The causal drivers of short-term changes (days to months) in human-, wind-, and wave-driven sand transport on a sandy beach are not often considered in an integral and data-driven approach. However, improving current knowledge on (urban) sandy beach topographical change requires the incorporation of multi-scale, cross-sectional and human factors. In this research we process a time series of 21,194 hourly point clouds, obtained in a Permanent Terrestrial Laser Scanning setup. From this 3D time series we extract 5,102 short-term temporary surface dynamics, through a method called 4D objects-by-change (4D-OBCs). The causal drivers of two of these 4D-OBCs are investigated in detail. One is interpreted as an aeolian depositional surface dynamic (1), and one as a bulldozer deposit, that consecutively eroded under high wave energy conditions (2). The dynamics show clear correlation to a particular combination of wind direction and intensity (1), and wave height and wave period (2), indicating that point cloud time series derived 4D-OBCs are useful data to study causality of short-term surface dynamics of different origins. However, to study these surface dynamics systematically and derive statistical proof of causal relations we must consider multivariate correlations, as well as spatiotemporal dependence between sediment dynamics and larger scale morphological changes on the beach.

Botanical trees demonstrate complicated motion patterns under wind forcing. Studying these motion patterns is relevant as tree parts may become hazardous during storms, and the movement of these trees affects the surrounding wind field, providing, for example, wind shadows. Eventually, wind enforced tree motion may even be incorporated in wind simulation models used to design liveable neighbourhoods. These motion patterns are, however, difficult to assess as in situ sensors lack coverage while video imagery lacks full 3D capacity. Recently, fast 3D laser scanners became available. For one such scanner, the Livox Avia, a workflow is presented that allows scanning tree movement at frequencies of up to 2 Hz. The scanner, possibly triggered to scan by a wind-meter, is building up a space-time array, consisting of range measurements at a give time, in a given direction. In a consecutive step, this space-time array is decomposed into single point clouds of a certain duration. The decomposition can be done over different time intervals; point clouds of shorter duration show lower coverage of measurements, but are also less affected by moving tree parts during scanning. After decomposition, motion vector fields between consecutive point clouds representing different time frames are estimated using the existing PlantMove algorithm. Applications of the workflow on different moving tree scenarios demonstrate how this novel measurement strategy is able to reveal complex and detailed spatial and temporal patterns of tree movement, with observed displacements on the order of decimeters, under wind stress. Extending the setup to e.g. three scanners will enable to capture wind induced complex tree movement at branch level, which is essential input for running reliable wind flow simulations at neighbourhood level.

Dutch beaches are increasingly urbanized with both permanent beach pavilions and seasonal sheds and holiday houses. The effect of these buildings on long term dune development between 1999 and 2024 is studied in this paper along ~ 100 km of coast on the outer delta in the south western part of the Netherlands. A total of ~ 7000 beach buildings have been manually identified in this period based on satellite images and the time line function of Google earth desktop. The effect of the buildings is determined and analyzed at 477 cross-shore profiles with dune volumes and properties like dune toe, top and heel based on airborne lidar datasets of 1999 and 2024. On natural beaches the dune toe position is derived from profile information, whereas on urbanized beaches near buildings the dune toe is based on the location of the buildings. Yearly volume changes at the profile locations vary between -10 m³/m/y and up to 40 m³/m/y. The results indicate that smaller and standalone buildings allow for larger variations in dune volume changes and suggest that larger buildings and connected buildings impede natural dune dynamics which could impact coastal resilience in the long run.

Dynamic beaches world-wide are monitored by so-called Argus cameras. Their automatic capturing results in large databases of ∼ 30' interval coastal images acquired during different illumination conditions. We present a lightweight method to automatically extract sand and supporting classes from ∼ 1 million Argus images, spanning ∼ 10 years, of the Sand Engine, The Netherlands, a nature-based solution for beach erosion. The workflow consists a series of neural networks. First, a ResNet18 model selects images of sufficient quality. Second, 24 pixel-wise shallow multi-layered perceptrons (MLPs) classify pixels into 5 classes, Water, Foam and Eolian, Wet and Armored Sand. The 24 MLPs correspond to 12 cameras with two different lighting conditions. The results from the 24 MLPs are used as pseudo labels for 24 CNNs that improve the initial classification by including spatial awareness. These 24 CNNs are fused in one ensemble CNN, robust to camera choice and lighting condition. At this stage, still a separate CNN is used to additionally detect vegetation pixels. Results show in most cases good agreement with human interpretation, with an overall accuracy of ∼ 88%. Most promising is that ∼ 150.000 images per day can be processed on a high-end consumer PC at a quality difficult to obtain by a human operator. Future work should focus on exploiting the large database of results for improving our understanding of dynamic processes at this challenging environment. As the workflow is generic in nature, it should be easily applicable for other image based monitoring databases.