Questions from the field...
“Trees that change category in nested plots”
Here is a question Jason Teraoka asked (sorry about the delay – website problems)
I have a question regarding nested plots in permanent fixed-area plot sampling.
Do you change expansion factors of the trees in the nested sub-plot over subsequent inventories when the tree records are large enough (in dbh or whatever the criterion is) to be recorded in the main plot?
If that doesn't make sense, here's an example.
All trees greater than 10" dbh are measured in a 1/10 ac plot (the main plot).
All trees 10" dbh and less are sub-sampled in 1/100 ac plot (the nested plot).
Let's say at year zero we observed 10 trees in the 1/10ac plot (exp factor = 10) and 10 trees in the 1/100ac plot (exp factor = 100). So, (10*10)+(10*100) = 1,100 trees per acre.
We re-inventory the plot 10 years later and find that all the trees in the 1/100 acre plot are now larger than 10" DBH. Do the expansion factors for those trees, from this point on, now change to 10 to reflect that the trees have ingrown into the 1/10 acre plot? Or do you keep the original expansion factor of 100?
Sorry if this is a silly question!
Ohhhh no … it’s not a silly question at all. It’s a tough decision, though. Here is at least one person’s opinion, from a mind subject to change by actual data.
If you change plot size, you get sudden increases or decreases. This happens even when there is no change in the item measured. I have always called this the “plot size recalibration”. This is just the opposite of what you want on long term growth plots – you want to measure a gradual change that indicates what is happening for a slowly moving trend like growth There is no easy solution to this that we know of. If anyone has sufficient data (over 4 remeasurements, perhaps) to show the size of the effect here we would love to see it.
If you are trying to stabilize the result from the number of trees per acre, you should probably keep the plot sizes constant. The situation here is that you chose the trees with a particular probability, and keeping that probability will minimize sudden change. How you initially decided on that probability is not the issue. Sudden change is the issue.
As an alternative, consider using Variable Plot Sampling for your permanent plots. It will smooth out some of the effects you are computing – and you will not have the sudden “jump” in results when the plot size changes. The tree plot size gradually changes as the diameter does, and it minimizes this issue – and in addition puts the best estimate of change on larger tree sizes (usually, but not always, an advantage).
It will probably help to use the “Critical Increment” estimate of change for anything you are trying to measure. This procedure, from about 1980, was the essential solution to most of the growth “problems” with Variable Plots used as PSPs. There is still the problem with mortality, and nobody has yet come up with a good solution for that issue.
Originally published January 2018
Question from the Field