Rotation Madness and estimation screw-ups

What is it with rotation conventions and mining software providers?

Every company seems to want to use a unique, individual method of defining the rotation and then  provide a set of instructions guaranteed to confuse anyone who attempts to read them. I might not be the sharpest tool in the shed and I like to think I am not stupid but working out what rotation has been used to estimate a resource sometimes does my head in.

If you use Surpac, and if someone gives you a model  estimated using Vulcan the rotations discussed in the paperwork (assuming the author got it correct in the first place) are different. But it is different again in Datamine where it seems you simply make it up as you go along! I reviewed a model done using Datamine where every domain used a different rotation convention (Dom1 = 3:2:1, Dom 2 = 3:1:3, Dom 3 = 3:1:2). And when I tried to create a search ellipse to match these rotations I simply could not get one to match the smoothing trend I could see in the grade estimate (which certainly didn't match the rotations in the paperwork!). Conclusion was either the original author had poor record keeping skills, didn't have a clue what they were doing or I was simply unable to understand the intricacies of Datamine's search rotation rules (would not be a bad bet to place money on the latter!). I reviewed another estimate recently where you could see they simply screwed up, used a +ve rotation rather than a -ve on the Y axis rotation and pushed the search 90 degrees in the wrong direction (and then obviously never bothered looking at the blocks in 3D!).

Then you can always go to Isatis with its flash build you own (I know most software programs have this option but not many use it) with a bunch of rotations called Fred, Fred2, Datamine, Datamine2 (both of which do not seem to match Datamines own), GSLIB - which I am sure is correct because I have never been able to get the GSLIB rotation right in my own mind - took me years to workout the major axis is the X axis and not the Y axis like most other programs! They even have a Leapfrog rotation now - which somehow never seems to give me the rotation I get out of Leapfrog! Then there is Isatis Mathematical - a real do ya head innerer??

Why do we persist with this ridiculous pattern of arbitrary (read stupid) rotation conventions? When we are taught at Uni to take a measurement of a plane with an internal pitching lineation we take a strike and dip or a dip and dip direction, and then measure the pitch in the plane as a rotation from the horizontal. Only 3 programs I have come across offer you this option - and no I am not talking about the "But we offer you the tools to convert it!" option, I am talking about actually being able to enter a strike (Micromine/Isatis) or dip direction (Leapfrog) and a dip, with a pitch. I can take a field measurement where I might see the mineralisation sitting in the fault plane (say striking 340, dipping 70 to the NE with a pitch of 40 degrees from the North) and I can enter those details into either of these three programs and know - implicitly - that the directions are correct! Someone says "What search direction did you use?" and I can say "Striking 340, dipping 70 NE and pitching 40 from the north end" and they will know - implicitly - exactly what the search is, better yet they can visualise it - In their heads!! I do not have to prefix if with Vulcan rotation - or Surpac rotation or mathematical rotation, or ZXY LLR or any other ridiculous and meaningless convention. Please - when will mining software companies give use rotation conventions that actually make sense from a geological context rather than some ridiculous mathematical - artificial process dragged up from 30 years ago?

As an indication of what can go wrong I have created a search ellipse on a little no name gold deposit and ran a simple little ID2 estimate. I have used Micromine because I can use proper geological search definitions!

After assessing the data and obtaining a geological measurement of strike, dip and pitch I have entered it into the search parameters and checked it against the data. 

Here we are looking down the semi-major axis - "down dip" - alignment looks good

Here we are looking down the minor axis - normal to the dip - yep me thinks the plunge matches the general orientation of the grade.

Here we are looking down the plunge to the South East - nice little clumps of grade indicating we might be onto something!

All looks good, so I ran the estimate into a block model - just a very basic ID estimate with not a lot of effort put into refining the parameters.

You can see the grades tend to conform with the composite grades and search ellipse we determined above.

All good - BUT what happens if some Muppet screws up the search parameters using some stoopid mathematical rotation he (or she) got from some paperwork or a report and runs the estimate? I re-ran the estimate with all things equal and simply used a -69 dip rather than a 69 degree rotation on the Y axis (which then throws the third rotation out and changes everything!).

Damn - that don't look right - doesn't match what we expect at all!

Now said Muppet would probably not display the correct search ellipse in view like this as a validation step (it is amazing how few people do!) and if you just did some basic statistical checks you may not pick it up at all.

As you would expect the GT chart shows a difference - but not a huge change, without something to compare with you probably would not pick up an error (the wrong model is the lighter colours with the dots);

Fewer tonnes and lower grades because we have broken up our HG tonnes

In total metal terms that is a 6% fall in ounces - just because I screwed up the rotation dip, but again this is not a significant change globally if you had nothing to check against!

	Cut-Off	Tonnes	Gold g/t	Ounces
Correct Model	1 g/t	141,283	7.14	32,429
Wrong Model	1 g/t	138,121	6.89	30,589
	Percent Difference	97%	96%	94%

Wish I had grades like that in my normal life - instead of "Who Hoo, we got 2.8g/t!!!"...

This is the two models side by side (correct on the left, wrong on the right);

They are certainly different and one is clearly incorrect!

So you can see that locally there are some significant differences and this will impact any pit optimisation run!

So why did we end up with a screwed model - because we use stupid mathematical rotations that bear no resemblance to diddly squat in the real world. If said Muppet had simple read in the paperwork that the mineralisation was striking 024/69E/26S (or striking 024 deg dipping 69 deg to the east, pitching 26 deg from the south) or 69/114/153 (which is dipping 69 towards 114 pitching 153 around from the north end) I am sure there would not have been a screw up!

That doesn't mean you get rid of the mathematical rotations, keep em on the form such as that seen on the Micromine form for the sadomasochists ...

This is the Micromine form, look at that geological orientations!!!

I can enter the details I want - and wonder of wonders - all the other fields update automatically!! Wonder of modern technology hey! If I want to do my head in calculating multiple different rotations I can, for the rest of us normal people who just want to get out job done in the quickest, easiest and least painful way you can't go past the Leapfrog setup!

Got to love simplicity!

I challenge any geologist to NOT understand this form! I really don't mind the Micromine one - or the Leapfrog one, in either case I can intuitively understand what I am looking at! Please STOP using stupid mathematical rotations carried over from the days of Fortran77 or DOSv2.0; times have moved on people and computers aren't dinosaurs anymore!!

OH and for those interested strike\Dip\Pitch in mathematical axes convention assumes Y is the major axis, and is Z(Clockwise), Y(Clockwise), Z(Clockwise) - it is not hard!!

Happy Modelling!

Ron Reid, is the Group Resource Geologist for Harmony Gold in Brisbane, and has been contributing posts for orefind.com/blog for the last several years. This post, and many more like this, can be found at www.Orefind.com/blog and on Ron’s own blog at https://wantokgeoscience.blogspot.com.au/ . Comments contained within are the authors alone and do not in any way represent the opinions of Harmony Gold.