ENTS Dimension Rating Index

RE: Formulas for points MY MEAGER TAKE ON THINGS	Will Blozan
	Jan 22, 2005 08:33 PST

Paul and other ENTS,

I find myself with very limited time to respond to such an interesting
discussion. However, I want to present an idea I have discussed with Bob L.
in the past, and it is a system that reflects Paul's desire for a relative
score. It is also independent of units, but is variable as new maximums are
found and does not allow for inter-specific comparisons unless superimposed
upon an absolute maximum "base". Naturally, the system can only be applied
to ENTS measured trees, further limiting it's usefulness in the big tree
lists. Oh well, I will propose it anyway.

With an existing database (ENTS) a set of maximums of girth, height, and
spread are established. The maximums are given a rating of 100, which
represents 100% of the known maximum. For example, let's look three big
tuliptrees; the Sag Branch tuliptree, the Mill Creek Monster, and the
Greenbriar Giant.

Known tuliptree maximums:
Max girth 24.25' =100 pts (Jess may have a larger one)
Max height 178.2' =100 pts
Max spread 113' =100 pts (maximum, not average- treated same
as height which we do not average for separate tops)

With the above numbers, a tree has the potential to have 300 points if it
contained all the maximum dimensions. Here is a comparison of three giants:

Tree Girth Height Spread Points
Sag Branch 91.7 94.3 100 286
Mill Creek 94.4 87.5 88.5 270.4
Greenbriar 100 87.5 85.5 276

How do these trees compare to the best we know of in the east (relative
bigness)?

ENTS maximum dimensions:
Girth 31.8' (Middleton Oak?)
Height 187' (Boogerman Pine)
Spread 154' (Maximum above ground- Cherrybark oak measured last week)

(Max spread estimated) Girth Height Spread Points
Sag Branch Tuliptree 70 89.8 73.3
233.1
Pinchot Sycamore 86.8 52.6 94.2
233.6
Sunderland Sycamore 81.8 61.8 99.4
243.0
Pine Plains Sycamore 75.5 57.8 94.2
227.5
Middleton Oak 100 34.8 87.7
222.5
Cherrybark oak 61.6 85.7 84.4
231.7
Cherrybark oak 62.9 72.2 100
235.1

This system gives much more equality with respect to differing tree forms.
It can be modified to compare within a species or within only conifers. The
Middleton oak, with its huge trunk and wide spread compares favorably to the
Sag Branch Tuliptree. The immensely huge Sunderland Sycamore scores high in
all measures, and justly so!

I believe this ranking system is called "hyper-volume" or something, and is
used in ecology to represent three variable niche fulfillments of species
and habitats. I like it because it is independent of units and the above
numbers generally represent my visual ranking of the trees.

Naturally, as we discover new records the numbers will change slightly for
"saturated" species, and more quickly for less measured species (an active
database would be required to continually update the numbers). The relative
ranking for one or more variables may be useful for latitudinal analyses.
For example, the black birch (and birch family in general) may not change
more than 10 points over a huge latitudinal range, whereas white pine,
tuliptree and northern red oak will change by much more within the same
range. A graph of these relative numbers would be very interesting.

Anyway, there it is!

Will Blozan

RE: Formulas for points MY MEAGER TAKE ON THINGS	Edward Frank
	Jan 22, 2005 10:34 PST

Will,

A couple of questions or points. For the species maximums, what about
using the tallest ENTS measured specimen even if it had since died? Thus
the maximum height would still be Boogerman Pine, only at 207 feet rather
than 187.

It will be an annoyance to recalculate the values for each tree every time
a new maximum for that species is established. It will also be a problem to
compare values if there has been one or several adjustments to the base
maximums had occurred between published figures. There is no reason that a
percentage can not exceed 100%. I would suggest a table of maximum values
compiled on a specific date for all species in the dataset. This could
then be used for an extended period of time, perhaps 5 years. Then the
dataset could be recalibrated using the latest values from that calibration
date. The figures could be published for trees listing the calculated
value and the year of the base maximum set used to derive the figures.

This would help fix some of the implementation problems with the proposed
measure. If other datasets listed the height, cbh, and canopy spread
values for trees, we could calculate the same values for them, but there is
still the problem of inaccuracies inherent in other measurement standards.
There is no reason that a new and better standard for calculating big trees
should not be developed by ENTS. I also think this method generates
numbers with real value and not just cosmetic meaning.

Ed Frank

RE: Formulas for points MY MEAGER TAKE ON THINGS	wad-@comcast.net
	Jan 22, 2005 11:16 PST

Will, Bob, ENTS

Using an excel spreadsheet, or even Access, you could group the trees by species, and have a cell dedicated to the current known maximum for each species. I agree it should be historical, as in the example of the Boogerman pine. A master list would be kept, and when a new dimension record is found, the cell that all the formulas use could be changed, then all the calculations would change automatically.

Scott

Standardized tree hypervolumes	tpdig-@ysu.edu
	Jan 22, 2005 13:16 PST

ENTS,

I really like the idea that's developing here (Paul started this, I think...) of
standardizing tree measurements in relation to known maxima. Units become
irrelevant (as long as they're consistent for each dimension), unlike in the AF
formula where the use of inches for girth dominates the additive formula.

I have a suggestion - don't add the percentiles for each dimension, average
them. That way the standardized maximum is always 100, no matter how many
dimensions are measured (I presume most people would stay with girth, height,
and crown spread, at least initially).

