As a geophysicist, it is very important to develop the habit of not trusting a contour map until I've seen the data distribution. Even a single suspicious blip on the contour map, I would take the trouble to look into the issue to the extent of checking and verifying the field data.
I've observed numerous times and again, this common mistake of ignoring data distribution. Very common amongst newbies and those who are not trained in geophysics but are are engaged in carrying out geophysical surveys. Most of the time they trusted and accepted the outputs from the software without giving any thought to the data distribution. I have seen someone generating contour maps with only two survey lines located quite a distant from one another !
The image below shows another example from an ERT survey I personally completed recently. Earlier, my focus was only on the inverse resistivity section. Since I realized the importance of these pseudosections, I've made it a practice to include them when presenting ERT results. . It was obvious that there were bad data removed as observed on the measured and calculated apparent resistivity pseudosections. However, this was not reflected on the inverse model resistivity section. The question is : how much trust and confidence do you have on the inverse resistivity model, especially at the section where data were missing ?
Do not just simply trust and accept what the software is churning out. Pay special attention to data distribution; do not ignore them as they play a crucial role in deciding whether or not to trust the outputs.
It goes without saying that if you are into potential field modeling and inversions, you still need to have a very well distributed data set. f your data are sparse and not well distributed, the models would be very questionable.
English
