A methodological approach to modelling in geohistory. From the (spatialized) chronology of events to the system behaviour through spatio-temporal mapping Une méthodologie de la modélisation en géohistoire : de la chronologie (spatialisée) des événements au fonctionnement du système par la mise en correspondance spatiale et temporelle En Fr

Résumé En

Within the frame of the spatio-temporal analysis of natural hazards, geohistory is a meaningful approach to obtain and exploit (geo)chronologies. This article formalizes a scientific approach aiming at capturing the risk system in its entirety (factors, processes, and dynamics at play). It is based on systemic analysis, already promoted by geographers, but currently little used in geohistory, and it involves four successive steps: construction of a database of events, documentation of their socio-historical and bio-physical contexts, mapping between the events and their drivers, and finally determination of a qualitative model of how the system works. These steps are discussed and illustrated using examples from the literature. The review shows that numerous chronologies have already been proposed for natural hazards, but these cannot be directly interpreted in terms of the evolution of phenomena or risks. To apprehend these, it is necessary to replace documented events in their social and bio-physical context, that is to say, to analyze the main factors that can explain the evolution of their spatial and temporal distributions. The article proposes an operational checklist for this contextualization step. Once this is done, it is possible to consider the risk system in its entirety, each of its subsystems (the natural subsystem and/or the societal subsystem), and to confront the spatial and temporal distributions of events with the evolution of potentially explanatory factors. Spatial and temporal mapping allows concomitances or offsets to be highlighted and, as a result, provides explanations for the evolutions observed. A final step that identifies the main effects at play leads to a qualitative modeling of the risk system provided through a simplified graphical representation. The latter, even if it remains for now little used in practice and may be difficult to obtain, is a reflection and communication tool that is both efficient and elegant.