Coastal monitoring has historically been expensive, sparse, and slow. Satellite imagery covers vast areas but lacks the temporal resolution to catch day-to-day sand movement. Ground-based LIDAR surveys are precise but cost thousands per deployment. CoastSnap, launched in Australia in 2017, routes around both constraints by turning casual beachgoers into a distributed sensor network — one phone cradle and a sign at a time. The mechanism is elegant in its simplicity. A fixed cradle ensures each photo is taken from the same position and angle. Software then aligns the images, performs a geometric transformation to produce a bird's-eye view, and an algorithm maps the shoreline position, correcting for tidal state using the timestamp. No training required from participants. No expensive hardware. The hard work is in the back-end image processing, not the data collection. The network now spans more than 200 stations in 21 countries, making it the largest coordinated land-based coastal monitoring programme in the world. That scale matters because erosion is intensely local — what happens to one beach after a storm tells you nothing about the beach 10 kilometres down the coast. You need density of observation points, and density costs money unless you crowdsource it. The data serves multiple purposes: tracking long-term erosion trends, assessing damage from individual storms, and evaluating whether beach replenishment schemes — where authorities dump sand or shingle to rebuild a coastline — actually work. That last use case is particularly valuable. Replenishment is expensive, and without sustained monitoring, authorities are essentially guessing whether their investment held. Large waves and high tides strip sand from beaches; calmer conditions rebuild them. Sandbars shift shape and position with waves and tides. Some changes are temporary, some permanent. The value of CoastSnap is converting these observable-but-unrecorded dynamics into longitudinal datasets that researchers can actually analyse. The generative logic here is textbook: take a task that was previously locked behind expensive professional equipment and institutional budgets, reduce the marginal cost of each observation to near zero, and distribute the workload across willing volunteers. The data generated is a genuine public good — it doesn't diminish when shared and it serves multiple downstream uses. The fragility is also real. Citizen science projects live and die by sustained participation, and participation depends on the cradles being maintained, signs remaining legible, and the novelty not wearing off. The project's durability will ultimately depend on whether the data becomes load-bearing for planning decisions — once councils rely on it, they'll have reason to maintain the infrastructure.
