EPA/600/4-91/012
March 1992
MONITORING AND RESEARCH STRATEGY FOR
FORESTS - ENVIRONMENTAL MONITORING
AND ASSESSMENT PROGRAM (EMAP)
U.S. Environment*! Protection Agency
Office of Research end Development
Washington, DC 20460
Environmental Monitoring Systems Laboratory • Las Vegas, NV 89118
Environmental Research Laboratory - Corvallle, OR 87333
Atmospheric Research end Exposure Assessment Laboratory - Research Triangle Park, NC 27711
-------
NOTICE
The information in this document has been funded wholly or in part by the United States
Environmental Protection Agency under cooperative agreement (CR81470) with the Environmental
Research Center of the University of Nevada at Las Vegas, Contract No. 68-C0-0049 to Lockheed
Engineering & Sciences Company, Contract No. 68-C8-0006 to ManTech Environmental Technology,
Inc. in Corvallis, Oregon, and Contract No. 68-00-0106 to ManTech Environmental Technology, Inc.
in Research Triangle Park, North Carolina. It has been subjected to the Agency's peer and
administrative review, and it has been approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
Proper citation of this document is:
Palmer, C.J.', K.H. Riitters2, T. Strickland3, D.L. Cassell3, G.E. Byers", M.L. Papp4, and C.l. Liff1. 1991.
Monitoring and Research Strategy for Forests-Environmental Monitoring and Assessment
Program (EMAP). EPA/600/4-91/012. U.S. Environmental Protection Agency, Washington, D.C.
University of Nevada-Las Vegas, Environmental Research Center.
ManTech Environmental Technology, Inc., Atmospheric Research and Exposure Assessment
Laboratory - Research Triangle Park.
'ManTech Environmental Technology, Inc., Environmental Research Laboratory - Corvallis.
4Lockheed Engineering & Sciences Company, Environmental Monitoring Systems Laboratory - Las
Vegas.
ii
-------
TABLE OF CONTENTS
Notice ii
Figures iii
Tables vii
Acronyms viii
Acknowledgements xi
1. INTRODUCTION
1.1 Content and Organization of EMAP-Forests
Strategy Plan 1-1
1.2 Overview of EMAP1-2
1.3 Goals and Objectives of EMAP-Forests 1-3
1.4 Assessment Endpoints 1-7
1.5 Legislative and Agency Mandates 1-7
1.6 History 1-8
1.7 Coordination 1-10
2 APPROACH AND RATIONALE
2.1 Research Design 2-1
2.2 EMAP Conceptual Framework 2-6
2.3 Forests as an Evolving Program within EMAP 2-16
2.4 Five-Year Strategy 2-16
3 STRATEGY FOR INDICATOR DEVELOPMENT AND
IMPLEMENTATION
3.1 Developing a Conceptual Framework for Assessment 3-1
3.2 Indicator Development Process 3-17
3.3 Implementation Plan for FY91 3-26
4 STRATEGY FOR MONITORING NETWORK DESIGN
4.1 General Statistical Requirements 4-1
4.2 Definitions of Populations and Sampling Units 4-1
4.3 Existing Forest Service Inventories and Monitoring
Programs 4-2
4.4 Overview of the Forest Service FIA Designs 4-3
4.5 Sampling Frame 4-5
4.6 Higher Grid Densities 4-8
5 STRATEGY FOR FIELD SAMPLING DESIGN
5.1 Introduction 5-1
5.2 Plot Selection Rules 5-1
5.3 Plot Design in the 1990 Pilot Studies 5-3
5.4 Plot Design Development 5-5
6 STRATEGY FOR STATISTICAL ESTIMATION AND ANALYSIS
6.1 Introduction 6-1
6.2 Status and Extent 6-1
6.3 Change and Trend 6-6
6.4 Associations 6-7
6.5 Methods for Integration of Information 6-8
-------
TABLE OF CONTENTS (continued)
7 STRATEGY FOR ASSESSMENTS
7.1 Introduction to Assessment 7-1
7.2 Status of Forest Assessments 7-1
7.3 A Strategy for FHM Assessments 7-3
7.4 Strategy Elements and Goals 7-11
6 QUALITY ASSURANCE PROGRAM
8-1 Introduction 8-1
8.2 Quality Assurance Policy 8-1
8.3 Total Quality Management 8-1
8.4 Organizational Structure 8-4
8.5 Quality Assurance Objectives 8-7
8.6 QA Documentation and Reporting 8-11
8.7 QA Operations 8-14
9 LOGISTICS APPROACH
9.1 Introduction 9-1
9.2 Logistics Issues 9-1
9.3 Organizational Structure 9-2
9.4 Logistics Implementation Components 9-3
10 STRATEGY FOR THE INFORMATION MANAGEMENT SYSTEM
10.1 Introduction to Information Management 10-1
10.2 Goals and Objectives 10-1
10.3 Design of the FHM IM System 10-2
10.4 Strategy to Move Toward the FHM IM System 10-8
11 STRATEGY FOR REPORTING
11.1 Current Status of Reporting 11-1