The cool thing with standardized data is we can now rank trees at different
geographical and/or taxonomic scales. E.g., an insanely big hop hornbeam
(Ostrya) at 9' CBH and 90" tall is ranked right at the top for its species. It
would, however, drop way down when ranked among all eastern trees. A 131' x 11'
CBH x 70' CS northern red oak in Massachusetts would rank near 100% among forest
grown NRO in New England, perhaps ~80% among all forest-grown trees in New
England, some unknown but certainly much less than 100% among all NRO in the
East (forest and open-grown), and so forth. The scale of analysis is set by the
investigator, and could range from "all eastern trees" right down to "all trees
of a certain species at a single site". We also have to remember that with all
biological data, no matter what analyses, manipulations, etc. are performed, the
original measurements can always be reported too.

Tom

Re: Standardized tree hypervolumes	Edward Frank
	Jan 22, 2005 14:39 PST

Tom,

I don't see any problems inherent in averaging the percentiles rather than
adding them. It doesn't really affect the outcome. One concern would be
trying to mix percentiles derived by 2 parameters, with those from 3
parameters, with those using 4 parameters, and so forth. By adding an
additional parameter the fundamental ordering of the list could be changed.
By mixing the two sets of information the ordering of one set would be
different from the ordering of the other.

For example a tall tree with a small crown spread might rank very high on a
ranking including only height and girth. If a third parameter for crown
spread were added, a forest grown tree with a narrow crown would be
averaging a relatively small percentile for crown spread into the mix,
while an open grown tree that placed farther down the initial list, would
be adding a relatively large percentage for a broad crown spread into its
mix. This isn't bad of itself, but would make it difficult to compare the
two parameter versus three parameter trees in a single listing, even
thought they both are expressed as a single percentage number.

Ed Frank

RE: Formulas for points MY MEAGER TAKE ON THINGS	fores-@earthlink.net
	Jan 22, 2005 15:10 PST

What about the whole idea of treating in-forest and field trees separately?
If you really don't like to separate them it would be easy to combine the
two separate tables into one master table. I just think that the form of the two
types is so different and the experience of seeing a tree in the forest or
on a lawn is so different....

RE: Standardized tree hypervolumes	Will Blozan
	Jan 22, 2005 19:22 PST

Tom,

I like your thinking, and I did toy with the idea of a "100" max scale. The
300 point scale I proposed gives more depth to the average observer, and a
few more points between trees close in size. It also facilitates separating
the maxima per variable, and better illustrates the relative ranking. A
maximum height ranking of "32" may not mean anything, but a "92" would mean,
"this tree represents 92% of the maximum known for the species, or the East,
or the site..." I.e. - "Wow, Man, that is way tall!"

Roll with it, man!

Will

RE: Formulas for points MY MEAGER TAKE ON THINGS	Will Blozan
	Jan 22, 2005 19:50 PST

Larry,

I think the relative ranking equalizes the extremes in growth form, since a
forest grown tree is usually taller and narrower than an open-grown tree
which is typically shorter but wider. The girth may be the most significant
variable between an open-grown and forest-grown tree, and one that may be
equally outweighed by height.

Will B

RE: Formulas for points MY MEAGER TAKE ON THINGS	Will Blozan
	Jan 22, 2005 19:50 PST

Ed,

...
As far as the updating of the "base", I would want to do it regularly, but I
see your point with regard to simplicity. Either way, the "base" should not
change dramatically since we have such a massive set of data for so many
species from so many ENTS measurers. We are certainly talking only a
fraction of a percent in general, with a 2-4% jump being very rare.

I must say though, that Jess Riddle blew the socks off the NA Carpinus
height record last weekend when he spotted a tree nearly 10 feet taller than
anything previously known (except for an unreported tree Ed Coyle measured
in NY recently- post it, DUDE!). Such a jump will be rare, but not uncommon
for species we do not intensively sample.

Will B.

RE: Formulas for points MY MEAGER TAKE ON THINGS	Will Blozan
	Jan 22, 2005 19:55 PST

Scott,

The bottom line is, we need to develop a formula that is independent of
growth form. But why? I am nagged by the "need" to quantify trees for a "big
tree list". That is not the mission of ENTS. Quantifications of tree
dimensions over latitudinal or whatever gradients we need are what ENTS is
all about. The understanding of trees, not the competition of superlatives.

Will

RE: Formulas for points MY MEAGER TAKE ON THINGS	fores-@earthlink.net
	Jan 23, 2005 11:12 PST

Well, much to my shock, it seems I am alone when it comes to this on this list, but perhaps it could at least be made standard practice to mark down along with other data whether it's in forest or not. It only takes about 2 seconds to make note of this (although certain rare cases might be non-trivial), so it wouldn't seem to be troublesome to make it standard practice, height and spread and stuff are what take effort. So even if the official ENTS and AF tables (not that the AF tables seem reliable enough for anything) don't split lists, others who care about this fact would still be able to make their own customs data tables.

Personally to me, it seems if you don't split the list then why even bother making separate lists for different states, locations, single vs. multi-stemmed, etc. From an biological or ecological perspective in forest and in front lawn are completely different beasts, as different as Smokies vs. northern reaches of Maine, recent third-growth and old-growth. Plus, isn't it more exciting to say find a tree that is 95% of the known in forest max girth than say only 70%, and where probably 100% of the top ten lists for girth will be from front lawns and 100% of heights from forests? Plus, isn't the experience of a great tree in a front lawn with cars rushing by, the sun beating down, and branching starting at 4' way (not that all lawn tree have this form, some are fairly forest looking in form) different than some forest monarch, not that lawn tree may not be impressive and neat, there are some nice old trees in my town, but still it's such a different thing for so many different reasons.

Biologically, ecologically, I don't see how it makes much sense (although, granted, various forest disturbances and histories can complicate things, depending upon what you are after). What can we gain about max girth and height in different forests and open parks and lawns and so on if everything is mushed together? Anyway, I'm beating a dead horse. Hopefully people will at least mark the difference so those who wish can look at the relevant data for particular needs.

