Rucker Site Index   Ed Frank
  Aug 10, 2003 16:50 PDT 

The Rucker index is a common topic in many posts made to the ENTS
discussion list and in the Eastern Native Tree Society Website. Colby
Rucker (2003) describes the index in his prolog to the Tall Tree
Preserves of the East list on the website tall_tree_preserves.htm As

“The following sites are among the most important examples of tall-tree
habitat in the eastern United States. They are listed according to the
average height of the tallest examples of the ten tallest species found
at each site.   This index, often called the “Rucker index,” provides a
numerical evaluation of both maximum height and diversity of the
dominant species. High index values are the result of many factors,
including climate, topography, soils, and a lack of disturbance. While
the most extensive sites benefit from a greater variety of habitat and
more individual trees, some exceptional sites are quite small. Although
many of these sites have been recognized as National Natural Landmarks,
or are located within national forests, neither designation provides
actual protection. Some sites are on private property, and are
vulnerable to logging or clearing for development.” Copyright 2003 by
Colby Rucker.

As this is such a high profile topic for ENTS, I emailed Colby with
numerous questions about the index. The quote above and a personal
email resulted. The overview below draws both from my observations and
questions, and Colby’s replies.

The Rucker Site Index or Rucker Index has a numerous merits that make it
a useful measurement when comparing various tall tree sites.

1) The formula is straight forward, unambiguous, and easy to apply. The
measurement is simply the average height of the tallest examples of the
ten tallest species found at each site. Anyone who can add ten numbers
and then divide by ten can calculate the figure.

2) The index can be applied to forests in any area with any make-up of
trees. One of the biggest problems faced when comparing different areas
is that the same tree species are not found in all areas to be compared.
The index is not species dependent. It does reflect to some degree the
species included, because all tree species do not reach the same height,
but none-the-less the formula will produce a useful concrete number.

3) The index requires a fairly diverse mix of trees in order to generate
a high index value. This means in order to achieve a high index rating
that the forest patch being evaluated must not be primarily dominated by
a single species, but be reflective of a more complete and by inference
a more intact forest ecosystem. This has some drawbacks however. Robert
Van Pelt, author of "Forest Giants of the Pacific Coast" writes in a
post dated October 29, 2002, “The low diversity of trees in some Western
forests quickly reduces the Index to below 200. Humboldt Redwoods SP,
for example, has the world's tallest tree, and 86 trees over 350'. Due
to the overwhelming dominance by redwood, the Index drops below 200
after only six species are included!”

4) To get a sufficient diversity of trees of great height requires a
fairly large plot of forest. This limits sites with a very small area
and few trees generally from having a high index. The value in this, as
I see it, is small sites with only a couple of spectacular trees, or
only a single big tree do not falsely appear from the index value alone
to be the equal of larger sites with larger numbers of trees.   

The Rucker Site index is essentially a foreshortened version of a
complete profile of all the species found on a particular site. Colby
(July 27, 2003) writes: “I enjoy constructing forest profiles for all
the tree species present on a site, right down to the mountain laurels.
It's extremely interesting to see how the numbers (based on a single
representative specimen of each species) arrange into groups (usually
six) according to habitat. I think the Chase Creek study on the webpage
shows that. It's clear that we couldn't expand an index to include more
than ten species, because a diversity of small species could skew the
whole works. I selected ten species, not perhaps twelve, because I was
interested in diversity, but wanted to keep within the dominant species,
and ten meant a convenient decimal point.” I would like to see someone
prepare complete profiles for all the species on several sites and
compare the results. At this point we have significant amounts of data
on The Great Smoky Mountains, Chase Creek, Cook Forest, and Mohawk Trail
State Forest. Perhaps the chart could be done as a 3d map, arranged
based upon a common x-axis by species, y-axis by latitude, and z-axis as
tree height.