11.2 Reporting Descriptions 11-1
11.3 Forest Monitoring Plans 11-5
11.4 Operations Reports 11-6
11.5 Data Base Summaries 11-6
11.6 Data Quality Report 11-6
11.7 Statistical Summary 11-7
11.8 Interpretive Reports 11-8
11.9 Technical Proceedings 11-8
11.10 Organization of Reporting Effort 11-10
11.11 Action Plan 11-14
12 REFERENCES
13 GLOSSARY
iv
-------
FIGURES
Figure 1.1 Need for coordination among agencies involved in
EMAP-Forests 1-8
Figure 1.2 EMAP cross-cutting activities 1-11
Figure 2.1 EPA EMAP tier outline 2-2
Figure 2.2 The baseline grid for North America 2-5
Figure 2.3 The landscape characterization hexagons 2-6
Figure 2.4 Enhancement factors for increasing the base grid density 2-7
Figure 2.5 The environmental risk assessment program 2-7
Figure 2.6 TTie EMAP risk assessment framework 2-9
Figure 2.7 Structure of EMAP-Forests resource group as a module 2-12
Figure 2.8 Resource groups - hub relationship 2-13
Figure 2.9 EMAP conceptual organizational showing terrestrial resource
groups in context 2-14
Figure 3.1 Regional assessment 3-3
Figure 3.2 Forest health monitoring (FHM) assessment framework 3-9
Figure 3.3 Application of integrity index 3-13
Figure 3.4 Flow diagram representing the indicator development process 3-18
Figure 3.5 Indicator development and implementation strategies 3-19
Figure 3.6 Indicator development issues 3-22
Figure 3.7 Developmental indicator assessment objectives 3-23
Figure 3.8 Indictor reporting framework 3-24
Figure 3.9 Pointing toward causality 3-25
Figure 3.10 Status of EMAP-Forests indicator development 3-26
Figure 3.11 Relationships of measurements, indicators, and assessment
endpoints for maintenance of plant and animal diversity 3-28
Figure 3.12 Outcome of a joint planning meeting with the Forest Service 3-30
Figure 5.1 Plot design for the FY90 field season 5-4
Figure 6.1 Example of estimated distribution plots with upper confidence
bounds, generated both as numbers (upper plot) and area (lower plot)
(Linthurst et al. 1986) 6-4
Figure 7.1 Current status of assessment framework for forests 7-2
Figure 7.2 Conceptual forest ecosystem model 7-4
Figure 8.1 Relationships between quality assurance and total quality
management's primary tenet - "customer satisfaction" 8-2
Figure 8.2 EMAP commitment to total quality management 8-3
Figure 8.3 Proposed organizational structure for EMAP-Forests
quality assurance staff 8-4
Figure 8.4 Hierarchy of data quality objectives 8-8
Figure 8.5 The DQO process for continuous communication and feedback
among decision makers and scientists 8-9
Figure 8.6 The DQO continuous improvement process 8-10
Figure 8.7 Assessment and control of process errors within a measurment 8-13
Figure 8.8 Data collection process 8-15
v
-------
FIGURES (continued)
Figure 9.1 Partial example of a flow chart for describing
logistics staffing 9-3
Figure 9.2 FHM/EMAP 3-year Gantt chart of activities 9-6
Figure 9.3 Partial example of a flow chart for describing field staffing 9-8
Figure 9.4 Estimated time requirements for indicator implementation 9-8
Figure 9.5 Example of a sampling site access plan 9-13
Figure 9.6 Example of a communications plan 9-14
Figure 9.7 Flow of information to and from regional project leads 9-15
Figure 9.8 Development of a sampling schedule 9-17
Figure 11.1 Sequence of reporting in EMAP-Forests 11-4
Figure 11.2 EMAP-Forests reporting organization and liaisons 11-11
Figure 11.3 Forest Service mega-regions 11-12
vi
-------
TABLES
Table 2.1 A comparison of characteristics of the phases of monitoring 2-4
Table 2.2 Forest Service - FHM implementation schedule as currently
planned and the predicted number of FHM plots 2-18
Table 3.1 Examples of relationships between societal values, assessment
endpoints, and indicators 3-6
Table 3.2 Indicator selection criteria 3-11
Table 7.1 Auxiliary data bases: uses, sources, and acquisition intervals 7-10
Table 7.2 Atmospheric constituents of interest 7-10
Table 9.1 EMAP logistics elements for implementation of forests
monitoring programs 9-4
