A recent visit to a client site brought to light a somewhat common misconception about how accurate their data is (or was). For reasons which will become obvious I will not be naming anyone involved. And I’m going to simplify the following to an extent simply because I don’t have the time or space to indulge in creating the equivalent of a post-secondary course on land surveys.
The problem was centered around the issue of Well placement on the map. The client was facing some issues when planning a directional well program in a fairly tight well spacing area. At one point in the meeting the client said “I know my data is accurate, it is measured from the survey grid to a fraction of a metre.”
Well, no. Your data is precisely measured from a survey grid, however the real question is “How accurate is that survey grid?”.
Let’s talk about the difference between accuracy and precision. While the terms are frequently used interchangeably, they really describe two different things, especially when you get into a technical realm.
At it’s simplest distillation, accuracy = truth. The question “How accurate is your data?” translates to a question about how much can I rely on your data – “How true is the information that you are providing me?” In a very real way, accuracy is a qualitative metric.
When we talk about precision then we are talking about a quantifiable level of detail. The question “How precise is your data” is a request for some sort of metric that could be used to compare one source of data with another.
When we get to well positioning or well placement on a map, then we are entering a realm that can be defined by aspects of both precision and accuracy.
In Western Canada wells are licensed based on the bottom-hole location, whereas in the US generally the well license or permit is assigned based on the surface-hole location. In all cases the coordinates that define the location are based on a survey that is tied to known markers. In Canada those survey coordinates are then referenced back to a survey grid (the same is true of many regions in the US and elsewhere in the world) to derive measurements from a known corner of the survey grid (metes and bounds in Canada, footages in the US).
The survey grid is administered by the Provincial government (for example, in Alberta the grid used by the Oil industry is the Alberta Township Survey Version 4.1[ATS 4.1]). As the name implies, the ATS grid has gone through several revisions and recalculations over the decades to reach a version number 4. Each revision has resulted in corrections to previously derived grid points with the result that each successive grid is more accurate than the one that preceded it.
The Provincial government include a statement of accuracy, as well as a technical discussion on how the grid is derived and maintained. For the ATS 4.1 file, the published accuracy is +/-3 metres (about 10 feet).
Now, if you look at a typical well Ticket, the metes and bounds for a well location are expressed in metres to one decimal place, so a well may show metes and bounds of N 435.6, W 353.4. This means that the bottomhole location of the well would be offset by those metes and bounds relative to the Southeast corner of the Section in that Township.
So, how precise are the coordinates? The precision is 0.1 metres (about 4 inches).
How accurate are the coordinates? The accuracy is +/- 3 metres, based on the grid that is used to derive the metes and bounds.
So for the client who was certain of his well placement to a sub-metre accuracy, that would translate to an actual 6 metre (20 foot) cone of uncertainty based on the published accuracy of the grid that he was trying to tie this all to.
Just to add that tiny bit of extra interest to the situation, the grid the client was using was based on NAD27, while everything is actually licensed based on NAD83. In the case of geoSCOUT, we convert all of the NAD83 data to NAD27 and back again as required, but the client was trying to use the data in another piece of software and was trying to tie it into data from a 3rd party source. Unfortunately the client was unsure whether the 3rd-party data was in NAD27 or NAD83, so there was the potential of misconvesion of one of the data sources.
And since the wells were going into 2 adjoining Sections, things were actually a little more uncertain, since each section had it’s own corner coordinates with a separate +/- 3 metre accuracy…
Update July 22, 2011 – Someone pointed out a small spelling error that I fixed.