3P Sampling

"How do you handle cull trees in 3P sampling?"

Published January 1998

This is an excellent, and practical question. Whether the trees are live or dead makes no difference in the general approach.

One basic problem is that cull trees have a zero estimate, and are therefore never checked. The solution to that problem is to give everything at least some estimated volume.

The causes a problem unless you put these trees into their own strata. Create a strata for "cull trees," meaning that they have are expected to have no net volume (or a very low percentage of net volume).

What should you use for an estimate for these trees? Well, the only important point is just to make sure that you do NOT use the net volume as an estimate (KPI) for that tree. Use the gross volume, or just give them all a constant percentage of the gross volume (say, 25%). The reason for this is that you do not want to start dividing by very small numbers when you get a sample tree ratio. This makes the ratio very unstable.


 Next Column

The ratio will not be particularly consistent in this strata, but this in not a big problem, since the net volume

In this strata is so small that it cannot contribute very much to the overall sampling error of the net volume for the cruise.

This is an example of an important principle. If you cannot get rid of variability by doing a "better job of estimating," you can always confine the problem to a particular stratum – thereby limiting the damage.

Our thanks to Rex Friend, in Alaska, for bringing this to our attention. We greatly appreciate these practical questions. The ability to ask good questions is critical. You can always find someone to answer your questions if you can figure out what those questions are.

Return to Home
Back to Contents