Table 9.2 List of supply needs for logistics 9-10
Table 9.3 Base site technical support requirements 9-11
Table 9.4 Safety information to be logged by field personnel 9-16
Table 11.1 Summary of reports to be produced by the
interagency forest monitoring program 11-2
Table 11.2 Summary of purposes of reports to be produced by the interagency
forest monitoring program 11-3
Table 11.3 Typical outline of a statistical summary 11-9
Table 11.4 Laboratory and national reporting unit assignments to produce
documents described in this chapter for EMAP-Forests 11-13
vii
-------
ACRONYMS
ADQ - audits of data quality
ADP - automated data processing
AFA - American Forestry Association
AREAL-RTP - Atmospheric Research and Exposure Assessment Laboratory, Research
Triangle Park, North Carolina
ARNEWS - Acid Rain National Early Warning System (Canada)
BIFC - The Boise Interagency Fire Center
BLM - Bureau of Land Management
C - element of carbon
cdf - cumulative distribution function
CPR - cardio-pulmonary resuscitation
DBH - diameter at breast height
DDRP - Direct/Delayed Response Project (acid deposition)
DLG - digital line graph
DQO - data quality objective
DRIS - Diagnosis and Recommendation Integrated System
DSI - data set index
EIC - EMAP information center
EMAP - Environmental Monitoring and Assessment Program
EMSL-LV - Environmental Monitoring Systems Laboratory, Las Vegas, Nevada
EPA - U.S. Environmental Protection Agency
EQO - ecosystem-level quality objectives
ERL-C - Environmental Research Laboratory, Corvallis, Oregon
FHM - multi-agency forest health monitoring program
FIA - Forest Inventory and Anatysis (USDA-FS program)
F1C - Forest Information Center
FPM - Forest Pest Management (USFS)
FRP - Forest Response Program (USFS)
FS - USDA Forest Service
FS-FHM - USDA Forest Sen/ice ongoing forest health monitoring program
viii
-------
FWS - Fish and Wildlife Service
FY - fiscal year (federal:October 1 - September 30)
GIS - Geographic Information System
GPS - Global Positioning System
HT - Horwitz-Thompson estimation formulas
IAG - Interagency agreement
IFB - fixed price contract
IM - information management
IQO - indicator-level quality objectives
K - element of potassium
LQAO - Laboratory Quality Assurance Officer
LTER - long-term ecological research
MOI - memorandum of intent
MQO - measurement-level quality objectives
MSR - management systems review (audit)
N - element of nitrogen
NAPAP - National Acid Precipitation Assessment Program
NE-FHM - New England forest health monitoring
NQAO - National Quality Assurance Officer
NAS • National Academy of Science
NFS - National Forest System units
NOAA - National Oceanic and Atmospheric Administration
NPS - National Park Service
NRC - National Research Council
NVS - National Vegetation Survey (USFS)
ORD - Office of Research and Development (USEPA)
P - element of phosphorus
PAR - Photosynthetically Active Radiation
PDR - portable data recorder
PEA - performance evaluation audits
QA - quality assurance
QAARW - quality assurance annual report and workplan
ix
-------
QAC
quality assurance coordinator
QAMS
quality assurance management staff
QAPjP
quality assurance project plan
QAPP
quality assurance program plan
QC
quality control
RDBMS
Regional Data Base Management System
RPA
Renewable Resources Planning Act
RQAO
Regional Quality Assurance Officer
RQO
resource-level quality objectives
RFP
competition negotiation
SAB
Science Advisory Board
SAF
Society of American Foresters
SCS
USDA Soil Conservation Service
S!
le Systeme International d'Unites (standard scientific units of measure)
SOP
standard operating procedure
TC
Technical Coordinator
TD
Technical Director
TQM
total quality management
TSA
technical systems review (audit)
USDA
U.S. Department of Agriculture
USDA-FS
U.S. Department of Agricutture, Forest Service
X
-------
ACKNOWLEDGEMENTS
The authors thank the members of the Peer Review Panel, Dr. Dan Binkley, Dr. George Furnival,
Dr. James Hornbeck, and Dr. Joseph Yavitt for their time in reading the document and their many
constructive comments and discussions.
