This package contains R codes and data related to the article: B. Renard. 2023. Use of a National Flood Mark Database to Estimate Flood Hazard in the Distant Past. Hydrological Sciences Journal. DOI: 10.1080/02626667.2023.2212165 CONTENTS R scripts used to set up models, analyse results and prepare figures. France207_MAX.RData contain monthly maxima at 207 stations. See references below for the original sources. Folder data_raw/ contain exports from the flood marks database. See references below for the original sources. funk.R is a library of functions used by other scripts. 0_setUp.R prepares the runs to estimate the model for one component. Estimation is performed using the RSTooDs package (https://zenodo.org/record/5075760). This script extracts the data, specifies the probabilitic models and write STooDs configuration files. 1_margin.R analyses the runs that have been performed and estimates marginal distributions at each station. Note that the runs set up in the previous point need to have been performed before using this script. 2_probMap.R computes estimated flood probabilities over the whole 1705-2015 period. The previous script needs to have been run before. 3_sensitivityAnalysis.R prepares configuration files for sensitivity analysis experiments. 4_ZenodoRelease.R prepares the released data and probability maps. The previous scripts needs to have been run before. Scripts named fig_XXX.R create the figures shown in the article. REFERENCES Streamflow series at stations: https://www.hydro.eaufrance.fr/ developed by the "Service central d’hydrométéorologie et d’appui à la prévision des inondations" (Schapi) from the Ministry of the Ecological Transition Flood marks database: https://www.reperesdecrues.developpement-durable.gouv.fr/ developed by the flood forecasting network "Vigicrues" from the Ministry of the Ecological Transition