Validation of Semi-Automatic Extraction of the Pluviographic Signal Using NUNIEAU: Application to Nine Synoptic Stations in Côte d’Ivoire (West Africa)

The rescue of paper pluviographic archives is necessary, but the validation of semi-automatic digitization algorithms for pluviograms remains limited in terms of the number of climatic contexts and types of pluviograms tested. This article applies to the Ivorian pluviographic network. Nine synoptic stations were selected. They cover the four climatic regimes of Côte d’Ivoire. The data come from the archives of the Direction de la Météorologie Nationale of Côte d’Ivoire (DMN) and cover the period from 1955 to 2023, with gaps of varying importance depending on the station. The time series extend from 34 to 55 years. The NUNIEAU software (version 5.01) was used to reconstruct heights at sub-daily time steps. These steps are compatible with risk analysis on small catchments, as is often the case for urban catchments. The daily height totals estimated by NUNIEAU were compared to the visual readings over 24 h performed on the same pluviograms. A total of 31,310 pluviograms containing at least one rainy episode were digitized. Beyond the overall statistics (R2 ≥ 0.97 and mean bias < 0.3 mm), the analysis targeting extreme events using physical rainfall thresholds (20, 50 and 100 mm) demonstrates robustness for cumulative totals ≥ 50 mm, the threshold set by the DMN to define extreme events; the absolute bias remains below 0.5 mm for eight out of nine stations, and the coefficient of determination remains ≥ 0.94. These results confirm that semi-automated extraction with the NUNIEAU software correctly restores daily heights. They therefore validate the use of the NUNIEAU software for the Ivorian climatic context (4 climatic regimes) and for pluviograms with curvilinear reference marks. The use of this software is therefore to be promoted for the rescue of paper pluviographic data threatened with disappearance in West Africa, in the continuity of international data rescue efforts carried by the Data Rescue (DARE) initiative.