Appreciation goes to the following individuals who contributed to various sections: Steve Cline,
Barbara Conkling, Ray Czaplewski, John Hazard, Chris Maser, Robert Mickler, Hans Schreuder, Don
Stevens, Rick Van Remortel.
The following individuals are acknowledged as the U.S. EPA Work Assignment Managers: Ralph
Baumgardner, Dan Heggem, Ann M. Pitchford, and Parker J. Wigington.
The authors also gratefully acknowledge the reviews of the following individuals: John Baker, Joe
Barnard, Robert Brooks, Bill Burkman, Dean Carpenter, Don Charles, Ray Czaplewski, Robert Graves,
Mark Hansen, John Hazard, Karl Hermann, Bruce Jones, Robert Kucera, Gene Meier, Tony Olsen,
Susan Peck, and Terry Shaw.
Appreciation goes to Barbara Conkling and John Nicholson for their many contributions as technical
editors and to Paula Showers, Lillian Steele, Pam O'Bremski, Celenthia McPherson, Angelina Viray,
Kelly Fuller, Steve Garcia, Lynn Gurzinski, and Frank Insinga for their high quality word processing and
graphics assistance.
xi
-------
1 INTRODUCTION
To protect, manage, and use forest
resources effectively, the condition of these
resources must be known. Concern about
documented and potential effects of air
pollutants in combination with other multiple,
interacting stresses has been a major
impetus behind the development of moni-
toring programs in forests. During the
past two years, the forest component of
the Environmental Protection Agency's
Environmental Monitoring and Assessment
Program (EMAP-Forests) has been working
closely with the Forest Service's Forest Health
Monitoring (FS-FHM) program and other
government agencies to develop a multi-
agency program to monitor the condition of
the nation's forested ecosystems. In this
document, this future multi-agency program
will be simply referred to as the Forest Health
Monitoring (FHM) program.
The purpose of this document is to
present a strategy that can be used as a
starting point by all government agencies
interested in participating in a nation-wide
FHM program. Monitoring issues such as
design, indicator selection, and assessment
are presented along with approaches to
resolving these issues. We ask your
assistance in evaluating whether or not these
approaches are sound and the strategy is
adequate to evaluate the ecological condition
of our nation's forests.
The purpose of this introductory
section is to provide an overview of the scope
and purpose of this document. The contents
of each section will be reviewed. An overview
of the overall Environmental Monitoring and
Assessment Program (EMAP) will be provided
along with the fundamental research
questions motivating the development of the
program. Specific EMAP-Forest goals and
objectives designed to answer these questions
for forested ecosystems will be presented. A
short historical background of the
development of the EMAP-Forests program
and the FS-FHM program will be given to help
the reader put the present planning process in
perspective. An important theme of this
section is that a multi-agency FHM program
can be successful only through effective
coordination.
1.1 CONTENT AND ORGANIZATION OF
EMAP-FORESTS STRATEGY PLAN
This plan is organized into twelve
sections. Sections 2 - 7 present the scientific
approach currently proposed for the
establishment of a forest health monitoring
(FHM) program, including design, indicators,
and assessment concepts. Sections 9-11
present aspects of logistics, quality assurance
(QA), and information management (IM). A
description of each section follows:
• Section 2. Approach and Rationale -
highlights the overall proposed strategy for
EMAP-Forests.
• Section 3. Strategy for Indicator
Development and Implementation -
explains the process for selecting and
testing indicators of forest condition.
• Section 4. Strategy for Monitoring and
Network Design - describes the statistical
issues related to sampling of forests on a
regional basis.
• Section 5. Strategy for Field Sampling
Design - describes specific design issues
related to plot establishment.
• Section 6. Strategy for Statistical
Estimation and Analysis - details the
statistical approach for the evaluation of
status and trends of indicators.