I agree. I don't like to hear that a tree is open grown, that is why it ranks so high. The capacity of a species is what we are after. Someone had said that we should list forms separately. I think we need to proceed with a formula that doesn't consider form. The trees should equal themselves out, as forest trees are taller and open trees have more girth.

RE: Formulas for points MY MEAGER TAKE ON THINGS	Edward Frank
	Jan 23, 2005 11:30 PST

Larry,

I don't think you are alone on this question. I would be in favor of
noting whether a tree was open grown or a forest tree. If that note was in
a sortable field in the database, the list could be sorted to generate
separate forest and open grown lists. There may be some problems with
annotating the existing database, but there should not be a problem for new
trees, or when trees are remeasured. It would be worth the effort in my
opinion. Bob, what do you think about a forest or open grown field in the
database?

Ed Frank

RE: Formulas for points MY MEAGER TAKE ON THINGS	Will Blozan
	Jan 23, 2005 19:40 PST

I fully support a distinction. Truthfully, it was not on my mind since I
measure 99% or more forest trees, which is likely true of all the ENTS
members. I take keen interest in yard trees, but focus on the forest
attributes of trees in their "native" habitat and growing conditions.

Great points!

Will

RE: Formulas for points MY MEAGER TAKE ON THINGS	dbhg-@comcast.net
	Jan 23, 2005 14:28 PST

Ed, Larry, et al:

    It might appear from what we routinely report to the list that we don't record anything about a measured tree but its height, circumference, and maybe spread. That is not the case for some of us. We have room in the database that several of us share to record all the following:

     1. Tree's condition (good, fair, poor)
     2. Age class (young, mature, old)
     3. Number of trunks
     4. Growing environment
           a. forest grown, partially open growth, open grown
           b. old growth, second growth, etc.
           c. moist, dry
     5. Height above base at lowest point of branching
     6. General comments

   We intend to add a terrain index factor when we figure out how we want to go about computing it.
    In truth, I get lazy when it comes to recording the above factors and even under good circumstances, time or location of a tree often precludes recording all the data we want. What some of us have been steadily moving toward is a system of identifying important sample trees on a site for which we will fill all the data fields. Non-sample trees would have some added items recorded, but not necessarily all the above.

     Ed, for ENTS website reporting, maybe we could agree on the following:

     1. The number of trunks at point of major branching
     2. Height at point of major branching
     3. Tree condition
     4. Age class
     5. Growing environment: Open, partially open, forest

   There may be other items we should report, but the above don't require more time or equipment. Tree condition and age class are judgment calls. So we may want to discuss these further. What do the rest of you think?

Bob

RE: Standardized tree hypervolumes	Darian Copiz
	Jan 24, 2005 08:03 PST

ENTS,
I think a relative size scale could definitely be more than a novelty.
During the development planning for forested sites there often is, or
should be, a forest stand delineation in which significant trees are
noted so that they might be preserved. These are always the very
largest trees on the site. It has always bothered me that a site could
potentially have a giant Carpinus, Cornus, Ostrya or such, yet these
giants of their species would be cut down without a second thought. If
a scale relative to species was used, some very cool trees might be
saved. Admittedly, it would be a long time before such a scale would be
adopted as part of the development process, but things always have to
start at some point. My two cents on the scale is that I prefer a 100
point max continuously recalibrated. A percentage based off of 100 is
the most readily understandable. For a better understanding of a
particular tree, the three variables could easily be observed as is the
case in most current lists. I don't like separating forest grown and
open grown trees. There are always gray areas. Trees should be
compared to trees. The challenge would be to balance the variables to
best reflect size - if that's not possible, big deal, we would still get
a pretty good idea of the tree's size.

Darian

RE: Standardized tree hypervolumes	edward coyle
	Jan 24, 2005 09:51 PST

Darian,

The 'novelty' comparison was for the arbitrary scores for state and other
champion tree lists. This suggested hyper volume scale is a means by which
comparisons may be made to any order. Eastern North America, our focus,
represented by the three largest dimensions we have for height, girth, and
spread, would be the super set against all trees would be measured. Of
course, the number represented by 100% could adjust upward, if a greater
specimen was found.

Additionally, sub sets for regions, states, sites, species could also be
made. Will detailed this in a Jan. 22 post to the site.

Ed C

Will's WAY COOL proposed new system	Robert Leverett
	Jan 24, 2005 10:32 PST

Will, Ed C., and others:

Oh Man, Will, I've already fallen in love with your proposed 300 max
point system. Look at how it treats the Sunderland, Pinchot, and Pine
Plains sycamores. I may be off a little on the maximum spread of the
Pine Plains tree. Can't find the original max spread measurement, but it
wasn't much off of what is being shown below. Note that the maximum
height for a northeastern sycamore is presently that of the Vanderbilt
Tree and I'm assuming the PA sycamore will yield a legitimate 30-foot
circumference. Scott will have to be the judge on that one, but for now,
we can use the 30-foot circumference as the max.

Comparison of 3 northeastern sycamores against northeastern maximums
Dim --> Hgt Spread Cir Pts
Maximum ………… 136.1 153.0 30.0
Sunderland………. 114.4 153.0 24.9
Pct Max……………. 84.1 100.0 83.0 267.1

Maximum ………… 136.1 153.0 30.0
Pinchot……………. 98.5 149.0 27.6
Pct Max……………. 72.4 97.4 92.0 261.8

Maximum ………… 136.1 153.0 30.0
Pine Plains………. 114.2 135.0 26.2
Pct Max……………. 83.9 88.2 87.3 259.5

Will, your system is WAY COOL. It has to be adopted .... like, right
now! The Sunderland sycamore rules. YES!