There are some negatives with the use of the index. Some of its
limitations are the same as some of its strengths. The index is not
measuring the same tree species in every different forest. White pine
is present in some stands and not in others. As a general statement the
Rucker index is dependant on the size of the site. It is not any way a
measurement of the age or maturity of the forest, just size. Old
growth forests on poorer soils will receive indexes lower than second
growth forests on good soils. The index is an amalgamation of which any
of a number of factors
could cause the index to be lower, so again the factors that affect the
index are variable from location to location, and not clear from the
generated index value alone. I posed these concerns to Colby, and in
most cases his perspective ameliorated my concerns.

The most important concept is to understand what the index portrays and
what it does not.. It is in effect a measurement of the potential of a
given site to produce tall trees. Differences in the age of the forest,
and other parameters such as site slope, water availability, latitude,
and various environmental factors can not be deduced from the index
alone. It is not meant to address these issues and should be used in
conjunction with a site description to deduce the cause of height
variations between forests.

The Rucker index is not in and of itself an indicator of old growth.
The question of old growth and height is also a subject of debate.
Colby has suggested for consideration, that perhaps as a forest matures,
taller, thinner trees might be replaced by broader, heavier limbed
species. Tall species like tulip tree and white pine would become less
dominant as old growth is approached. The stage of tallest trees may be

The example cited above, with regard to the coastal redwood forests,
whereby mature forests with a less diverse tree population being
undervalued is still a question in my mind. Colby (July27, 2003)
writes: “The index itself is subject to the whims of diversity. At Belt
Woods, here in Maryland, the first four species are over 140 feet, but
the tenth tree is only 100. It's a remarkable site, but may have been
subjected to some sort of "forest improvement" cutting of "inferior"
species a hundred years ago. So, the index does favor sites having a
diversity of habitat supporting numerous species of all sizes.”   As he
says, the index is what it is, and this is true for whatever measurement
option you choose.

I posed a question about how the indexes could be artificially inflated
by including larger areas, the equivalent of gerrymandering in designing
political precincts. Colby (July 27, 2003) writes: “Habitat is much more
important than acreage...Yes, we can increase index ratings by
incorporating more territory. By lumping a lot of Jess Riddle's sites
for Station Mountain or the Andrew Pickens Ranger District, the numbers
went up, but it showed that the Great Smoky MNP wasn't so unique.
Sometimes smaller sites make a good showing, as at Ice Glen (ca. 40 a.),
Belt Woods (43 a.) or Chase Creek (ca. 150 a.). I don't think there'd
be any point segregating sites according to acreage at this point. We
don't have much info on privately-owned sites; Chase Creek is the
primary example.” I believe as we collect more data perhaps some
evaluation may need to be made of this factor. But at the moment we
simply need to rely on the people making the evaluations to fairly
delineate the boundaries of the study area.

I questioned whether a single specimen of a tree was a fair measure of
the entire population of the tree species in a forest. Colby (July 27,
2003) writes: “I've become comfortable with a single specimen
representing that species on a site. The maximum number does show the
potential for the species under the inherent conditions. The validity
of a single representative specimen is shown by the way the numbers
become arranged in forest profiles, the Chase Creek study on the webpage
being a good example.” Well, maybe.

I questioned the use of a hardwood index to compare various sites. My
concern was that any number of indexes could be constructed with or
without certain species. Colby argued, and I must agree, that hardwood
indexes are useful because they allow fair comparisons to be made
between sites that do not have the tall species of White Pine and
hemlock with those that do. These species have often been removed by
selective timbering in the past, or in the case of hemlock, are
currently being killed off by the hemlock woolly adelgid.

The final concern I raised was about statistical regression analysis on
the height data being conducted by Bob Leverett. One example is his
post to the discussion list dated Dec 05, 2002 18:24 PST, entitled
Rucker Height Index. My concern is multifold: 1) The same people are
not measuring the data at all the sites. Assuming the methodologies of
the measurements are the same, there is still personal bias in what
trees are selected for measurement; 2) I believe there is some selection
bias in which trees are measured. Some individuals may focus on certain
species more than others...etc. 3) I believe th selection methodology
changes over time as big trees are found and measured; 4) the
measurements are too dependant on individual trees. The death or
discovery of a single specimen can significantly affect the overall
index value.

