Putting Big Tree in Context

Putting Big Tree Size in Context

== 2 of 4 ==
Date: Sat, Dec 1 2007 3:27 pm
From: doug bidlack

I agree that the trees need to be mature. I'm not
sure where you'd draw the line though. It seems like
alot of the old growth forests that I've visited have
many really big trees and alot of really small ones
waiting for their break. I'm thinking of a forest
floor in the porkies that is mostly little sugar
maples and most of the remaining sugar maples are way
up in the canopy. At least that's what I remember.
If that's true, then maybe it works best for these
type of forests.

Doug

== 3 of 4 ==
Date: Sat, Dec 1 2007 7:22 pm
From: dbhguru

Doug,

I've been thinking of an appropriate human analogy whereby a comparison between an individual and either the tallest or some kind of average is made. In an apples to apples context, an adult would be compared to a subset of all adults. But the comparison possibilities appear endless. I like the notion of a comparison of an individual to a mean and to a maximum.

Bob

== 4 of 4 ==
Date: Sat, Dec 1 2007 9:09 pm
From: doug bidlack

Bob,

Yes, I like both comparisons as well. I was also
thinking of a human analogy. I was thinking of a
group of say 50 men and 50 women. Lets say the mean
height of the women is 5'4" and for the men 5'10" and
the tallest woman is 5'10" and the tallest man 6'4".
Seems likely, but what if 5 more average sized men
walk in and 5 women that play professional basketball?
Lets say the maximum and mean male height doesn't
change, but the mean female height goes up slightly
and the max hits 6'5". Looking at the means and
maxes, you can say that the men are, in this
population, taller on average, but that the tallest
woman is taller than the tallest man. You can also
say that the tallest woman in this population is much
more exceptional than the tallest man, because you
know how much her height deviates from the mean.

Doug

== 2 of 7 ==
Date: Sun, Dec 2 2007 12:44 pm
From: doug bidlack

Ed,

As long as you adequately define your population and
you randomly select individuals within that population
to measure, you will have obtained a valid mean
(average). Trees aren't the only organisms that never
stop growing. Before becoming an entomologist, I
seriously thought about becoming a herpetologist. I
especially love snakes. Snakes also never stop
growing, though they tend to slow down quite a bit as
they get older. Herpetologists measure sexually
mature snakes all the time, and the mean lengths that
they obtain are perfectly valid. You could narrow
your population by only measuring, for example, beech
trees that are part of the canopy in a particular
forest.

Doug

== 3 of 7 ==
Date: Sun, Dec 2 2007 1:11 pm
From: dbhguru

Ed and Doug,

Doug, I think your point is an especially important one to ponder. Also, I understand where Ed is coming from. I've thought a lot about averages as they might relate to a particular species and find that a lot of qualifying has to take place. White pine is the perfect species to speak about averages versus maximums and where comparisons to averages provide added insight and where maximums are more valuable. I think we need them both. I'll give examples in a future e-mail and let the rest of you take shots at them.

Bob

== 4 of 7 ==
Date: Sun, Dec 2 2007 4:23 pm
From: doug bidlack

Bob,

I certainly understand where Ed is coming from as I
have the same problem with trying to figure out where
to draw the line about which trees to measure. With
humans, we generally don't consider people to be
mature even though they are sexually able to
reproduce, so we typically use somewhat arbitrary ages
to denote when a person is mature. I suppose we could
do the same with trees, and just measure those over 4"
dbh or some other number just for practical reasons.
I'm not sure that I like this though. It may not be
easy to tell when some trees are starting to
reproduce. I don't know the answer.

Another thing to consider is that maximum tree size
and average tree size are not the only numbers to
consider. Just as important as the mean is the
variation around that mean. A forest with alot of
really tall, fat trees will have high means for both
height and girth, but the variation around those means
will be relatively small and the maximum height and
girth may not be far from the means. In contrast, it
is also very possible to have a forest where there is
a great diversity of tree sizes and here the means
would be lower while the variation around those means
would be higher. Such forests might easily have
greater maximums than the previous forest type.
However, knowing the maximums, means, and standard
deviations would give you a much better picture of how
the forest looks.

