Questions from the Field ...

“How should you Cruise Clumps of Stems?”

Sprouting trees are a particular problem. Maple, as well as Redwood come to mind (young Redwood being very nasty because it looks like a hedge anyway and you can't see the individual stems in many cases - one of those times when fixed plots make sense).

When you can see the stems, however, how should clumps be handled? First, it is not a great problem with count plots. That's another reason to use them more in your work. Second, you might not want to do all of the small stems individually. Nothing wrong with that, of course, but it gets pretty tedious and for little benefit.

After determining the stems that are in with your prism or Relascope, you can just sub-sample. If you have 7 of them in with the prism, choose one of them to measure. If you choose 1 of the 7, then it just gets written down 7 times (if your compilation cannot handle a directly weighted sample). This approach will always work with any compilation system, as long as it does not kick back an error because several of the items are identical.

In some compilation systems, individual trees have their own BAF's and you could also write down a BAF 7 times as large for the one tree out of 7 that you sampled. You would have to verify that this works with your own compilation system. It should - but you never know unless you can check the computer code. You ought to know the phone number for the computer jock that is supposed to know this stuff. If you have not done anything to thank him for his previous help, of course, he might not care as much (hint, hint).

It is best to do this tree selection from the cluster randomly if you are to be check-cruised or are working for someone else. If you are doing your own work you could always choose a biased sample that you considered typical of the group. If you are wrong, you are the only one who suffers.

The rules of statistics can always be suspended when you are risking your own money. In a case like that you are substituting the guarantee of your professional judgment for the guarantees that proper sampling provide.

This kind of sub-sampling is automatic with the "Big BAF method", which might make much of this problem go away for practical work.

The variability that clumps like this cause is likely to be restricted to a particular species (or at least a size class within that species). If it is a serious problem with your work, you might consider putting small trees that occur in clumps into an individual strata of their own. This will isolate the problem from the rest of the data. You might have a high Sampling Error % in that group, but it will apply to a small volume and value.

PS : our thanks to Kevin Oliphant, who asked this question.