z tolerance in "Raster to TIN" tool

[MangeantBaptiste]

Hi !
I want to test different values of z tolerance with the "Raster to TIN" tool on an 1-meter-grid raster. I obtain acceptable TIN when the z tolerance is equal or higher than 1 unit. But when i test 0.5, 0.9 or even 0.999 values, my TIN becomes completely irrational (for example, the first half of my TIN is a perfect regular TIN with 1-meter edges, but the second half does not represent nothing, with 200-meters-edge triangles). It seems that 1 is a limit value for z tolerance, is this correct ? What can i do to test lower values than 1 ?
Thank you for your attention.

Mangeant Baptiste
French Engineering Student

[curtvprice · Accepted solution]

  I obtain acceptable TIN when the z tolerance is equal or higher than 1 unit. But when i test 0.5, 0.9 or even 0.999 values, my TIN becomes completely irrational (for example, the first half of my TIN is a perfect regular TIN with 1-meter edges, but the second half does not represent nothing, with 200-meters-edge triangles). It seems that 1 is a limit value for z tolerance, is this correct ? What can i do to test lower values than 1 ?

Are you sure you have this much Z accuracy in your elevation data? Is this going to create more than few million nodes?

If this is indeed a bug with the tool and not your data, you could scale your data by using a Z factor. This will scale your values by 100 so you can use smaller tolerances, for example, if the z_factor is set to 100,  a z_tolerance of 50 would represent 0.5.

[MangeantBaptiste]

Indeed, I tested this method (taking a z-factor of ten and a z tolerance value of 2) and it works well. I just wanted to understand the limit in the z tolerance value. But effectively, it seems that my z accuracy is around 1 meter, which creates more nodes than my limitation of points (1500000).
Thank you for your advice.

[curtvprice]

You may want to look into terrains, which can handle larger numbers of points and do this conversion on the fly when needed.