Bob

Re: Will's WAY COOL proposed new system	Lee E. Frelich
	Jan 24, 2005 10:56 PST

Bob:

I like the 300 point and % max scale you demonstrate.

FYI the University of Wisconsin developed a 300 point scale for comparing
the abundance of tree and herb species among stands in forests during the
1950s, and it is still known today as the Wisconsin School statistics. It
included relative frequency (species present at 0-100% of all points within
a stand), relative density (species accounts for 0-100% of all stems within
a stand), and relative abundance (species accounts for 0-100% of cover, or
in the case of tree, basal area within a stand). For trees they also used
a 200 point scale that included relative frequency and relative abundance.

Lee

RE: Will's WAY COOL proposed new system	edward coyle
	Jan 24, 2005 11:21 PST

Bob,

I was toying with just how to display the numbers. Your sycamore ranks 267
among known NE max, and the Sunderland would rate 235 against the hypermax.
235/267 or 235.267. Any thoughts on how best to distinguish numbers for sub
sets, ie. region, state, site, species.
Ed C

RE: Will's WAY COOL proposed new system	Darian Copiz
	Jan 24, 2005 11:46 PST

ENTS,

Not meaning to push a personal preference, but just noting another way
of putting it: 89.0%, 87.3%, and 86.5% of "maximum potential". One B+,
but no A's.

Darian

RE: Will's WAY COOL proposed new system- 300 vs. 100 points	Will Blozan
	Jan 24, 2005 12:24 PST

ENTS,

Ed Frank has proposed that the three scales be combined into a 100 point
maximum point total for all three attributes. I personally like the 300
point scale as it give more readily understandable resolution and where the
tree stands with respect to any given attribute. However, I am open to ideas
and enlightenment. It is just a fledgling idea after all, and free to play
with.

My mind is racing with how many ways we can apply the rating system. It can
really give meaningful depth to the "bigness" a tree can have relative to
it's peers. Much more so than the AF points system can. A 300 AF point
hemlock in New England is way more significant than a 300 point tree in the
Smokies. The new system would illustrate that when it is compared "apples to
apples". Such a tree in New England may score the same as or even surpass a
"larger" 350 AF point tree in the Smokies when viewed against the
appropriate regional scale. Likewise, the 111' "shrimpy" hemlocks of the
Porkies can now be viewed as the relative giants that they are- no less
significant than the Tsali hemlock in the Smokies. That is, when viewed
"apples to apples" against their peers. Neat! Credit where credit is due!

Defining the regional scales is the next hurdle, as is naming this new
system. Much to my surprise, the system seems to have instantly gained
support (even by the stalwart ecologists out there), and may be on the way
to adoption within ENTS. This is exciting! Any ideas for a name? Does it
have a "real" name? Is my memory correct in the "hyper-volume" memory?

ENTS Hyper-volume Rating System? Long, no acronym...
Multivariate Abyss of Confusion...

Maybe something with MASS as the acronym?

Regional assessment of Massiveness?

Regional Magnitude of Massivity?

Burl-belly Hyper-volume Fulfillment Index?

Now we're talkin'...

Will B

RE: Standardized tree hypervolumes	Edward Frank
	Jan 24, 2005 13:01 PST

ENTS,

Will's Meager Proposal seems to have reached immediate acclaim among ENTS
who have responded. We should adopt it immediately. There are only two
factors I see left to be decided:

1) Should it be the sum of the three percentiles or the average with the
system based on 100%. Tom Diggens, Darian Copiz, and myself have
supported a 100 percent scale. Others, including Will Blozan and Lee
Frelich, favor adding the three figures together.

I like the 100% scale, because as Tom pointed out: "That way the
standardized maximum is always 100, no matter how many dimensions are
measured." Will suggested that the 300 point scale would to an average
person appear to have a greater resolution - 300 versus 100. I would
counter that the percentages should be carried to 1/10 or 1/100 of a
percent, giving a virtual scale from 0 to 1000. If the percentage were
multiplied by 1000, then a perfect tree - one that was the tallest,
fattest, and greatest crown spread - would have a score of 1000. Consider
the three sycamores Bob cited in his post (Jan 24, 2005) the Sunderland,
Pinchot, and Pine Plains sycamores. In the additive formula the scores are
respectively 267.1, 261.8, and 259.5. Using a 100% scale with scores of
89.03%, 87.27%, and 86.5% of "maximum potential" would translate to 890.3
points, 872.7 points, and 865 points.

2) Should the baseline maximums from which a species percentage is derived
be updated continuously, or on a longer term basis?

If they are updated continuously, then the advantage for doing it this way
is that the values for each individual tree are always up-to-date and
represent the most recent findings. If they are updated on a longer term
basis, and a larger tree was found you would have a percentage greater than
100% (not necessarily bad IMHO). The advantage for this method would be
that if the base line numbers were updated continuously, any published
figures would become out-of-date and differ from each other every time the
base line was updated - in general I think a longer term basis for the base
line would provide information that would allow analysis made using it to
be directly comparable over a longer period of time, and hence have a
greater utility.

There was some discussion about whether a separate list should be
maintained for open grow and forest grown trees. This was a concern
expressed by Larry Baum. I also agree that information would be
interesting. However the general tree database includes a field lists
whether a tree is open grown or forest grown. Seagate lists could be
generated from a single master by sorting from the master database using on
that field. So really only one master list needs to be maintained.

Lets hear opinions on these two questions, and if anyone has any additional
questions or concerns I would like to hear them at this initial stage.

Ed Frank

RE: Will's WAY COOL proposed new system- 300 vs. 100 points	Edward Frank
	Jan 24, 2005 13:09 PST

Will,

I think EHR is a fine acronym.

