United States
Environmental Protection
Agency
Atmospheric Research and Exposure
Assessment Laboratory
Research Triangle Park, NC 27711
Research and Development
EPA/600/S3-91/051 Feb. 1992
Project Summary
Forest Health Monitoring Plot
Design and Logistics Study
Kurt Riitters, Mike Papp, David Cassell, and John Hazard
Concern over the condition of for-
ests in relation to natural and manmade
stresses has led to an interagency For-
est Health Monitoring program. To im-
prove the efficiency of forest monitor-
ing, the forest group of EPA's Environ-
mental Monitoring and Assessment Pro-
gram conducted a field test of selected
measurements. The objectives of the
field test were to decide statistical, plot,
design, and logistical issues.
Measurements of soil, vegetation
structure, foiliar chemistry, mensura-
tion, light transmittance, and visual
symptoms were made at 40 plot loca-
tions in New England and Virginia. The
data were used to derive optimum multi-
stage sampling intensities for different
cost assumptions. The field test also
provided a realistic test of logistics.
The number of different types of mea-
surements are recommended for moni-
toring in these forest types and regions.
Specific recommendations are also
made to streamline field sampling. In
general, the plot designs and sampling
intensities currently used for forest
modeling are adequate for the mea-
surements tested.
This Project Summary was developed
by EPA's Atmospheric Research and
Exposure Assessment Laboratory, Re-
search Triangle Park, NC, to announce
key findings of the research project
that is fully documented in a separate
report of the same title (see Project
Report ordering information at back).
Background
The Environmental Protection Agency's
Environmental Monitoring and Assessment
Program (EMAP-Forests) has joined the
U.S. Department of Agriculture Forest Ser-
vice and other government agencies in an
effort to monitor and assess the condition
of the nation's forested ecosystems in re-
lation to natural and man-made stresses.
A long-term and multi-tiered strategy for
monitoring and assessment includes ele-
ments for detecting, evaluating, and ex-
plaining changes in forest condition. To
improve the efficiency of monitoring, part
of the strategy is to test and optimize field
measurement procedures.
A preliminary set of measurements was
chosen for testing in the detection phase
of monitoring based on workshops, a re-
view of the literature, expert opinion, and
reports from studies done elsewhere. The
interagency Forest Health Monitoring
(FHM) program is now conducting research
to expand that set and to verify the capa-
bilities of measurements to accurately rep-
resent and respond to changes in forest
condition over time. Research is needed
to optimize the deployment of selected
measurements, because any per-unrt cost
reductions will be multiplied many times in
a nationwide program.
Thus, the FHM program conducted a
field test of plot design and logistics for
previously selected measurements in 1990.
Not all of the possible measurements were
tested, and not all of the questions that
have been asked about the selected mea-
surements were asked in this study. The
objectives of the field test were:
• to evaluate plot design and
subsampling procedures,
• to quantify time and resource re-
quirements,
Printed on Recycled Paper
-------
• to assess the relative efficiency of
competing methods in some cases,
and
• to supply information to improve the
national FHM program.
The purpose of this report is to summa-
rize the results that were obtained. Rec-
ommendations are made to guide further
planning of the monitoring program.
Section 1 provides an overview of the
rationale, objectives, and approach to the
field study. Section 2 describes the site
and stand characteristics of the forests
that were sampled. Sections 3 and 4 are
the main focus of the report; they summa-
rize the results obtained for plot design
and logistics. Recommendations are based
on the detailed results obtained for differ-
ent groups of measurements as reported
in Sections 5 through 9. These measure-
ments include the following.
• Visual Symptoms - selected
mensurational variables such as tree
size and the percentage of live
crown, and tree crown dieback,
transparency, discoloration, defolia-
tion, and density.
• Soil Productivity - soil chemical and
physical properties.
• Foliar Nutrients/Chemical Contami-
nants - chemical analyses of foliage
from sample trees.
• Vegetation Structure - vertical veg-
etation structure.
• Growth Efficiency - canopy transmit-
tance of photosynthetically active ra-
diation.
The starting point for measurements and
plot design was based on several
interagency committee reports, the FHM
measurement strategy, and the current
plans for monitoring. From this starting
point, a suite of measures of forest site
and stand "condition," broadly defined,
were selected for testing. The potential
efficacy of these measures has been es-
tablished through peer review. The ques-
tion of their efficiency is important be-
cause many are likely to be included in
the collection of measurements deployed
in the field. The type of information needed
about these measurements includes, for
example, how many sites should be mea-
sured, how frequently should measure-
ments he made, how should measure-
ments he physically arranged in the field,
how much do different methods cost, and
what infrastructures are required to make
the measurements.
The field test was conducted in the New
England and the Southeast regions of the
United States. In New England, 20 field
plots were located in northern hardwoods
forest types. In Virginia, 20 field plots were
located in loblolly pine-hardwoods forest
types. The plot locations were selected
systematically (Virginia) or randomly (New
England) from candidate locations defined
by the EMAP sampling grid and by the
U.S. Forest Service forest inventory sys-
tem. At each selected location, a plot
consisting of four subplots was established.
