Tools and methods for constructing 3D geological models in the urban environment
|Location||International Geological Congress,oslo 2008|
|Author||Bourgine, Bernard۱; Prunier-Leparmentier, Anne-Marie۲; Lembezat, Carole۱; Thierry, Pierre۱; Luquet, Claire۱; Robelin, Christian۱|
|Holding Date||29 September 2008|
Urban development is a world-wide phenomenon that accelerates daily. The best way for successfully mastering the resulting modifications to the surface and subsurface of cities, and the interactions of Man with his natural habitat, is an excellent understanding of the geological environment on which the city develops. It is necessary to dispose over a 3D geological model of the subsurface, in order to be able to anticipate any difficulties or to give direction to the civil-engineering work to be carried out. Geostatistics propose a range of methods that help in the construction of 3D geological models, but many obstacles present themselves when working at the scale of a conurbation:
(1) The need to manipulate a very large number of data, e.g. several thousands of boreholes, imposes maximum automation in data processing;
(2) The common presence of redundant clustered - or contradictory - data, hinders rapid identification of similarities or inconsistencies between similar data in terms of context and geological hypotheses;
(3) The formations to be modelled have a strong spatial variability, because the part of the soil of interest to developers usually covers the surficial formations, such as fluviatile alluvium. The geological rules that govern the emplacement of such formations thus have to be integrated as much as possible, for a reliable interpolation of data in areas with too little information. Such rules cover type of soil-formation process (e.g. deposition, erosion, or unconformity) and the resulting geometric constraints (e.g. maximum height of an alluvial terrace, or position of the colluvium related to slope angle or type of ground);
(4) We are in a situation were most data are inequality constraints (not hard data), that have to be processed correctly: most boreholes do not intersect all formations to be modelled and provide only partial information on the position of interfaces.
We present a set of methods developed by BRGM over the past years, with the aim of resolving the above-mentioned difficulties where they are poorly resolved by habitual modelling. These methods, whether new or based on existing geostatistical methods, facilitate the automated construction of 3D models. They enable the systematic and automatic search for incoherent data, the taking into account of geologic constraints for constructing the model, the management of calculation parameters, and the production of standardized graphic documents. We illustrate the implementation of these methods with the example of the City of Paris, were we discuss the geometric aspects of the modelling. The successive data-processing phases are: data preparation, search for anomalies, homogenization of the data (boreholes, geological maps, DEM), construction of the different interfaces, incorporation of the geological constraints, and validation of the model. We also discuss the present-day technical limits and we suggest some research subjects that might further improve the results.