Questions about new system?

3) Should we be using the current biggest tree, or a historical measurement
that is larger if it was an ENTS measurement with good information?

The example that comes to mind is the Boogerman pine. Should we be using
the max height at the current value - 187 estimated, 186 last measurement -
or should we be using the maximum measured height of the tree - 207 feet -
measured before crown loss a few years ago. I would favor the historical
bigger number, but don't view this as a critical question.

Ed Frank

RE: Will's WAY COOL proposed new system- 300 vs. 100 points	Robert Leverett
	Jan 24, 2005 13:26 PST

Ed:

At this point, I would vote for historical maximums since they tell us
a specie's story over time much better than current maximums do. The
concept of historical maximums should be apply to local, regional, or
eastern wide maximums.

In response to your prior e-mail, I'm still thinking about the 100%
scale versus the 300 point maximum scale. I'm also developing a sample
of 20 New England white pines that I know very well. I plan to put the
scores earned by these pines on each of the 3 point systems ( AF, ENTS
Pts, and EHRS) side by side with comments. I'll pass the results to you
in a spreadsheet in a couple of days so maybe you can post it to the
website for others to look at and comment on.

Bob

RE: Will's WAY COOL proposed new system	Robert Leverett
	Jan 24, 2005 13:32 PST

Ed Coyle:

Good questions. We may want to consider breakdowns of eastern wide,
region, state, and site. We would have to agree on the regional
definitions. That can get messy. The Northeast is commonly consider to
include New England, New York, Pennsylvania, and New Jersey. I'm unsure
of what states belong to what other regions. The pie can be sliced in so
many ways and for so many differing purposes.

Bob

RE: Standardized tree hypervolumes	Darian Copiz
	Jan 24, 2005 14:55 PST

Ed and ENTS,

In regards to topic of discussion 2 below, I think there are some
problems with not continuously updating a database. If a larger tree is
found, and given a score of say for example 303 or 101 or 1001 (whatever
is decided), the maximum potential score of all other trees of the
species are immediately out of date. The new discovery is up-to-date in
comparison to the old base tree/trees. However, although all the other
trees are also accurate relative to the baseline, they do not reflect
the new discovery and a new potential maximum. How clear would it be to
the reader that the tree maximums are set by a baseline, which might or
might not be the actual known maximums for the species. Additionally, a
score above a set maximum implies that the tree's size is actually
greater than the maximum for the species. But the tree in question, by
existing, has displayed otherwise. I think that in order for published
material to never be out of date, scores would have to be chosen that
the species could probably never actually attain so that the baseline
wouldn't change. That, however, doesn't sound like a very appealing
option to me. I think part of the nature of published material is for
it, at some point, to become out-of-date. A web based database,
however, could always be up-to-date. One way to help solve the problem
of an old base line would be to use the scoring system in conjunction
with the actual measurements. Almost all big tree lists I can think of
list the measurements in addition to the final score.

My vote would be for a 100 point system (with possible decimals), second
would be a 1000 point system, and third would be a 300 point system. I
think most people immediately understand a scoring system between 1 and
100 and can more quickly grasp the relation between a given number and
100. I immediately have a good idea what 89.03 is in comparison to 100.
It takes me a little longer to understand the relationship between 267.1
and 300 and even after thinking about it for a while, its still not
quite as clear as on a 100 scale. That's my vote, although it should
probably not be weighed very heavily since I have not regularly
contributed in the past and I don't think anyone in ENTS knows anything
about me.

So as a real quick introduction, I'm a landscape architect/urban
designer living in Maryland right outside Washington D.C. I have a
great interest in trees (of course) and nature in general. I am also
interested in photography and history. Sorry for the tardiness in the
introduction.

Darian

Will's WAY COOL system - problems	abi-@u.washington.edu
	Jan 24, 2005 15:19 PST

All,
I mostly like where this is leading, but you are not there yet.

The new system makes the three measurements equal - are you sure you want that?

I still think average crown spread is arbitrary and not very useful.

All of the methods that convert what a tree is into some arbitrary point system will always have problems with certain trees. The AF system favors big trunks over tall trees and falls apart with super large trees.

What is your objective - do you want this new system to reflet volume? If not, then what else?

In my scientific research I must predict volume all of the time. My equation for Piea sitchensis, for example takes the form of:
(((2.596*(dbh^2.3))+(1.78*(HT^2.697))).

This is based on 61 trees that run the full range of size to 401 cm diameter and 92.7 m tall.

Enough for now.

Cheers,
- BVP

Re: Will's WAY COOL system - problems	Edward Frank
	Jan 24, 2005 16:09 PST

Bob,

I wanted to address your general comments. I am sure others have opinions
to express as well. The question of whether tree canopy should be given
equal weight was ask earlier by Larry Baum.

I don't think this formula is intended to reflect volume, but to reflect an
overall generic idea of "bigness." This is perhaps not a tightly defined
concept, but reflects a balance of all the parameters that contribute to
the aesthetic concept of bigness.

Should all of the factor be weighed equally? For a couple of reasons I
would argue that yes they should be treated equally. You think average
crown spread is arbitrary and useless. I am not sure it is. The crown
spread represents to some degree the size of the tree canopy. This is
where all leaves are located, where all of the branches are located, where
all of the photosynthesis is taking place. It is hardly a trivial area of
the tree in a biological sense, although it may be in terms of volume.

You certainly know far more than I do about tree canopy
structure from your detailed mapping of giant trees around the world. If I
were to look for an expert on the subject, you would be the one I would
call, but on this question about whether canopy spread is a useful
measurement or not, I still disagree with your assessment.

