Colloquium

Colloquium internship ERIC VERHOEFF

For this internship a tool was build to semi automatically create a 3D-model of the subsurface based on available borehole descriptions from DINO loket. Different interpolation techniques are applied and tested.

Organisator Leerstoelgroep Hydrologie en kwantitatief waterbeheer
Datum

vr 14 juli 2017 11:00 tot 11:30

Locatie Lumen, building number 100
Droevendaalsesteeg 3a
100
6708 PB Wageningen
+31 317 481 700
E112

 

Since the beginning of civil engineering the importance of geology for construction is acknowledged. The development of GIS rapidly increased the ability to model and visualize the subsurface geology and subsurface fluid dynamics. In the Netherlands, the government is since 2008 obliged to digitize all subsurface data and make it online available for free use. One of the data products that is offered is REGIS 2, an hydrogeological model of the subsurface of the Netherlands. Although the model has been used successfully for a variety of purposes, the model also has shortcomings. The model offers no flexibility to the user to adapt it to specific cases and the newest data. Recently demand arose for new methodologies that allow flexible subsurface modelling. Users want to be able to easily create models for specific tasks while using as much data as possible. For this reason this study developed a workflow for the (semi)automatic creation of hydrogeological models of the subsurface based on all available data in the Dutch subsurface database DINO loket. The developed workflow was tested by creating an hydrogeological model of the subsurface in Woensdrecht, the Netherlands. Four different interpolation methods (Ordinary kriging, bilinear interpolation, bicubic interpolation and nearest neighbour) were used and their performance was tested. The statistical parameters used to assess the quality were the mean error (ME), mean absolute error (MAE) and the root mean square error (RMSE). The resulting interpolated grids were also evaluated on shape, size, mean and standard deviation, to compare which one resembled typical layers of sedimentary deposits in the Netherlands the best, based on expert judgement. During the presentation de functioning of the tool will be demonstrated and the results of different interpolation.