Rucker Iterations/Rucker Mania   Robert Leverett
  Nov 10, 2004 06:14 PST 

John and John:

   Last night I updated the MTSF Rucker index through 13 iterations. I
guess curiosity got the better of me. The iteration values follow.

    134.45
    132.27
    130.34
    127.76
    127.59
    126.93
    126.09
    125.33
    124.13
    122.97
    121.97
    121.91
    120.99
    120.32

    The 14th iteration drops below 120, so I stopped.

   It would be super sweet if we could get the 4th order index up to
130, but I seriously doubt we can. I do think we can extend the 120s out
to maybe 16 or 17, but probably not much beyond.

   The bread and butter species for us will continue to be white pine,
white ash, sugar maple, red oak, black cherry, red maple, and maybe
basswood. Beech would be, but too much of it has died and the
distribution of bigtooth aspen, a solid performer through 10 iterations,
is on the low side. Most of the performers come from a single small
stand.

    Even allowing for the death of some of the current champions, it is
unlikley that Mohawk's 1st order index will drop below 133 over the next
few years. But are there opportunities for it to rise above its present
level? Well, with continued growth from the pines, ashes, maples, and
oaks, it might conceivably climb back to 134.7 or 134.8, but I just
can't see 135, now or in the future.

    What are the opportunites? It is conceivable we could eventually get
a 130-foot black cherry. That would replace the aging beech champ, but
the beech will be gone, so gaining the cherry would only be a holding
action. A few hemlocks could climb up into the low 130s. We might gain
one or two tenths of a point.

    Overall, I believe we are at the pinnacle of height development for
the Mohawk forests and we get to experience them at their height zenith.
Rather exciting when you think about it.

Dale:

    How far out the iteration path do you feel comfortable going with
the Cook Forest? I don't even want to ask Will how far out he can hold
up the low 150s. It's not fair, darn it. It's just not fair.

Bob


Robert T. Leverett
Cofounder, Eastern Native Tree Society
RE: Rucker Iterations/Rucker Mania   Will Blozan
  Nov 10, 2004 15:58 PST 

Bob,

I will update a new Smokies RI soon, but an RI of 150 will likely hold for
many hundreds, if not thousands of reiterations. All told, Mike Davie, Jess
Riddle and I have only looked at a mere 5% or less of the Park, and much of
the new stuff will be in the VAST acreages of second-growth that is now
maturing (at a young age) into super forests that defy description.

Sorry BVP, but HARDWOODS RULE, TOO!!!

Will
RE: Rucker Mania   Dale J. Luthringer
  Nov 10, 2004 17:10 PST 

Bob,

You MUST have a lot of time on your hands... 14 iterations?!?!

I'd be willing to wager a case of BVP's favorite beer that Cook should
be able get 130 on the 5th order RI. Top species would most likely be
ranked in the following order (high to low):

White pine, hemlock, black cherry, tuliptree, white ash, cucumbertree,
red maple, white oak, N. red oak, Am. beech

Dale
RE: Rucker Mania   Gary A. Beluzo
  Nov 11, 2004 14:29 PST 

Bob:

Most impressive computing! I am wondering if there is a way to standardize
the field work used to collect numbers for the Rucker Index calculations to
compare different sites more fairly. My concern is that those sites that
get far more effort (like MTSF by you or Zoar by Tom) will obviously do much
better in RI reiterations than sites poorly visited/measured. I also wonder
if there may be other "superlative" sites in the East that just haven't been
identified yet. Is a different kind of sampling warranted? Perhaps the
shear pride of individual dendromorphometrists at their home turf will
ensure adequate "search models" for all of the maxima. The discussions
about size distribution and statistical deviation have been interesting!

Gary

Gary A. Beluzo
Professor of Environmental Science
Holyoke Community College
303 Homestead Avenue
Holyoke, MA 01040
RE: Rucker Mania   Edward Frank
  Nov 11, 2004 19:20 PST 

Bob (and John and anyone else with 2 cents to contribute),

I have some questions concerning the Rucker Index Iterations. I am not
trying to denigrate the usefulness of the process, just trying to
understand. I have re-read everything published on the discussion list
concerning the iterations. I understand how they work mathematically, my
questions concern interpreting what they mean.

Bob Leverett wrote, (March 09, 2004): "The Rucker index process when
applied iteratively reveals the role of each species in the index. Fast
growth pioneer species that form stands tend to play a significant role
for many iterations. Therefore, fairly mature second growth forests can
have the highest indices, especially in areas where a severe climate tends
pare back the canopy. Colby has often mentioned the monotony of relatively
young, tall, skinny second growth forests."

So if there is a relatively even aged fast growing stand of pioneer
species, they will show up for several iterations. I understand that. Why
does doing an iteration, or multiple iterations of the Rucker index show
this better than a simple plot of height range versus number for each
individual species? I can see advantages for a series of height versus
number plots, but really can't see an advantage for the multiple iterations.

Bob also wrote, (March 09, 2004): "The iterative process also allows
anomalies to be distinguished from the norms. There may be a statistical
outlier that unduly impacts the
index and promotes misleading images of a place. Successive iterations
reveal the role of outliers."

This harkens back to a comment I made a few days ago when I pointed out
that by measuring only the tallest trees (for the first iteration of the
Rucker Index) you are not really measuring the expected growth of a typical
or average tree on the site, and may be getting statistical outliers from
the general distribution plot. I would expect if all of the trees in a
certain age range were measured they would form a generally bell shaped
curve, but one that may not be symmetrical, may be skewed to a higher or
lower median value, and be of varying broadness in shape depending on the
species. The measurement of the tallest trees doesn't show what this
distribution may be, how broad the curve would be, the median growth value,
nor the skewness of the distribution.

