Question: From the help files it appears that the .aidv format is intended for groundwater data that is collected across screened intervals. I’m wondering if it is also an appropriate format to use for subsurface sediment data that is also collected across intervals (i.e. x, y, z_top, z_bottom, concentration).
I have also looked at the .apdv format. It appears that this format will associate each sample’s concentration to a single z value rather than across a z interval.
To complicate things, I understand that to do cross validation via Python I would need to use the .apdv format.
Please let me know if I’m on the right track here. It makes sense that I would only be able to cross validate using the .apdv format, but we would prefer to use the interval-based data for our 3d interpolations of chemical data in the subsurface.
Kriging uses data samples at points. (emphasis on the period)
Our aidv format allows you to view data as tubes with the appropriate interval distance, however, when you krige with aidv data, we use the max gap parameter to determine how we convert the intervals into one or more points.
The reason I have previously said to use apdv for cross-validation with the process proposed, is that you would want to drop all points associated with an interval. If you drop an entire screen, you could do cross-validation with aidv dat as well as apdv data.
Remember that if you use too small of a Max Gap parameter, you are creating a situation of oversampling which will result in kriging artifacts and problems which our Data Reduction Tool was designed to mitigate for data such as MIP.
However, if it is too large, a long screen may not have a constant concentration enforced over its length. Unfortunately, large screen intervals are themselves a problem in that the concentration is highly unlikely to be constant over the length but will be averaged and “smeared” by its nature.