Mapping Measurement Differences in Türkiye’s Rain Gauge Network Across the Manual-to-Automatic Transition: A Spatial Data Quality Perspective

Automatic gauges supply almost all daily precipitation data in Türkiye, while long climate and hazard series begin in the manual era. This study evaluates the automatic against the manual record at 308 co-located stations (2005–2024) using the ISO 19157 quality elements, maps the results with their reliability, and validates screens and transfer functions on unseen stations and years. On clean days, automatic gauges recorded 92.9% of the manual total, with the largest relative deficit below 5 mm and at high stations in winter. At high stations observed in both periods, the winter deficit had largely disappeared by 2019–2024, when station ratios no longer showed detectable elevation dependence or clustering. Maps showing reliability as transparency reveal that the Eastern Anatolian deficit cluster rests on less reliable records and belongs mainly to the transition years. Learned screens ranked missed events well (AUC 0.96, average precision 0.36) and potential spurious totals poorly (0.81, 0.22). A gradient-boosting transfer function restored the network totals and reduced the daily error but did not outperform a constant ratio for future station totals or heavy-day frequencies at unseen stations. For 2013–2024, the implied aggregate manual-equivalent adjustment is about 5–6%, and a station-level quality layer carries the measures with the data.