Location: here From this publication: https://link.springer.com/article/10.1007/s10346-019-01208-5 The 3D model of the area was obtained by using a technique that combines aerial photogrammetry from a UAS (González-Jorge et al. 2016) to obtain a complete point cloud of the study area and a terrestrial survey from a ground-based LiDAR to obtain a better accuracy of the critical areas, such as the unstable blocks. Some targets were added to the structure to facilitate the registration process. The photogrammetric flights were made using a Phantom 3 UAS equipped with a 12 Mpx digital camera (f/2.8, 94°). Flight height was about 87.5 m. 220 photographs were taken for the photogrammetric restitution, a step carried out with the photogrammetric suite Photoscan. After the restitution, the software allows exporting a 3D point cloud. It was levelled and scaled using a Global Navigation Satellite System Trimble R8, with which a horizontal and vertical accuracy of 1.97 and 3.24 cm, respectively, was obtained. The photogrammetric flights covered an area of 0.102 km2, with a ground resolution of 3.59 cm/px and a reprojection error of 0.886 px. With this process, a point cloud with 29 million points was obtained. LiDAR data acquisition was done using six scan stations to provide a coverage of the unstable blocks area. A Faro Focus 330X LiDAR was used for this survey, which has a resolution of 2 mm at 25 m (González-Jorge et al. 2018). Each point cloud has between 9 and 15 million points. Targets on the structure provide control points for the registration of the point clouds and for joining all the data in a single coordinate reference system. After registration, this resulting point cloud is again merged together with that one provided by the UAS. The point cloud provided by the UAS was used as reference because it was the most complete one. Distance histograms between the points of the reference point cloud and each LiDAR point cloud registered are provided by Fig. 3.1 in the Online Resource 1. As can be seen, the registration error is below 6 cm in almost all the registered LiDAR point clouds. Finally, a global and detailed point cloud from the structure and its surroundings was obtained (Fig. 2)