Within subplots, further subsampling tules
were devised according to the particular
objectives for each set of measurements.
This sampling design established a multi-
stage sample framework to address the
objective of sources of variance. It also
established a realistic setting for the test
of logistics.
In New England, the field plots were
located on a wide variety of site condi-
tions, but all were contained in the maple-
beech-birch forest-type group. In Virginia,
the coastal plain sites were less variable,
but both the loblolly- shortleaf pine and
the oak-pine forest-type groups were
sampled. The field plots are representa-
tive of uniform, fully stocked, mature
stands on typical soil types in these forest
types and regions. Stand density (basal
area) ranged from 19 to 48 nWha in New
England and from 13 to 47 mVha in Vir-
ginia. Stem density was between 500 and
1300 trees/ha in New England and be-
tween 400 and 1550 trees/ha in Virginia.
Species composition was different for the
two regions, but the range and average
number of overstory species were similar
between regions. The New England
stands had structures approaching uneven-
aged, and the Virginia stand structures
were suggestive of multistoried stands.
Plot Design
Standard statistical procedures were
used to estimate the optimum number of
sample units for different measurements
and stages of sampling under two sets of
cost assumptions. These results should
be considered guidelines rather than rules
for sampling. In most cases, the sample
designs developed for the 1990 and 1991
field tests are adequate for a regional
monitoring program, but generalizations
to untested species and regions may be
tenuous. The optimum solutions were not
particularly sensitive to cost reductions in
the final stage of sampling. In comparison
to locating field plots and establishing sub-
plots, the costs of the final stages are less
important than the information gained, sug-
gesting that the measurement effort not
be unduly constrained by logistical con-
siderations once personnel are on the plot.
Spatial correlation was not an important
factor to consider when calculating opti-
mum sample sites for measurements that
were made on systematic grids within
plots.
The recommended sample allocations
suggest that the current design of four
subplots per field plot location is more
than adequate for most of the measure-
ments that were tested. The estimates of
sample allocation suggest that two trees
per each of three or four subplots should
be sufficient for visual symptoms mea-
surements for any single species. If sepa-
rate statistics are desired for each spe-
cies, then the total number of trees mea-
sured at each location depends on the
number of species present. Analyses of
typical mensuration variables suggested
that two subplots were sufficient for char-
acterizing average tree sites and total
stand basal area, but the subplots may be
too small because not enough trees were
present on them to adequately portray
stand structure.
The soils data suggested that two to
three soil pits will be sufficient. These pits
should be systematically arranged so as
to represent the entire field plot location.
The foliage chemistry data were quite vari-
able and only two species were tested.
The suggested allocation is for five to six
branches from each of one to two trees
taken from each of two to three subplots,
or a total of between 15 and 30 branch
samples per species per plot location. If
interest centers on the subset of macro-
nutrients, then only one-third as many
branches are required. The larger sample
sites would be required mainly to charac-
terize the heavy metals in foliage.
A conservative estimate of the number
of vegetation structure measurements re-
quired in the forest types tested is six
subplots and four measurement stations
per subplot. The results for photosyntheti-
cally active radiation (PAR) suggest be-
tween two and six subplots, more on
cloudy days and fewer on clear days, and
two measurement stations per subplot.
Both the vegetation structure and PAR
measurements need to be tested under a
wider variety of forest canopy types be-
fore firm recommendations can be given,
because the results obtained were highly
conditioned upon the subplots being "ro-
tated" into similar canopy conditions.
Logistics
The elements of logistics that were ex-
amined included: staffing and personnel,
procurement and inventory control, train-
ing, reconnaissance, sampling, communi-
cations, and safety. It was feasible for a
-------
five-person crew to make the measure-
ments on a plot in a ten-hour day. Six
specific recommendations are made to
improve the efficiency of field activities:
• Specify the criteria for eliminating
sampling sites prior to reconnais-
sance.
• Provide reconnaissance prior to field
sampling to reduce the time it takes
a field crew to locate a sampling
site.
• Review all equipment and consum-
able items with measurement coor-
dinators to determine exactly what
is required.
• Determine staff requirements early
and ensure that contracts are estab-
lished before field sampling activi-
ties commence.
• Provide an assistant to the field
crews to stock equipment and
consumables and to receive, main-
tain, transport, and track samples.
• Provide a better communications net-
work. The regional project leader or
an identified assistant should be re-
sponsible for all communication
with the field crew leader. The lo-
gistics personnel should have early
and close communication with other
groups (design, indicators, QA, in-
formation management) to develop
efficient field implementation.
Visual Symptoms
Measures of visual symptoms and men-
suration are important elements of forest
monitoring in most countries, but care must
be taken to obtain comparable measure-
ments for different species and locations.
