Automated Digitization and Integration of Legacy Geophysical Maps: A Reproducible Python Workflow for Analog Archive Recovery Applied to the Southern Marmara Sea, Türkiye

In many national geophysical surveys performed during the period from the 1950s to the 1990s, large analog archives are still not accessible to modern analysis techniques because of their printed, non-digital form. This paper presents a Python-based, fully automated workflow for the systematic digitization, georeferencing, and multi-source integration of legacy gravity and magnetic survey maps, applied to a critical geophysical archive in the southern Marmara Sea region of Turkey. The study analyzes 27 printed maps, including 15 gravity and 12 magnetic maps, from 1972 to 2000, created by three different agencies (MTA General Directorate, C&K Petroleum, Norexplo Consultant) using two self-developed Python modules. The fully digitized database contains 24,510 Bouguer gravity measurements (86 survey lines) and 25,230 reduced-to-pole (RTP) magnetic measurements (87 survey lines), amounting to 49,740 integrated data points. Cross-validation for the gravity data set gives r = 0.9996 and RMS = 0.426 mGal, and for the magnetic data set r = 0.9991 and RMS = 7.485 nT, with mean residuals close to zero in both cases, indicating the absence of systematic errors. The reconstructed Bouguer gravity field varies from −37.47 to +74.01 mGal, and the RTP magnetic field from 746.5 to 2085.6 nT over the study area (27.0–29.4°E, 40.2–40.7°N), resolving anomaly patterns at a resolution of 150-200 m, roughly 10-20 times higher than the resolution of global maps available so far. The spatial coverage is above 99% for both data sets. The workflow reaches an accuracy of 94-97% in contour recognition and saves up to 80-90% of processing time compared to manual digitization. The fully open-source Python code is made available for the reproducible recovery of printed archives on a global scale.