Testing geological conceptual models with GPR simulations
Abstract
We present here a method for testing conceptual models of geological heterogeneity. This method is motivated by the modeling of the Tagliamento river, Italy, for which surface GPR data and sedimentological observations are available. Modeling heterogeneous aquifers issued from highly dynamic environments such as braided rivers systems requires some strong hypothesis concerning the morphology of the heterogeneities. Object shapes and sizes in object models, Training Images in Multiple Points Statistics models, or rules within process-based methods are some examples of these assumptions. These hypotheses result in smoother or stronger lithological contrasts in the simulated media. GPR data offer the advantages of underlining porosity and therefore hydraulic conductivity contrasts and being a non-intrusive method. With some treatment it can be used to identify main objects structures. By running forward GPR simulations on the simulated geological media, and by comparing the resulting patterns with the post processed GPR field data, we can test the validity of the prior conceptual model. To illustrate the method, we used different scenarii of a pseudo-genetic algorithm embedding DS simulations to obtain different types of geological heterogeneity. The GPR response is simulated for each geological model scenario. Post processing of those images are then statistically compared with those from the Tagliamento dataset, to accept or reject the conceptual model implied by the scenario.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2012
- Bibcode:
- 2012AGUFM.H33A1276P
- Keywords:
-
- 1825 HYDROLOGY / Geomorphology: fluvial;
- 1847 HYDROLOGY / Modeling