Satellite remote sensing is transforming coastal science from a “data-poor” field into a “data-rich” field. Sandy beaches are dynamic landscapes that change in response to long-term pressures, short-term pulses, and anthropogenic interventions. Until recently, the rate and breadth of beach change have outpaced our ability to monitor those changes, due to the spatiotemporal limitations of our observational capacity. Over the past several decades, only a handful of beaches worldwide have been regularly monitored with accurate yet expensive in situ surveys. The long-term coastal-change data of these few well-monitored beaches have led to in-depth understanding of many site-specific coastal processes. However, because the best-monitored beaches are not representative of all beaches, much remains unknown about the processes and fate of the other >99% of unmonitored beaches worldwide. The fleet of Earth-observing satellites has enabled multiscale monitoring of beaches, for the very first time, by providing imagery with global coverage and up to daily frequency. The long-standing and ever-expanding archive of satellite imagery will enable coastal scientists to investigate coastal change at sites vulnerable to future sea-level rise, that is, (almost) everywhere. In the past decade, our capability to observe coastal change from space has grown substantially with computing and algorithmic power. Yet, further advances are needed in automating monitoring using machine learning, deep learning, and computer vision to fully leverage this massive treasure trove of data. Extensive monitoring and investigation of the causes and effects of coastal change at the requisite spatiotemporal scales will provide coastal managers with additional, valuable information to evaluate problems and solutions, addressing the potential for widespread beach loss due to accelerated sea-level rise, development, and reduced sediment supply. Monitoring from Earth-observing satellites is currently the only means of providing seamless data with high spatiotemporal resolution at the global scale of the impending impacts of climate change on coastal systems.

Structure from motion (SFM) has become an integral technique in coastal change assessment; the U.S. Geological Survey (USGS) used Agisoft Metashape Professional Edition photogrammetry software to develop a workflow that processes coastline aerial imagery collected in response to storms since Hurricane Florence in 2018. This report details step-by-step instructions to create three-dimensional (3D) spatial products from both singular and repeated collections of shoreline aerial imagery. The products can be used for real-time hazard guidance and future forecasting and recovery endeavors.

The application of deep learning, specifically deep convolutional neural networks (DCNNs), to the classification of remotely-sensed imagery of natural landscapes has the potential to greatly assist in the analysis and interpretation of geomorphic processes. However, the general usefulness of deep learning applied to conventional photographic imagery at a landscape scale is, at yet, largely unproven. If DCNN-based image classification is to gain wider application and acceptance within the geoscience community, demonstrable successes need to be coupled with accessible tools to retrain deep neural networks to discriminate landforms and land uses in landscape imagery. Here, we present an efficient approach to train/apply DCNNs with/on sets of photographic images, using a powerful graphical method called a conditional random field (CRF), to generate DCNN training and testing data using minimal manual supervision. We apply the method to several sets of images of natural landscapes, acquired from satellites, aircraft, unmanned aerial vehicles, and fixed camera installations. We synthesize our findings to examine the general effectiveness of transfer learning to landscape-scale image classification. Finally, we show how DCNN predictions on small regions of images might be used in conjunction with a CRF for highly accurate pixel-level classification of images.