Abstract Reconstruction of past topography in palaeo-DEMs serves various geomorphological analyses. Constructing a palaeo-DEM by stripping young elements from a LiDAR DEM can provide results for large study areas at high resolution. However, such a ‘top-down’ approach is more suited to recent periods and geomorphologically static parts of the landscape than to geomorphologically dynamic areas and periods farther back in time. Here, we explore this approach by reconstructing the early medieval (circa 800 CE) topography of the Lower Rhine river valley and upper delta in Germany and the Netherlands. The large (4500 km2) study area contains abundant anthropogenic terrain modification and stretches across geomorphologically active as well as inactive zones. We first removed all anthropogenic relief elements from the LiDAR DEM, using separate procedures for linear and non-linear elements. These steps were sufficient to obtain the palaeotopography of the inactive zone, characterized by inherited natural relief. Then, we reconstructed the topography and bathymetry in the fluvially-reworked active zone by incorporating geological and historical geographical information. We present and evaluate zonal averages of elevation differences between the modern and past valley floor topography in this densely populated area with complex land-use history, which allows us to approximate total anthropogenic volumetric change. Further comparisons with the modern LiDAR DEM elucidate changes in floodplain negative-relief connectivity, showing the potential importance of investing in palaeo-DEMs when assessing past river flooding. Our palaeo-DEM construction workflow is deployable at diverse spatial scales and widely applicable to other lowland areas, because of its top-down and generic nature. The relative importance of different workflow aspects depends on the time period that is targeted. Beyond a target age of 10–15 ka, valley floors that are too large in area are to be considered geomorphologically dynamic and the top-down approach to palaeo-DEM construction is no longer advisable.

中文翻译：

---

LiDAR 衍生的高分辨率古 DEM 构建工作流程及其在中世纪早期莱茵河下游河谷和上三角洲的应用

摘要 古 DEM 中过去地形的重建服务于各种地貌分析。通过从 LiDAR DEM 中剥离年轻元素来构建古 DEM 可以为大型研究区域提供高分辨率的结果。然而，这种“自上而下”的方法更适合于近期和地貌静态部分，而不是地貌动态区域和更远的时期。在这里，我们通过重建德国和荷兰的下莱茵河谷和上三角洲的中世纪早期（大约公元 800 年）地形来探索这种方法。大型（4500 平方公里）研究区包含丰富的人为地形改造，横跨地貌活动区和非活动区。我们首先从 LiDAR DEM 中删除了所有人为救援元素，对线性和非线性元素使用单独的程序。这些步骤足以获得以继承自然地貌为特征的非活动区的古地形。然后，我们结合地质和历史地理信息重建了河流改造活动区的地形和水深。我们展示并评估了这个人口稠密、土地利用历史复杂的地区现代和过去谷底地形之间的带状平均海拔差异，这使我们能够估计总的人为体积变化。与现代 LiDAR DEM 的进一步比较阐明了洪泛区负地势连通性的变化，显示了在评估过去的河流洪水时投资于古 DEM 的潜在重要性。由于其自上而下和通用的性质，我们的古 DEM 构建工作流程可在不同的空间尺度上部署并广泛适用于其他低地地区。不同工作流方面的相对重要性取决于目标时间段。超过 10-15 ka 的目标年龄，面积过大的谷底将被视为地貌动态，不再建议采用自上而下的古 DEM 构建方法。