To elaborate on item three above what is measured and recorded is not
the same from the first measurements to the last. When first visiting a
site, every large tree is measured until a really big one is found.
Thereafter measurements are still taken, but I would think tend to focus
on trees that are very big and have the potential to be a new height
champ... more modest large trees are not measured as aggressively. Also
Once a sufficiently big example of one tree is found, I am sure there is
a tendency to focus on other species, in order to provide data for a
Rucker index. Big trees are not completely ignored, but I am not
convinced that as much diligence is paid to them once the feeling a
species has topped out is generated. All of these could seriously skew
the statistics.

Colby (July 27, 2003) writes: “Both Bob and Dale have made reiterations
of the indexes for MTSF and CFSP. These indexes, down to the ninth and
tenth-place representatives, have
been interesting. Of course, you have to have a lot of trees/acreage to
find enough decent specimens to see much. At Cook, the reiterations are
quite close, being in line with the rather monotonous aspect of the
hardwood forest canopy. Chase Creek is quite the opposite, with many
specimens having dramatic potential, but great vulnerability to a single
windstorm or lightning strike."

I am not completely convinced by these arguments on the validity of
statistical analysis, but it does provide an interesting framework for
evaluating the evolution of the index for a particular locality.
Overall I have a greater understanding of the Rucker index and how it
may be successfully applied to the tree height data being collected
using the Eastern native Tree Society protocol. I want to thank Colby
Rucker for his invaluable input and patience with my questions.

Ed Frank
Rucker Index Continued
  Aug 11, 2003 16:37 PDT 


Thanks for reintroducing the Rucker index as a subject for discussion.
Colby has given us a lucid explanation of the value of the Rucker index and his
observations pave the way for further discussion. I'll go first, albeit in a
somewhat different direction.

All of us who use the Rucker index do so with an eye toward its
limitations as well as its strengths. None of us see the Rucker index as an end-
all. Although we haven't been focusing on it here on this list, other data need
to accompany the index if we are to adequately describe the forest sites that
we study. As Colby explains, one of our objectives in applying the Rucker index
is to facilitate site to site comparisons, individual species comparisons to
assess a site's overall growth potential. To these ends, height is most
valuable, but completeness requires that we other measures including age and
diameter distributions. For more extensive research, we need environmental,
geographical, and topographical data to include latitude, longitude, altitude,
aspect, precipitation distribution, soil type, bed rock, etc. In pursuit of a
scientific objective, we would not inflate the role of any single statistic,
even charismatic ones like the Rucker index. In promoting sites for other
reasons, a single measure like Colby's index can receive the lion's share of
the attention. In time, Colby's index will seek its true level of scientific
importance, but for me, the Rucker index has already paid handsome dividends.
It has been an important break-through to getting a handle on an illusive
aspect of forest development and I think it is will eventually gain wide
acceptance as one measure of site fertility.

A primary ENTS objective has become development of easy to use methods
for rating a site's raw tree-growing capacity both as an instantaneous snapshot
and as a way of measuring capacity over time as an aggregate measure and for
each represented species. To this end, a frozen image of a site as encapsulated
by a single computed number based solely on tree heights tells only a piece of
the story. Mean stand age tells another story. Site disturbance history tells a
third. This admission seems to fly in the face of all the hype we've been
giving the Rucker index, but that is definitely not the case. We've always
understood site complexity. Yet regardless of how sophisticated we get,
accurate determination of tree heights will always remain an important
component of our analysis. The reason for our past heavy concentration on this
one measurement is that, despite its obvious importance to the profession of
forestry, accurately measuring tree height has been poorly handled by all
concerned - perplexingly so. Well, on second thought, perhaps not so

Before lasers, measuring height accurately was labor intensive. Many
hours spent with a transit taught me this. Many foresters already knew the
labor required. Measuring every tree of interest with a transit was just not
an option. Simple, field-efficient methods had to be developed to provide
height approximations that were good enough for forestry purposes of timber
volume assessment. That does not require whole tree measurement. The principle
method developed that uses the slope concept became widely adopted. If
carefully applied, the method can get you within +/- 5.0 feet of true tree
height for a surprising number of trees. In the case of straight conifers,
accuracy can be +/- 2 feet. So slope measurements in a planted spruce stand
make sense.