• Section 7. Strategy for Assessment-
depicts the strategy for integrating the
statistical information into ecologically
meaningful statements regarding forest
condition.
• Section 8. Quality Assurance Program -
identifies a total quality management
1-1
-------
(TQM) approach to ensure that the data
collected are of sufficient quality to meet
the data quality requirements of the data
users.
• Section 9. Logistics Approach - explains a
strategy for conducting a field program of
this magnitude.
• Section 10. Strategy for a Joint
Information Management System - defines
how information management supports
data collection, data evaluation, and
reporting functions.
• Section 11. Strategy for Reporting -
describes the type and organization of
reports needed to disseminate Information
to clients.
1.2 OVERVIEW OF EMAP
As the 1990s begin, the United States
has begun the task of protecting the integrity
and sustainability of the nation's ecosystems.
This is critical because the incidence and scale
of reported environmental problems have
increased during the past two decades.
Scientists and the general public are
increasingly concerned about environmental
problems such as global climate change,
acidic deposition, and loss of biological
diversity. Scientific studies have heightened
environmental awareness and indicate that
ecological processes determining how
ecosystems respond to natural and
anthropogenic disturbances are complex.
Research has also framed several
fundamental questions about the severity of
these disturbance effects and whether or not
they are changing in response to government
policies. Unfortunately, the answers to these
questions are not readily apparent because
the status of the nation's environment is not
well documented. A baseline is needed
against which we may evaluate measured
changes in the condition of resources and the
overall effectiveness of national environmental
policies.
In 1988, the U.S. Environmental Protec-
tion Agency's (EPA) Science Advisory Board
(SAB) recommended the implementation of a
program to monitor ecological status and
trends and to develop innovative methods for
anticipating emerging environmental problems
before they reach crisis proportions. In
response to these recommendations, EPA's
Office of Research and Development (ORD)
began planning the Environmental Monitoring
and Assessment Program (EMAP). Several
key questions were formulated to guide the
program toward meeting the goals set by the
SAB:
1) What is the current extent of our ecological
resources, and how are they distributed-
geographically?
2) What proportions of the resources are
currently in acceptable ecological condi-
tion?
3) What proportions are degrading or improv-
ing, in what regions, and at what rates?
4) Are these changes correlated with patterns
and trends in environmental stresses?
5) Are adversely affected resources improving
in response to control and mitigation
programs?
In 1989 an EMAP approach emerged to
address these questions. The EMAP approach
will be presented in greater detail in Section 2
of this document, but the basic tenets are
presented below:
1) Monitoring should be on a broad regional
scale to provide quantitative and unbiased
estimates of status and trends in
ecological condition.
2) All ecological components of the landscape
in all regions should be monitored (i.e.,
there should be no "orphan" ecosystems).
3) Successful implementation will require a
long-term commitment.
4) The scope of the program will require a
multi-agency effort.
1-2
-------
5) The analysis of ecological condition will
require measurements of habitat condition,
pollutant sources and exposure, and
biological condition.
6) The program will need to focus on ecology
as well as traditional monitoring activities.
EMAP comprises seven ecological
resource groups: Agroecosys terns, Arid
Lands, Forests, Great Lakes, Near Coastal
Systems, Surface Waters, and Wetlands. The
goal has been to ensure that EMAP monitors
all major ecological resources. The planning
efforts in these resource groups are supported
by a number of cross-cutting activities such as
design or logistics. This document describes
the development of EMAP-Forests toward the
FHM program.
1.3 GOALS AND OBJECTIVES OF EMAP-
FORESTS
To focus the EMAP-Forests program
development, goals and objectives have been
specified. The overall goal of EMAP-Forests is
to develop and implement a program to
monitor, evaluate, and report on the long-term
status and trends of the nation's forest
ecological resources as these resources relate
to changes in and among natural phenomena,
resource management practices, and
pollutants across the landscape.
To coordinate EMAP-Forests efforts
with those in other terrestrial resource groups
(EMAP-Arid Lands and EMAP-Agroeco-
systems), forest land has been defined as
land with at least 10 percent of its surface
area stocked by trees of any size or formerly
having had such trees as cover and not
currently built up or developed for agricultural
use (USDA Forest Service 1989).
