Hydrogeological model predictions are used by managers to define groundwater resource management policies. Long ignored or neglected, the problem of uncertainty about model outputs is essential: how to make a decision with a tool whose reliability is either unknown or very strong?
Title: Hydrogeological modeling: should physics be sacrificed to manage uncertainty?
Cost: Free
Time: 11 am to 12 pm
Location: Salle 4108, Pavillon Adrien-Pouliot, Université Laval, Québec (Québec)
Speaker: Alexandre Pryet, assistant professor, École nationale supérieure en environnement, géoressources et ingénierie du développement durable (ENSEGID), Université de Bordeaux
See the video of the presentation »
Event in French
Hydrogeological model predictions are used by managers to define groundwater resource management policies. Long ignored or neglected, the problem of uncertainty about model outputs is essential: how to make a decision with a tool whose reliability is either unknown or very strong?
Estimating the parameters of a model and quantifying the uncertainty implies having models whose calculation time is relatively short. Should we sacrifice physics to reduce computing time? For those who choose to quantify predictive uncertainty, the operation often reveals unpleasant surprises. The acquisition of various observations such as recharge, piezometry, and the geochemical signature of water then becomes necessary to constrain the parameters, reduce the uncertainty, and make the model operational. This approach will be illustrated on real cases.