For the trunk dimensions and the canopy dimensions the formula is using
linear dimensions. Using a volumetric measure might be more appropriate,
but a canopy does not form a perfect cone or sphere, the trunk tapers (as
per your formula). Compromises must be made. I feel the simpler method is
adequate for defining the space. This is especially so considering that we
are considering each the measurements as a percentile of the maximum. For
giant trees the formula may or may not breakdown. For a detailed modeling
of tree form, certainly a more complex set of descriptors might be needed,
but as a measure of overall bigness this formula is excellent.

We can calculate an index using just circumference and height as the only
parameters, and we can calculate a figure using circumference, height, and
crown spread. We could calculate a figure using those three plus an
additional undetermined parameter using this methodology. If we base the
numbers on a 1 to 100 or 1 to 1000 scale, all of these would generate
numbers in a comparable range. This would let us see how inclusion of each
of these different parameters affected the overall picture. Part of the
beauty of the formula is that all values are treated equally.

I agree that whatever method is chosen, there may be problems with
certain trees. The formula by comparing the measured parameter with the
largest known measure of that parameter, I believe minimizes that problem.
Please post additional thought you have on the issue, or propose an
alternate formula. I am sure many of us are interested in what you have to
say on the subject.

Ed

RE: Standardized tree hypervolumes	wad-@comcast.net
	Jan 24, 2005 16:20 PST

Ed

1) 300 point scale. I think it will appeal the masses, but is that what we want.

2) I think the data base should be updated at the find of a new champ. I don't think it will happen that often, after the initial year or so. As mentioned before, the data could be calculated easily using a spreadsheet. If not, then at least annually. I know the 10 year span for the Pa tree list probably frustrates alot of people who find a new tree one year after the latest results come out.

Scott

RE: Will's WAY COOL proposed new system- 300 vs. 100 points	edward coyle
	Jan 24, 2005 16:46 PST

Scott,

I too am inclined toward the tri-measure. I think more reduction leads to
a lesser impression, or picture, of an individual tree.

Ed C

RE: Standardized tree hypervolumes	fores-@earthlink.net
	Jan 24, 2005 17:15 PST

So all the coefficients are going to be 0.33333 (or 1) each? No more
thought of testing out various weights for the final output?
a(% of max species Height) + b(% of max species Width) + c(% max species
Spread) and a+b+c=1.0 (or 3.0)

anyway I suppose it doesn't matter anyway, as long as height, situation,
etc. anyone can make up there own tables
as they see fit, renormalizing % of each factor to max for location type
and setting coefficients as they desire, there isn't a way to make one
single list that's good for every purpose or desire anyway.

one thing
though with officially keeping open and forest completely tied together,
well won't this, with equal coefficients case in particular, mean that a
great many in forest trees might not even make it onto the table and
perhaps in other cases notable lawn trees, how many of each type will
appear? if just one from each region then obviously there will be no way to
rebuild the table since you'll loose whatever % of species for each table
type. even if say 5 of each species get in the list, maybe it is a species
that gets exceptional girth and spread in the open and so much less so that
no amount of height will let any forest tree make the list. what exactly is
going to be plugged into the list, all significant finds from each region
(many entries per region) or just one tree per region, etc.?

3) Sould we be using the current biggest tree, or a historical measurement
that is larger if it was an ENTS measurement with good information?

I would also favor historical.

1) Should it be the sum of the three percentiles or the average with the
system based on 100%. Tom Diggens, Darrian Copiz, and myself have
supported a 100 percent scale. Others, including Will Blozan and Lee
Frelich, favor adding the three figures together.

if we did go to varying coefficients, I might set spreads coefficient=0.0,
heh. In some ways I can see it being quite worthy though. In other ways it
seems more problematic and rather a pain. I think overall I could easily
live without it. I would certainly scale it way down at least, to be no
more than a modest tie-breaker (of course there are truly horrible counter
examples, one the worst of which would be the Live Oak, in general though,
I think it might make more of a mess than help, I can think of scenarios
were it could mess things up as badly as it would help for Live Oak, not sure.)

Re: Will's WAY COOL system - problems	fores-@earthlink.net
	Jan 24, 2005 17:37 PST

Ed,

They are good points.
OTOH, I keep envisioning this fat white-ash that grew to immense spread
open-grown, branching exceptionally low to the ground and how while it
seemed big it just didn't seem nearly as big as this one I saw off-trail in
Pine Orchard that just towered up soooo high, but a very narrow crown. I
think with the 1,1,1 (1/3,1/3,1/3) factors it would not even make the list
though and would appear to be almost incomparably smaller by these #'s which
isn't the feel that I get around them. And say there was some redwood 50'
taller and similar diameter to one that, for some reason, and maybe it
doesn't occur to this extent, had exceptional enough spread to come out as
high or even much higher in final score. still, I do see your valid points
regarding spread.

From: Edward Frank
This is perhaps not a tightly defined
concept, but reflects a balance of all the parameters that contribute to
the aesthetic concept of bigness.

Should all of the factor be weighed equally? For a couple of reasons I
would argue that yes they should be treated equally. You think average
crown spread is arbitrary and useless. I am not sure it is. The crown
spread represents to some degree the size of the tree canopy. This is
where all leaves are located, where all of the branches are located, where
all of the photosynthesis is taking place. It is hardly a trivial area of
the tree in a biological sense, although it may be in terms of volume.

RE: Standardized tree hypervolumes	Edward Frank
	Jan 24, 2005 17:37 PST

Larry,

Three equal measures seems to be the way almost everyone wants, but nothing
has been decided yet. I noted in my last email that you wanted to weight
canopy spread less than the other factors, and BVP ask about whether we
were sure we wanted to give all three values equal weight. I favor giving
them all the same weight, but it is not universal. I will be happy to
repost the questions, or post a new message, asking whether all factors
should be weighed equally or weighted in some manner.