Bob's statement that multiple iterations will allow anomalies to be
distinguished from the norm is what bothers me. When you do the second or
successive iterations, you are eliminating the highest values from the
previous iterations. If there was an anomalous high value, i.e. outside
the normal distribution curve, it is true it would be eliminated. But at
the same time you are eliminating the high values of less common, or less
measured species from the equation, even though they may well be part of
the normal height/frequency distribution for the species. You are as the
old saying goes "throwing the baby out with the bath water." In these
successive iterations, if you have managed to eliminate some anomalous
heights, you are then comparing a mixture of the top end of some species,
with the second tier of another species, and perhaps lower or anomalous
values not yet filtered by the process. Do you want to filter out the
contributions of the less common species at the site, and if so why?

What I see the multiple iterations doing is demonstrating you have a lot of
big trees and have done a lot of measuring to get a lot of height values.

Bob Leverett said, (March 09, 2004): "In the big picture, we utilize the
multiple iteration process to sharpen our awareness of the role that each
species plays in a site. Stand-oriented species like white pine that form
dense clusters often provide many trees of roughly the same height, so the
role of those species is fairly uniform through many iterations. A
disturbance species like bigtooth aspen is much more thinly represented in
MTSF, but widely distributed. It has an impact over a large percentage of
the Mohawk forest."

Certainly stand oriented species will be consistent over a number of
iterations and those that have grown from widely dispersed disturbances of
varying ages will pop in and out of the iterations. Isn't that pretty self
evident? In what way do the multiple iterations clarify this process? I am
not a forest ecologist so perhaps I am missing something here, I don't know.

John Eichholz wrote, (March 8, 2004): "It is a running question how the
Rucker index values 
would sort out if you actually used all the trees on your site to make up
sets of ten. It would be some kind of bell curve I imagine." I would think
if the forest was a mixed forest of both young and old stands the
distribution not forma bell curve but you would get a downward trending
concave line asymptotic to 0 as the number of trees in each size category
increased as the heights decreased. If it were an open floor of a mature
forest with little regrowth, you would have the same pattern but it would
have a sharp break at the height of any secondary canopy. As I said before
I am not a statistician, but have had some statistics courses, and I really
want to understand what you are thinking.

Ed Frank

RE: Rucker Mania   John Eichholz
  Nov 12, 2004 06:46 PST 

Ed, et.al.:

Let's see if we can make some more progress on this thorny issue.
First, I would like to clarify my understanding of some of these issues.

 
  If there was an anomalous high value, i.e. outside the normal distribution curve, it is true it would be eliminated. But at
the same time you are eliminating the high values of less common, or less measured species from the equation, even though they may well be part of the normal height/frequency distribution for the species.


When interpreting multiple iterations, you are right, successive
iterations will devalue the contributions of less common species, while
weeding out the outliers of common species.

 
  What I see the multiple iterations doing is demonstrating you have a lot  of big trees and have done a lot of measuring to get a lot of height values.

I think that is a valid point, i.e. the sporting aspect of the work.
Not a bad thing. Some of the more thorough stratification work, such as
counting trees into height classes would be much harder than doing
several iterations, though.

 
  The measurement of the tallest trees doesn't show what 
this
distribution may be, how broad the curve would be, 
the median growth value, nor the skewness of the distribution.


To get here, a stratification approach is the only way to go. We have to
get the data.

 
  "It is a running question how the Rucker index values would sort out if
you actually used all the trees on your site to make up sets of ten. It
would be some kind of bell curve I imagine." I would think
if the forest 
was a mixed forest of both young and old stands the
distribution not forma bell curve but you would get a downward trending
concave line asymptotic to 0 as the number of trees in each size
category increased as the heights decreased.

 
Yes, I misstated in that comment. Counting the rucker iterations into
height classes would likely yield a bell-type curve, though.

Now for some further analysis:

I would not advocate "throwing out" any of the iterations. Some of what
they all contribute is:
1) Species composition of the canopy. Types of trees present. Conifers
or hardwoods?
2) The maximum and co-maximum heights by species.
3) The relative heights of species i.e. is oak taller than hickory or
vice versa.
4) Diversity. Are more than ten species present in the set of
iterations?
5) Are there lots a that height or only a few? Likely to find higher
examples? (Dozens at 125+ but none over 130 suggests further searching
may uncover one.)
6) There should be some correlation between size of the site and the
spread of the iteration values. A small site may have a foot or two
between iterations, but one would expect a very large site (Smokies) to
have a very small difference between successive iterations. If this
does not happen, it says something about the uniqueness of the taller
trees. I'm not sure what, though.
7) I still think an index made out of the modal height by species would
show more about the overall site productivity, (but not about
composition.)

I'm sure this is just the surface of it.

John
RE: Rucker Mania   Edward Frank
  Nov 12, 2004 07:31 PST 

John,

If the Rucker index and Iterations of the index are going to continue to be
the mainstay of much of the discussion of ENTS it is manifest that we be
able to clearly state the methodology of doing the calculations and their
meaning. The methodology of the calculations has has been well described.
I have been trying to plead, cajole, and brow-beat a concise listing of
what these calculations show. This type of description has not been
forthcoming. There are three parts of that question: What do the
calculations show? What do they tend to imply? And what don't they show?
The listing you have provided is the best answer yet. I think my
discussions with Colby and others last fall provided a reasonably clear
explanation of the basic Rucker Index. Thank you and I hope you and
everyone else continues to try and refine the usage of these indices.

Ed Frank