During my course of studies I changed the emphasis from geophysics to geology, not only because I couldn’t stand all the physics and maths lectures any more, but also because I thought that as a geophysicist I was less likely to do any field work than as a geologist. Back then we invented the word power-socket geophysicist for some of my friends that did modelling 24/7 for months. Now, about four years later my geophysicist friends do all kind of exiting fieldwork (e.g. week-long helicopter based surveys) while I am stuck at my desk sitting in front of a screen all day long! Let’s see what power-socket geologists do exiting.
Making old data look pretty (and more useful)
As a researcher of any kind one often uses old sources of information. In my case that includes reports and publications that can be several decades old and regularly display their data in a way that does not soothe the eye if you are accustomed to modern graphics. Thus re-visualising the old data can help to get a better idea what you are looking at and at the same time you have some nice figures for reports etc.!
I wanted to redraw a geological map of east Arizona from the 50s. The software of choice for creating (geological) maps is ArcGIS (or any other GIS software) as the map will be spatially referenced and it will thus be easier to add any spatial data later on. The process of redrawing the map it rather time consuming and when I halfway finished my new map I found a MSc thesis for which all the hard work had already be done! Anyway, the original is on top, the new map on the bottom.
Additional to the map I created a new stratigraphic column of the area. For such kind of work I like to use Corel Draw while others prefer Adobe Illustrator. Both programms are used to create vector based graphics. The advantage of vector graphics over raster graphics (graphics you create in Paint and similar software) is that you can use the same image at any size without losing any clarity of your image. This is especially helpful if you are planing on actually printing the graphics. Here is the result:
On of the main problems of many geological (research) questions is that there is on limited information on the subsurface available.While the mapping of rocks in outcrops gives a distribution of rocks on the surface, the exact behaviour of rocks in the subsurface can only be extrapolated. There are two kind of information that can help to get a better grasp on what the subsurface looks like: Indirect data and direct data. Indirect data often comes from a range of geophysical methods such as reflection seismology. This kind of data needs to be thoroughly interpreted and there are often more then two explanations for the same result! The only way to get direct information on the subsurface is to drill a hole and have a look what is going on. However, drilling holes is quite expensive and no matter how many holes there are, between boreholes information is still missing and has to be interpolated (modelled).
For my recent project I want to know how the subsurface of a certain area looks like. As you can see above, I already have an idea how the surface geology looks like and what kind of rocks are likely to be found in the depth. Luckily, there are several boreholes in the area as well of which the well logs have been released which I use to model the geology of the subsurface. The software I am using is Petrel, an environment often used for by exploration and production departments of oil and gas companies. While it is very powerful and has loads of different modules that do all kind of exiting stuff, it is also quite complex. Most of the work I have done so far in Petrel is self-taught and I am still trying to find a good guide for it (if you know where to find one, let me know!).
To create a model in Petrel is (relative) simple: Load all necessary data into the program, adjust the modelling variables to your needs and wait for the program to calculate the horizons for you! Here you can see the input parameters:
The vertical lines represent the wells used as input parameters, the spheres in the wells correlate to the depth at which certain geological layers have been found in each well. The yellow rectangle illustrates a fault that separates the NE part of the area from the SW part. The 3D model created from these parameters looks like this:
While this looks quite nice the information content of this illustration is rather low. A better way to display the results is to create depth maps for single horizons or to create cross sections through the model:
My Petrel model is still highly inaccurate and needs loads of work but the first results are quite promising! Things I will try to include in future models are the surface geology and more and more complex faults. So far I haven’t been able create a model with more than one fault (help!).
The next blog entry will hopefully show some more advanced geological model and might even include some things about statistical modelling in R, something that is on my agenda for October/November.