It is my understanding that every tree with measurements will be included
in the list so that there won't be any forest grown or open grown trees
with measurements left out.

There may be situations where differences in girth overwhelm those of
height or crown spread, or the opposite. We just need to wait and see what
the data shows. If the results turn out to be too skewed for reasonable
rankings, the list can be sorted by forest or open grown, or the weight of
each parameter could easily be adjusted by changing a number in a
calculating cell. All factors being equal seems the most appropriate way
to start.

the formula is really:

(1/max height) specimen height) + (1/max girth) specimen girth) + (1/max
canopy spread) specimen canopy spread) = Points, the choice then under
this scenario would be raw points as listed, or a percentage value derived
by dividing the result by 3. (I also suggested expanding the percentage
derived by multiplying by 1000 to give a larger, more appealing, number
representation)

One of the ideas for the database is to give people access to the raw
information so that they can manipulate the information to whatever need
they have. So the information can be reprocessed to reflect whatever
sorting or weighting of the parameters you want.

Ed Frank

RE: Standardized tree hypervolumes - Weighting Parameters	Edward Frank
	Jan 24, 2005 17:40 PST

ENTS,

Larry Baum and Bob Van Pelt have both questioned whether or not we wanted
to weigh all three of the parameters under discussion evenly - Height,
Girth, and Canopy Spread. What are your opinions, please try to explain
the pros and cons of each option in your reply. I have previously stated
my opinion.

Ed Frank

RE: Standardized tree hypervolumes	edward coyle
	Jan 24, 2005 18:16 PST

Larry,

As I understand your concern for 'losing' a species, due to its being
forest or field grown, I can only think that it would depend on the list.
All trees would not, and should not, make a national list. It would be
exhaustive. However, a plot or site, might require recording every species
present.
As the list jumped in scope to state, or regional level, your particular
trees may be lost due to their relative unimportance. For example a state
champion something might disappear from an East coast listing.
The list can be made to any level you want,plot, site, state, region. It
will have ultimate comparison with the hyper volume tree values. The number
of each species listed could be unlimited, but would more likely, it would
be reduced to a managable number,and be representative
of the best the sub plot has to offer.

That's how I understand the concept at this juncture.

Ed C

RE: Standardized tree hypervolumes	fores-@earthlink.net
	Jan 24, 2005 18:17 PST

Darian Copiz wrote:
In regards to topic of discussion 2 below, I think there are some
problems with not continuously updating a database. ...

One way to help solve the problem
of an old base line would be to use the scoring system in conjunction
with the actual measurements. Almost all big tree lists I can think of
list the measurements in addition to the final score.

I guess it would be easy enough to keep up to date. Each % for the three
values would be calculated based off of excel cell location for that
species max girth, height, spread so it could automatically update for every
tree of that species without any work at all, since it is is recording all
the data for each tree. You could have a function that goes down the list
automatically finds the largest for each of the three values and stick it
in the storage box for that tree which the percentage boxes would then make
their calculations based off of. You'd have functions to grab this for
state, region, location type etc. and then other telling which to use.
actually this msg is getting garbled i don't have time to be coherent now,
back to work.

RE: Standardized tree hypervolumes - Weighting Parameters	Paul Jost
	Jan 24, 2005 18:18 PST

Ed,

I never liked spread since it assumes nicely symmetrical trees. It
penalizes trees with asymmetrical crowns. It usually doesn't tell you
anything about growth potential since trees shed branches too easily.
BVP once mentioned using the length of the longest branch. If we needed
to add a third dimension to the scoring parameters, then I believe that
the third dimension of longest branch is probably the fairest. Also,
the longest branch measurement is as easy to make as the height and
girth measurements, while spread may be difficult to make on steep
grades or on trees overhanging rivers. Spread on a line through a tree
trunk is also more prone to error to those not careful in their
measurements due to the fact that the measurer uses his opinion and
perspective to assume the axis of greatest spread. Average spread then
takes the spread of the axis at a right angle to the axis of the
maximum spread. In a large data set with many measurers, it is just
too sloppy of a measurement to consider of equal weighting to girth and
height.

As far as 100 vs. 300 points go, who cares - it's all relative anyway.
But the advantage to a 100% scoring system over a 300 point system is
that you can take a preliminary measurement without spread and see where
it lies. Most of us may never take spreads on the majority of trees
anyway unless we find that it's score warrants fine tuning due to it's
large size. Otherwise, large trees missing one dimension will be too
far down the list in a 300 point system. This has been one of my pet
peeves on the AF scoring system.

Paul Jost

Parameters - Max Crown Spread or Longest Branch	Edward Frank
	Jan 24, 2005 19:02 PST

ENTS

Question(4): What should we be using to measure crown spread: Average
crown spread, maximum crown spread, or longest branch?

Paul Jost suggested the following: "BVP once mentioned using the length of
the longest branch. If we needed to add a third dimension to the scoring
parameters, then I believe that the third dimension of longest branch is
probably the fairest. Also, the longest branch measurement is as easy to
make as the height and girth measurements, while spread may be difficult to
make on steep grades or on trees overhanging rivers."

Bob Leverett stated in a previous post: "It's my understanding that we
would be using maximum spread for a tree instead of average spread. How
does that impact your thinking about the 300 point system? If we were to
use weights, do you have a feel for how they should be determined to do
justice to volume?"

Average crown spread is a different concept than longest branch or maximum
crown spread. Much of the existing data may be in the form of average
crown spread, but we could dig out a longest branch approximation for many
of them. The biggest are remeasured frequently and could be redone using a
revised measure criteria. So this should not be an overwhelming factor.

