树篱的预测结果基于2022年的多时相PlanetScope影像以及搭载UNet骨干网络的卷积神经网络（Convolutional Neural Network, CNN），未进行任何后处理步骤。关于校准方法与精度的详细信息可参阅Muro等人（2025）的研究：https://doi.org/10.1016/j.rse.2025.114870。 本产品存在以下已知局限性： 1. 冠幅小于3米的树篱以及近期被砍伐的树篱（将产生假阴性结果）； 2. 与其他树篱或森林斑块间距小于约10米且呈平行分布的树篱（将产生假阴性结果）； 3. 宽度略超出20米定义范围的树篱，可能被纳入预测结果，也可能未被纳入； 4. 下方无树篱的密集林带（将产生假阳性结果）。 用户可在数据禁运期结束前申请获取该数据，只需点击下方的「请求访问」按钮即可。 联系方式： 哈维尔·穆罗（Javier Muro）：[e-mail] 斯特凡·埃拉斯米（Stefan Erasmi）：[e-mail]