Welcome to Petroleum Geology Forums

Dear static modelers,

Can anybody help me with understanding the use of variograms in my facies modelling workflow. I have stared myself blind on the data analysis tab in Petrel and do understand what the Nugget is and what variograms indicate (variation in relation to distance). I understand that the sill indicates the spatial distance at which there is no more relation between the particular property.
However, when I change the search distance or radius the nugget and the sill changes all the time. Even when I can see in the little map inlay that no extra well locations are included in the calculation (the radius only expands outside the area in which we have wells).
Why is this occurring? And more importantly... which values of the nugget and sill should I use in the subsequence krigging? I think I understand the principles, but I can't seem to get consistent variograms fro my data. How should I interpret/use this?

thanks!

Janine

Janine,

I understand your confusion as the data analysis tab is a little bit difficult to understand. It is also difficult to asses what's going on in your model, but here's a try...

The range you see in the little map window (you know, the cone with the circular lines) is only displayed for one data point. This is done to keep the map readable. It may thus be that by expanding your range other data points are actually included in the calculation. The most southern wells, may no be compared to the most northern wells (if your trend in N-S). You can check this by clicking on the southern wells and checking their range.

The thing to remember is that your data analysis exercise should have some geological backing and simple variogram analysis is not a magic tool to get the best results. The right approach is that the variogram model must conform the geology. The effect of a variogram should be related partly on the data set but mostly on the knowledge of the geological setting and other external facts. This geological understanding should also determine your range, direction, etc.

For instance in high energy environments, like fluvial systems, facies and hence petrophysical properties change over short distances and there may be no relationship at all within tens of meters (consider a channel fill vs floodplain). In lower energy levels, like along the shelf, variograms ranges may be chosen to be larger as the chances of finding similar petrophysical properties over large distances is more likely.

Of course variogram analysis can also be a method for identifying these environments, but there are many other uses that I think should be mentioned here:

-Finding possible outliers (together with other tools; eg histograms)
-Detecting anomalies and level of heterogeneity (nugget effect study by vectors)
-Detecting changes in function due to scale (specially geological trends)
-Determining anisotropies (different variograms in different directions)
-Use it to quantify geology (the link between geology and maths/statistics)
-Combining variograms (eg. short scale heterogeneity with larger scale trends)

Did this help???