3D geological model for Bucharest
Description
This dataset contains rasters of the 3D geological model for Bucharest, elaborated by Toma-Danila D., Bala A. and Ciugudean-Toma V., within the framework of a 2022 paper which is under review. Rasters contain information regarding base depth of the first 6 principal Quaternary complexes beneath Bucharest and thicnkess of these layers. There can be also found rasters clipped to the area of high confidence (where the model error is the lowest) - available as shapefile. The model was developed based on borehole data from S.C. Metroul S.A. and from the NATO-SfPProject 981882 (original data availability can be seen in a shapefile). The interpolation method was kriging. For the model, a recent digital elevation model developed by ANCPI in 2022 was used (MNT; https://geoportal.ancpi.ro/portal/home/).
Files
Steps to reproduce
In total, we used an amount of: • 969 point values for Layer 1; • 349 point values for Layer 2; • 340 point values for Layer 3; • 331 point values for Layer 4; • 274 point values for Layer 5; • 59 point values for Layer 6. The final kriging parameters chose in ESRI ArcMap 10.6, for the generation of rasters, were: - Simple kriging with Normal Score transformation type, with decluster before transformation (given that our data points are highly dense in some areas but limited in others) and a constant order of trend removal; - An optimized model, obtained through ArcMap capabilities; - Cell declustering method, with cells rotated 45 degrees to match the trend of geomorphological features on Bucharest map; - Student’s t base distribution; - 4 Sectors with 45 degrees offsets, a minimum number of neighbors of 3 and a maximum of 5, in Neighborhood Search. After interpolation, a check for lower layers piercing above layers was performed and, in some cases, values were adjusted. This was expected, given that geological characteristics highlight some layers to disappear in certain. As such, the layer thickness, when needed, became 0 for some layer areas (mostly for Layers 1-4). In order to show areas where there is high confidence in our 3DGM, we computed also prediction error values. Where the errors are lower than 10% of the difference between maximum and minimum values of a specific layers – these area is characterised as “area with values of high confidence”.