A multi-method dataset of shoreline position at Narrabeen-Collaroy Beach from sub-daily to decadal timescales (1930 to 2021)

This repository contains a 91‑year, multi‑method shoreline dataset for
Narrabeen-Collaroy Beach, New South Wales, Australia - one of the world’s
most intensively monitored sandy coastlines. The dataset comprises 10,334
individual shorelines (>2.5 million shoreline data points) spanning
1930–2021 and integrates 13 complementary survey techniques, including
historical and commercial aerial photography, Landsat and PlanetScope
satellite–derived shorelines, Argus coastal imaging (near‑daily coverage
for 19 years), RTK‑GNSS quad‑bike surveys, airborne lidar, a continuously
scanning terrestrial lidar transect, and community‑sourced CoastSnap
observations. All shoreline positions are referenced to a consistent
vertical datum and transformed into a unified alongshore–cross‑shore
coordinate system that accounts for the embayed (log‑spiral) planform.
Method‑specific uncertainties (bias, standard deviation, and RMSE) were
quantified against high‑accuracy RTK‑GNSS and airborne lidar benchmarks
and are provided per shoreline point to enable uncertainty‑aware analyses.
Shoreline data are distributed as a single CSV file
(Narrabeen_Shoreline_Dataset_v1.csv) with fields for alongshore
coordinate, cross‑shore position, UTM (EPSG:28356) eastings/northings,
timestamp (AEST), survey method (codename), and per‑point accuracy
metrics; flags indicate tidal correction availability and whether
alongshore interpolation was applied. Two accompanying time series -
bias‑corrected ERA5 waves (Bias_Corrected_ERA5_Waves_Sydney.csv) and
astronomical tides (Astronomical_Tide_Sydney.csv) for Sydney - are
included to support process studies and model forcing. This dataset offers
unprecedented spatiotemporal coverage (sub‑daily to decadal) suitable for
investigating storm‑scale erosion and recovery, long‑term shoreline
variability and climatic forcing, validation of remote‑sensing algorithms,
development/benchmarking of numerical and data‑driven coastal models, and
multi‑hazard coastal risk assessments. A companion Jupyter notebook
(GitHub) demonstrates data access and common analyses. Please cite this
repository and the associated data descriptor when using the dataset.