For purposes of clarity, a tree has
been defined as a woody plant with one or
more perennial stems at least three inches in
diameter at breast height at maturity, with a
more or less definitely formed crown of foliage
and a height of at least sixteen feet at
maturity (USDA Forest Service 1989).
It is important to review the overall
goal and highlight certain aspects. Although
the EMAP-Forests definition of forest land is
based on trees, the goal of EMAP-Forests is to
assess the health of forested ecosystems and
not just trees. This assessment must be
based on components of forest condition that
reflect the variety of values which society
associates with forests. Societal values
include the variety and abundance of plant and
animal life, clean air, clean water, and fertile
soils. Forested ecosystems have traditionally
provided economic values for those who
directly or indirectly derive their livelihood from
the utilization of forest resources. The values
which society now places in many forests for
recreation, aesthetics, and a place to observe
nature in its pristine state are also important
considerations for EMAP-Forests.
Many times, these different values
result in conflicts regarding the management
of forest lands and the assessment of health
or condition. For example, an old growth
forest may not be as efficient as a newly
planted forest in capturing light and converting
it to wood production, but it may be valued for
its habitat for certain types of wildlife. EMAP-
Forests must obtain objective measures of all
values society places in forests.
In addition to assessing forests for
ecological and societal values, EMAP-Forests
also intends to assess the various stresses
on forested ecosystems. These include
natural stresses such as climate, pests, and
anthropogenic stresses (i.e., management
action and pollution). The evaluation of
pollutant stresses and their relationship to
forest condition are of particular importance to
EMAP-Forests and the EPA.
1-3
-------
The overall goal has been developed
into some specific objectives. These detailed
objectives, which are presented in the
following subsections, have been designed to
parallel the overall EMAP objectives which are:
1) Estimate the current status, extent,
changes, and trends in indicators of the
condition of the nation's ecological
resources on a regional basis with known
confidence.
2) Monitor indicators of pollutant exposure
and habitat condition and seek
associations between human-induced
stresses and ecological condition.
3) Provide periodic statistical summaries and
interpretive reports on ecological status
and trends to resource managers and the
public.
1.3.1 Short-term, Mid-term, and Long-term
Objectives of EMAP-Forests
1.3.1.1 Status and Trends of Ecological
Resource Indicators
EMAP-Forests will provide unbiased,
regional estimates of the status and trends of
indicators of ecological resources in forests,
which includes productivity (food and fiber),
animal diversity, plant diversity, water quality,
and aesthetics on an annual basis for all
ecoregions of the United States.
1.3.1.1.1 Short-term Objectives (1-5 years)
By 1995, EMAP-Forests will resolve the
following technical issues:
• Development of a conceptual framework
for integrating indicator information into
assessment endpoints representative of all
values placed in forests.
• Resolution of frequency of plot sampling.
• Strategy for evaluating indicators as to
their accuracy in representing ecological
condition and pollutant exposure.
• Plot design for uniform and non-uniform
stands.
• Evaluation of multi-stage remote sensing
for characterizing landscape processes.
• Linkages of landscape indicators with plot
level indicators.
By 1995, EMAP-Forests will develop an
effective forest health monitoring program
that:
• Is a multi-agency program with the FS and
other federal and state agencies.
• Has the active support of EPA line
management at participating EPA
laboratories.
• Has an efficient indicator development pro-
gram which includes the participation of
the outside scientific community in
identifying and testing of new indicators.
• Is characterized by the collection of high
quality data through the coordinated
activities of QA, logistics, and information
management.
• Meets the needs of data users through
rapid turnaround of data by electronic cap-
ture, transfer, and easy access.
By 1995, EMAP-Forests will develop as
products of forest health monitoring:
• A set of at least four core indicators with
documented spatial, temporal and
measurement components of variance and
detailed methods manuals for each region.
• Annual statistical summaries of the
status and trends for these indicators in
regions where forest health monitoring has
been implemented.
• A QA program plan and associated project
plans for each region.
1.3.1.1.2 Mid-term Objectives (6-10 years)
By 2000, EMAP-Forests will improve
technical capabilities of the program by:
1-4
-------
• Testing additional indicators for incorpor-
ation into the set of core indicators.
• Expanding the multi-agency forest program
to include all other forest land manage-
ment agencies.
• Providing interpretive assessments on
critical endpoints for U.S. forests.
• Improving ability to obtain indicator
information from satellites such as NASA
EOS.
