Add to ORKGOur ability to anticipate future large earthquakes depends partly on our knowledge of the past events. Yet this knowledge is poor as we lack methods to identify the past earthquakes, hence data that describe them. Our work had thus two objectives: 1) Improve one of the rare available paleoseismological methods, - that based on cosmogenic 36Cl exposure dating of seismically exhumed fault scarps; 2) Apply this method to acquire a dense dataset documenting the few past earthquakes on a few major target faults. We have thus developed a reappraised numerical protocol to properly model the 36Cl concentrations measured in exhumed limestone fault scarps. The novelty of the code is to include all the factors, as their uncertainties, that may contrib...