There
has been an interesting back and forth discussion between Phil
LaBranche and Lee Frelich among others concerning a soil
sampling design for a study site in Mohawk Trail State
Forest. Excerpts from these discussions are presented
below... Ed Frank
Overview:
One goal is to collect data on the influence of the
chemistry of the underlying geology on tree growth.
One approach would be to conduct soil surveys and get close to
the same data (as rock chemistry). I'd take samples on
top of the rock field and samples as far away from that area at
the same time. I'd run both samples through experiments at the
lab and see what data we find. For instance, if the sample close
to the boulder is higher, much higher, in phosphates than the
outer samples it would begin to prove your hypothesis (concerning
variations in growth). We wouldn't have as precise data, but
we'd be going in the right direction. On top of that UMASS has
recently opened a state of the art soil science laboratory.
Some of the things I plan to do is to dig several soil
test pits and document what I find. There will be pictures and
samples taken. I m going to then take the samples to the lab and
sift out the particle sizes, run pH, phosphorus, nitrates, and
anything else I can think of. I've already secured the lab at
HCC for over the summer, to run my data analysis. The
initial site will be approximately 75 feet square. Once it
is sampled the procedure could be repeated at subsequent sites.
[Phil LaBranche]
Sampling
Density:
The
frequency of sampling depends on the purpose of the study. For
example, if you want to make a contour map of soil properties
and relate that to the growth rates of each tree, you may need a
fine sample grid. If you want to characterize the entire stand
as one unit, then the average of four soil pits is sufficient.
[Lee Frelich]
If you do want to study the growth of individual trees, then for
such a small site, you may have to extend your soil analyses
beyond the edge of the stand, because the environment off the
plot will influence the trees (i.e. edge effect).
[Lee Frelich]
The final
density of samples depends somewhat on the variability you
find in the early part of the sampling. I would start with a low
density
grid, say every 40 feet and then add samples at 20 foot
intervals or
perhaps 10 foot intervals in between if those initial samples
show a lot of
variability, but stop at 40 feet if they are mostly the same.
[Lee Frelich]
Sampling
Site Design:
The site
I'll start work on is about 73.5' square. I plan to even
things out at 70' square. I sketched out the plot and am
going to place 5 pits and 10 corings. I'm going to put a
pit in each corner and 1 in the middle. Then, conduct a core
sample in between. [Phil LaBranche]
A pit would
consist of using hand tools to dig a hole about 1.5-2' wide and
2-4' deep. Some of the soil taken from the hole would be bagged
for lab analysis later on. I'd also use the Soil Survey Manual
to help determine other field data, and conduct a field test for
texture and color. When everything was collected, the hole would
be refilled and covered over to minimize the look of the
disturbance. [Phil LaBranche]
For the soil
description, it is important to note depth of hardpans (if
any), since they can limit the depth of tree rooting, which is a
major
determinant of tree, growth rate, ultimate size and height
because volume
of soil determines how large the root system can be. [Lee
Frelich]
A coring, on
the other hand, would give us less data, but also create less
impact on the forest. I'd use a soil auger to core a hole about
3-4" wide and 3-4' deep. Some of the soil would be bagged
for the same lab analysis as the full test pit soil. When I was
finished, the hole would be filled in and covered over. Using
these two methods together will still give us the quality data
we're looking for, while drastically cutting down on the impact
I'd be leaving on the site. [Phil LaBranche]
I have some
recent experience digging a fairly narrow hole with spades down
to a depth of 3ft+, and, I can tell you, in stony New England
soil, it was NOT a pleasant experience. The rocks were very
uniform in size, like flattened potatoes, and, at different
depths, were maddeningly overlapped, I assume by the work of the
last of those pesky glaciers. Now that I think of it, those
large, flat stones may actually prohibit you from coring down to
any depth. [Susan Benoit]
Soil Description and Analysis:
Probably
early next week, I'll be talking to UMASS and begin
setting up the steps needed in order for me to use their new
soils lab.
Talking to my professor today, about half of the tests I'm
planning to run I
am going to have to do at UMASS, as HCC isn't set up for them.
One of the
other issues I'm going to discuss with the soil scientist is
what would be
good things to be looking for in the samples. I'm still reading
through the
research papers I found online, but am looking to cover pH,
nitrates,
phosphorus, potassium, calcium, magnesium, manganese, zinc,
copper, boron,
sulfates, organic matter, cation exchange capacity, and soluble
salts. Is
there anything in the list that we could rule out? And is there
anything to
add? I'll keep everyone up to date, as things progress.
[Phil LaBranch]
Nitrogen is
a difficult nutrient to work with because it is the rate of
supply of nitrogen in the form of NH4 and NO3 to the roots that
is
important for tree growth, not necessarily total N in the soil.
Most soil
analysis labs can do extractable NO3 and NH4, but it is a
separate analysis
from the other elements. Also, N availability in the horizon
with the most
fine roots (less than 2-4 mm diameter) is more important than N
elsewhere
in the soil profile. [Lee Frelich]
I assume you will send the soil samples to ICP analysis
(Inductively
Coupled Plasma), which is commonly used by soil labs and can
give you
concentration of a dozen elements other than N for several
dollars per sample. [Lee Frelich]
You should be able to obtain an electronic copy from your
library of a
paper I published recently:
Frelich, L.E. Jose-Luis Machado and Peter B. Reich. 2003.
Fine-scale
environmental variation and structure of understory plant
communities in
two old-growth pine forests. Journal of Ecology 91: 283-293.
This is an example of the type of fine scale analyses we do in
the Midwest,
and we used 5 and 10m grids for the samples. However, we were
examining
tree seedlings and small herbs. I don't think a grid that fine
is always
necessary when analyzing trees. [Lee Frelich]
|