Ed Frank

Re: RE: Will's WAY COOL proposed new system- 300 vs. 100 points	Jess Riddle
	Jan 24, 2005 20:17 PST

I often think of ENTS tree measuring as investigating the potential of
different species, and I believe several others have placed measuring in
the same context. From that perspective, this new point system makes
great sense. Keeping potential in mind suggests that we use all ENTS
measurements not just the greatest current dimensions of known trees; that
is, we should use historical data. That view also supports updating the
reference values continuously to best reflect the species' known
potential. That process would diminish the value of posting old trip
reports on the website, so I'm not nearly as firm in my position on this
point. If the list were updated with each new record, that event would be
a fairly common as evidenced by six new records in one day recent day of
measuring in the Congaree.

A 300 point system seems a significantly more useful format than a 100
point system. Reduced intuitive appeal has been cited as the main
drawback of the 300 point system, but no one on this list has requested
further explanation of how the 300 point system world work and it seems
unlikely that the system would remain perplexing to anyone after a moments
consideration; the 100 point system may be more intuitive, but the
difference is too small to cause any real difficulties. On the other
hand, the 300 point system does provide some unique flexibility. We have
little spread data for many species, so if we add the height and girth
percentiles for those species we essentially have a scoring system out of
two hundred points. Looking at those score would immediately reveal that
one dimension had been left out whereas with a 100 point system how much
data had contributed to the point score would be unclear. The 300 point
system would be analogous to the use of the Wisconsin School statistic or
Importance Values where three percentiles are normally combined to obtain
a value but the frequency value may be omitted if only one plot is used
and a note is simply made that the scores are out of 200. Anyone more
familiar with those systems feel free to correct me, or elaborate if that
is incorrect.

I'm still undecided on long branch vs. no spread vs. long spread; however,
if spread is used I would argue for long branch or maximum spread over
average spread. Again, either of the latter two dimensions relay
information about the structural potential of a species while average
spread obscures the raw data.

Having something to refer to the point system by rather than "300 point
system" would be nice. Hyper-volume doesn't appeal to me since I keep
thinking of hyper as a prefix, and that meaning does not apply well to
this use. Also, volume is not actually being measured, so that term could
easily be misinterpreted. Alternatives include percents of potential
scale (POPS), relative maximum dimension score (RMDS), or some other name
using percent, relative, maximum, or potential. Don't know if score or
scale should be kept on the end or not. Fortunately, that concern is
relatively trivial.

Jess Riddle

Re: RE: Will's WAY COOL proposed new system- 300 vs. 100 points	fores-@earthlink.net
	Jan 24, 2005 22:01 PST

On the other
hand, the 300 point system does provide some unique flexibility. We have
little spread data for many species, so if we add the height and girth
percentiles for those species we essentially have a scoring system out of
two hundred points. Looking at those score would immediately reveal that
one dimension had been left out whereas with a 100 point system how much
data had contributed to the point score would be unclear.

Personally, I don't really care much whether 100 or 300, since it is the
exact same
thing and 300 isn't something pain in the neck weird like out of 967.8,
although 100 seems a little
bit more natural since it has been normalized to a more natural mode for
our typical base 10 way of thinking.
But it's a very minor point, maybe 300 sounds bigger and more grand, maybe
100 sounds like perfect score, I don't know.
I don't see that your arguments make all that much sense though in
supporting either 100 or 300, a tree might be
70%,75%,40% for a 170% total and another might easily be 80%,95%
for a 175% total, so I don't see how this 300% system automatically reveals
whether one aspect has been left out or not, here both are below 200 AND
the one with only two
components is larger. Sure, in many cases if it has less
than 200 it would mean that, but I don't think it would be all that uncommon
for it not to either, it would be entirely unreliable way to tell at a
snap, although perhaps trees of such low triple
points would be rare (but then again a forest tree might have low spread,
big but not tops girth and decent height and often apear less than 200).
Also, with the 100 points then you just look if its under 66, same as
checking if its under 200. So whether you add the three (two)
together or add the three together and then divide by 3 (2) they both seem
equally easy and clear. OTOH, if we tracked all three, but very or mostly
only used 2 in the end, then 200 is more natural than 66 to deal with, so
that might favor the 300 system since there would be less converting thigns
back and forth, although then again if we deal mostly with out of 200 could
make the 200 a 100 and the few times you use 300 make that out of 150.

Jess weighs in	dbhg-@comcast.net
	Jan 25, 2005 03:57 PST

Jess:
Thanks for weighing in on the alternatives. Your writing is crystal clear. You've done a service for the less numerically focused on our.
Even though, I support the 300 point system slightly over the 100 and am partial to the longest limb slightly over the longest spread, we shouldn't rule out some weighting umtil we've applied the formula to lots of trees. I'm presently comparing 16 white pines in New England on all 3 systems: the proposed one, AF, and ENTS Points. From my perspective the new system is a slight improvement over the other two for reasons that are as subjective as objective. More to come on this topic.
Bob

Re: Jess weighs in	Lee E. Frelich
	Jan 25, 2005 06:28 PST

Bob:

I agree with Jess. 300 points is more honest, since it tells you right
away that three factors were used, and for those trees where you only have
2 measurements, you can compare them using the 200 point scale. This is
probably the reason that most indexes based on multiple factors, such as
the plant abundance system I explained in a previous e-mail, use 200 or 300
points.

The longest limb is probably the best measurement to take for purposes of
comparing big trees.

I take the crown radius in several directions and use elliptical formulas
to calculate crown area, since I am usually interested in area occupied by
trees in different age classes. However, that is much more detailed data
than necessary for big tree comparisons.

Lee

Random thoughts on measuring and formulas	Robert Leverett
	Jan 25, 2005 06:32 PST

ENTS:

The unprecedented burst of e-mails on where to take the m