This study sampled trees according to the
protocols that were used for implementa-
tion of FHM in New England in 1990, and
despite the comparatively small sample
sizes in thick study, the results were gen-
erally similar to those reported by the FHM
program based on New England monitor-
ing effort. There was no comparable
sample for Virginia, however. The vari-
ance analysis of crown density yielded
results similar to those obtained in a field
test in Great Britain. Most of the variabil-
ity of crown density can be attributed to
tree-to-tree variability, and most of the re-
mainder to stand-to- stand variability. The
quantitative analysis of these sources of
variance allow FHM to derive optimum
numbers of trees to sample within each
stand. Additional analyses are needed to
determine which of the competing mea-
surement methods are most efficient or
accurate. Additional analyses of root and
tree increment core samples are also
needed.
Soils
Measurements of soil physical and
chemical properties are fundamental to
forest monitoring. The study focused on
logistical concerns and on quantifying the
measurement variability that may be ex-
pected when soil measurements become
a routine part of monitoring. Two different
statistical techniques were applied to a
set of laboratory chemical parameters that
were measured on soil samples collected
from eight intensively sampled field plots
in each region. The results suggest that
the present systematic sampling design of
three soil holes per field plot is sufficient.
The variance among pits within clusters
was the same as the variance among pits
among clusters, indicating that a sampling
design with individual pits as the
subsampling unit provides a better alloca-
tion of resources.
Foliar Chemistry
Sugar maple (in New England) and
loblolly pine (in Virginia) foliage samples
were obtained from the upper crowns of
dominant and co-dominant trees on 10
plots in each region. A suite of chemical
analyses were performed which included
macro- and micronutrients, total C, N, and
S, and selected trace elements. Between-
plot, between-subplot, between-tree, and
between-branch variances were calculated.
Between-branch variances could not be
estimated for sugar maple because the
samples were collected incorrectly. The
sample optimization suggested five to six
branches from each of one to two trees
on two to three subplots for most ele-
ments, but fewer branches were required
to characterize just the macronutrients.
Most of the relatively large branch-to-
branch variability observed for the trace
elements may be attributed to concentra-
tions at or below the analytical detection
limit.
Vegetation Structure
Physical alteration of habitats is a threat
to biotic diversity. For this reason, the
structural features of land use, land cover
types, and animal habitats are candidate
indicators of biotic integrity. The primary
objective of this study was to compare an
ocular method to a pole method for mea-
suring the amount, arrangement, and com-
position of forest vegetation. These meth-
ods differ in their conceptual and proce-
dural approaches to estimate foliage oc-
cupancy and distribution. The results indi-
cated that ocular estimates of total foliage
occupancy are significantly (3 to 13%, p
<0.01) larger than pole estimates. The
advantages of the ocular method include
speed and a capability for generating a
complete stand profile (the pole method
was limited to the lowest 30 ft of the
profile). The advantages of the pole
method include lower measurement error,
more flexibility for data manipulation, and
better variance estimates. Both methods
were successfully implemented, but addi-
tional analyses of associations among
measurements and of existing data bases
are required before a definitive judgement
can be made about which method is pref-
erable for FHM program needs.
Photosynthetically Active
Radiation (PAR)
Ground-based measures of canopy den-
sity and processes are considered to be
more of a research topic than a candidate
for full implementation at this time. In the
uniform-canopy stands that were sampled,
it was possible to estimate the plot me-
dian percentage of transmitted PAR with
a relative standard error of between 0.1
and 2.7% using a portable integrating ra-
diometer in less than one hour. Under
ideal sky conditions, relatively few mea-
surements are needed to estimate plot-
level statistics with a high precision in
uniform canopies. In this situation, it may
be better to choose a plot design that
characterizes a larger plot area, so that
PAR measurements can be better related
to remotely sensed measurements.
Cloudy sky conditions reduced the sam-
pling efficiency but they did not invalidate
the sampling design. To alleviate prob-
lems associated with cloudy skies, fewer
samples per subplot but more subplots, or
the simultaneous measurement of both
ambient and under-canopy radiation,
should be investigated.
Summary
The field test of plot design and logis-
tics provided information to help plan the
detection phase of forest monitoring for
selected measurements. In general, the
plot design and sampling recommenda-
tions are within the realm of practical pos-
sibilities. Most measurements will require
further testing in other forest types and
regions before definitive national recom-
mendations can be made. Some mea-
surements will require a critical evaluation
to determine if it is practically possible to
implement them everywhere. It is antici-
pated that the indicator development re-
search currently underway in other FHM
projects will suggest additional procedures
for which plot design and logistical ques-
tions may be tested in the future.
•&LJ.S. GOVERNMENT PRINTING OFFICE: 1992 - 648-080/40177
-------
Kurt Riitters is with ManTech Environmental Technology, Inc., Research Triangle
Park, NC 27709, Mike Papp is with Lockheed Engineering & Sciences Co., Las
Vegas, NV 89119, David Cassell is with ManTech Environmental Technology,
Inc., Corvallis, OR 97333, and John Hazard is with Statistical Consulting Service,
Bend, OR 97701.
Barry E. Martin is the EPA Project Officer (see below).
The complete report, entitled "Forest Health Monitoring Plot Design and Logistics
Study," (Order No. PB92-118 447/AS; Cost: $26.00; subject to change) will be
available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Atmospheric Research and Exposure Assessment Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental
Research Information
Cincinnati, OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
EPA/600/S3-91/051
------- |