However, measurement within dense forest stands, measurement of broad-
crowned hardwoods, and measurement in rough terrain often leads to large errors
using the tangent-based slope method. That isn't an opinion. It is a fact. But
even when the method does pretty well, it doesn't satisfy the ENTS goal of
achieving +/- 1.0-foot accuracy levels. This degree of height accuracy is
necessary for ENTS purposes. Why? Because large errors prevent accurate site to
site comparisons, which enable us to identify exceptional forest sites that
have fallen through the cracks. We have to be accurate if we're going to lock
horns with potential adversaries who see nothing special about the sites. The
tree measurers of ENTS could write a book on this subject.

But not only can we identify exceptional sites, given enough data, we
can explain in detail what makes them exceptional, at least on a statistical
basis. Ultimately, we can rank the sites, locally and regionally and provide
valuable data to many programs. Zoar Valley, NY may be our best recent example
of our ability to do this.

A critical element of all this is the knowledge we add about a site
that was not pre-existing. Knowledge flowing from the comparisons we are able
to make fosters appreciation for the rare places and appreciation is a
precursor to protection. Zoar Valley, Mohawk Trail State Forest, Ice Glen, and
Bryant Homestead are just a few examples of sites where ENTS input has changed
public perception toward much greater appreciation.

By contrast, lack of proper appreciation of special sites, which can
include some of our most treasured icons, by the very resource agencies
assigned to protect them is not limited to isolated cases. Will Blozan and I
saw that back in the mid-1990s and our concerns have since been echoed by
others. Tom Diggins and Bruce Kershner have spoken eloquently to this point
with respect to Zoar Valley, NY. Their arguments ring equally true for other
exceptional NY places. If protection of the exceptional sites flows from
appreciation of them, appreciation is enhanced by comparative knowledge about
them and validating the site comparisons requires that we be squeaky clean,
i.e. accurate in our measurements - thus that needs to be our compulsion.

    However, the story does not end here. The negative impact of poor
measurement methodology goes deeper than lack of appreciation of our forest
icons by the resource agencies we entrust to protect them. Large areas of
private commercial forest have been under-valued with respect to growth
potential and that has encouraged the high grading epidemic we now witness,
with what I might add, the complicity of the state's forestry oversight
apparatus and the troubling diffidence of some forestry academics. ENTS members
are referred to past posts by Karl Davies who exposed flawed techniques in the
development of volume growth models by the U.S.F.S. Northeast Experiment
Station that can be traced primarily to two factors, one of which is, guess
what? Tree height measurement! Thus, our crusade continues, but we acknowledge
that it should not deter a fuller array of ENTS efforts and adoption of
disciplined measurement protocols that serve many scientific and promotional

I say this still emphasizing that ENTS is not principally about big
tree contests - as much fun as they are. Our mission is far more serious.
However, we shouldn't feel reluctant to promote the sporting side, especially
where a preservationist benefit can be derived. Contests are easy to report on
and have considerable publicity value, so they will remain an important part of
our daily grist. But in time and courtesy of what Dr. Lee Frelich plans to
accomplish, ENTS will make a far richer contribution to our understanding of
how environmental and geographical factors are correlated to the growth
potential of many eastern tree species, which will relieve us from a dependence
on anecdotal accounts with their inevitable exaggerations and more to the
point, expedient measurement practices. Uses of such knowledge? I leave that
open for other Many!

Finally, I'd like to return to the Rucker index. How has it helped in
our studies of MTSF? Well, it has shown the place to be exceptional at growing
trees for the latitude. Applying the index iteratively, we can quickly deduce
that the index isn't a fluke. We can also determine the canopy role of each
species for areas of different size and habitat type. In short we can obtain a
gold mine worth of data that puts a whole new light on the region. Although,
MTSF occupies 6,800 acres of real estate, in only 750 contiguous acres we get
the Rucker index of 134.01. In slightly less than 200 acres, our index is 131.
Adding 50 more contiguous acres gets us up to 133. These figures should the
raise eyebrows of anyone interested in forest growth potential.