Unlocking Historical Sea Level Records through Digitisation—A Review and Inventory

Christian James Kenwright, Ivan D Haigh, Joanne Williams, Robert J Nicholls, Duo Chan

Multi-decadal to centennial sea-level time series are essential for quantifying historic trends in mean and extreme sea levels, refining future projections, calibrating coastal models, and supporting coastal management and adaptation strategies. Data archaeology can increase the spatial coverage and length of timeseries through the recovery of historical observations from paper-based records. However, locating, scanning, digitising, and quality-controlling these records remains labour-intensive. Here, we review published sea-level data archaeology studies and compile a global database of historical tide records identified within them, alongside an inventory of tools available for digitising paper-based datasets. Across 53 published studies, 872 historical sea-level records were identified (representing ~15,264 years of sea-level data). Of these, 241 records have been digitised (~7,823 years of data), while 631 remain undigitised (~7,441 years of data). Most of these records are ledgers (systematic handwritten or printed logbooks of sea-level observations at discreet intervals); there are also marigrams (plots showing continuous records of sea-level variations over time). Manual digitisation remains the dominant recovery method used in most studies. Only one study applied citizen science to capture historic records, and 14 studies used semi-automated algorithms. A broader review of data archaeology across other disciplines identified 20 tools with potential application to digitising sea-level records. By compiling identified records, past recovery efforts and available tools, this inventory provides a foundation for more efficient, consistent, and strategic recovery of historical sea-level data in the future. We aim to update these databases as new records are identified or digitised and encourage researchers to contribute information on records not yet included, helping to improve the completeness of these databases.