Doug

==============================================================================
TOPIC: Putting big tree size in context
http://groups.google.com/group/entstrees/browse_thread/thread/b6cc78c0b837efef?hl=en
==============================================================================

== 1 of 1 ==
Date: Thurs, Dec 20 2007 8:24 am
From: Ed_Frank

Doug,

I don't want you to get me wrong, I like the ideas you are talking
about, I am just grappling with the problem of making the concept
work. I see a couple of things that won't work - but perhaps
something that will work. If you sample a complete population of
trees what you find is that at the very smallest end are inumerable
small sprouts. At the next size up there are fewer slightly larger and
older tree, then fewer yet sapling, then fewer yet young trees, and so
forth until you get to the single biggest tree found. These plot in a
a pretty smooth roughly parabolic shaped curve. The exact shape of
the curve is dependant on the rate of mortality from one size category
to the next. The graph if there are sufficient numbers of trees
sampled and they reflect the general population plot in a very good
looking curve. When Will was looking at the height data for 60+ tall
hemlocks we had a discussion back and forth about whether a less than
two foot gap in the plot where he had no trees in that size range was
meaningful or not. So what this means if you set a particulat height
as the minumum for your sampling, the average of the trees remaining
above that size will be dependant on three factors - the shape of the
curve, the cut-off point you defined, and the maximum height
measured. The average will shift for a species as you shift the cut-
off point. But this tells little about the population of the trees
themselves. Defining the shapeof the curve would be worthwhile, but
the average height based upon this, don't thin would be meaningful.
Another idea I thoght about was using a minimum girth as the cut-off
criteria. But since in a broad sense in a forest setting, the fatest
trees are often also the tallest, this would have the same limitation
as the hieght graph, although the data set would be much noisier.

One criteria that might work is age. That would give you a nice
distribution. You can't age every tree, but if ther was some event, a
derecho, a fire, a clearcut, or other event that knocked all of the
trees down, thenyou could get a relatively even aged stand. Then the
average size for this age of trees would be workable. After the
initial sprouting there would be some x-amount of time before the new
trees started to produce seeds, so there would be an age gap that
might be discernable in the tree heights - ie they may fall into two
height groups. Aside from the few that would sprout from seeds
outside the affected area in interum).

This isn't getting you to quite what you want yet. But it is an idea.

Ed Frank

== 3 of 3 ==
Date: Fri, Dec 21 2007 7:36 pm
From: "Edward Frank"

Doug,

Age is a problem. You can't just look at a tree and know how old it is.
Coring large numbers is impractical. That is why I suggested looking at a
site with a known "reset" date from a large scale disturbance. One thing
Bob has played with in MTSF is to do a Rucker Species Index - The average of
the ten tallest individuals of a species. You could average the X number of
tallest trees to get an average height of how tall the trees would get at a
particular site, if the site had not been logged. With X being whatever
seemed to be a practical number. For less common trees there would be a
bigger drop-off in height from one tree to the next, so you would need to
include a larger sample area to offset this problem. What this averaging
will do, is to lessen the impact of height outliers from the general
population trend. Just keep brainstorming and bouncing around ideas until
you get workable way to address the questions you are examining.

There is an old saying that statistics don't lie, but liars use statistics.
By manipulating the sample selection, sampling process and boundaries, and
processing the data in a particular way, you can figure out pretty much how
to generate whatever results you want to generate. Some of the things I see
put out there seem to indicate this can be done unconsciously as well as
deliberately. The caution for everyone being to make sure the parameters
chosen, and analysis method are the best way to examine the problem, rather
than simply a way to generate numbers to match a preconceived idea of what
they should be.

One of the neat things about this group is the variety of backgrounds of
different people on the list and their different perspectives on various
questions and issue. Gary B is an ecologist, but he was suggesting in a
post, that an Index of dendricity used to model lake systems might be
applied in some way to tree studies. For all I know someone with a
psychology background may know of a methodology from that field that could be
applied to some of the questions we are looking at regarding trees. Even
MAC people con contribute to the quest using the parts of their brains not
turned to mush by "The Cult of Jobs." Hang in there and get things figured
out. Coming up with good questions is often harder than figuring out how to
answer them.

Ed Frank

Home