1.3.1.1.3 Long-term Objectives (11+ years)
By 2005, EMAP-Forests will demon-
strate the effectiveness of the forest health
monitoring program by:
• Providing unbiased estimates of the status
and trends in forest condition for all
regions of the U.S.
• Selecting statistical estimators for spatial
patterns, trend detection, and ecological
modeling.
• Providing a full assessment of the
components of variance for all indicators in
all regions.
1.3.1.2 Status and Trends of Stress Effects on
Ecological Resource Indicators
EMAP-Forests will provide unbiased,
regional estimates of the status and trends of
indicators of stresses on ecological resources
in forests that include land management
practices, pollutants, and natural stresses on
an annual basis for all ecoregions of the
United States.
1.3.1.2.1 Short-term Objectives (1-5 years)
By 1995, EMAP-Forests will resolve the
following technical issues:
• Identification of key indicators for
monitoring for air pollutant stress and the
accuracy needed for these variables.
• Evaluation of the relationship between air
pollution and visual symptom indicators of
air pollution stress on forest vegetation.
• Selection and testing in a pilot mode key
indicators of land use practices and other
management stresses on forests.
• Incorporation of forest pest management
surveys into annual statistical summaries
of forest health.
• Evaluation of important climatic variables
and their correlation to forest condition.
• An evaluation of indicators of the
accumulation of toxins in foliage and soils.
By 1995, EMAP-Forests will develop an
effects process for incorporating stress
information into forest health monitoring by:
• Integrating climatic data and air pollution
data into all annual statistical summaries.
• Developing a strategy for completing
detailed landscape characterization around
established field plots to document
management stresses.
• Involving FS Forest Pest Management staff
in detection level monitoring activities.
By 1995, EMAP-Forests will provide the
following products:
• Maps of air quality for regions where
forest health monitoring has been
implemented.
• Maps of climatic stress in the same
regions.
• Maps of other natural stresses in those
regions.
1.3.1.2.2 Mid-term Objectives (6-10 years)
By 2000, EMAP-Forests will improve
technical capabilities of the program by:
• Field testing inexpensive air pollution
monitors at plot sites.
• Developing and testing models for im-
proving the estimation of climatic stresses
1-5
-------
at plots from off-frame monitoring
locations.
• Providing interpretive assessments for all
endpoints in U.S. forests.
1.3.1.2.3 Long-term Objectives (11+ years)
By 2005, EMAP-Forests wilt demon-
strate the effectiveness of the forest health
monitoring program by:
• Developing an integrated index of forest
stress.
• Identifying regions where excessive air
pollution or management stress is
associated with poor forest conditions.
• Initiating research projects to further
evaluate why certain forest resources are
in poor condition.
1.3.1.3 Assessment and Reporting
EMAP-Forests will provide to the
Administrator and the public annual statistical
summaries and assessments that evaluate
the associations between indicators of forest
condition and indicators of stress on these
ecosystems.
1.3.1.3.1 Short-term Objectives (1-5 years)
By 1995, EMAP-Forests will resolve the
following technical issues:
• Development of method for incorporating
off-frame indicator data into forest health
assessments.
• Development of method for reducing the
uncertainty in regional forest condition
estimates through the use of other
monitoring data such as forest inventory
and analysis.
By 1995, EMAP-Forests will improve the
ability to provide interpretative assessments
by:
• Identifying appropriate areas for program
integration (e.g. the development of
assessment infrastructure) among the
EPA, the FS and other cooperating
agencies.
• Obtaining appropriate equipment and
analytical tools for assessments.
• Identifying regional forest assessment
units to complement the national program.
By 1995, EMAP-Forests will provide the
following products:
• An annual statistical summary that
includes indicator data from all regions
where regional forest health monitoring
has been implemented.
• Summaries of all indicator development
projects funded by EMAP-Forests.
1.3.1.3.2 Mid-term Objectives (6-10 Years)
By 2000, EMAP-Forests will improve its
assessment capabilities by:
• Providing its first interpretative report of
forest condition for a region.
¦ Including landscape level assessments for
regions where landscape characterization
has been completed.
• Contributing to the integration of data
among EMAP resource groups.
• Providing assistance to other countries
interested in developing similar programs.
• Integrating forest health monitoring
assessments into the EPA regional risk
assessment framework.
1.3.1.3.3 Long-term Objectives
By 2005, EMAP-Forests will demon-
strate the effectiveness of assessment
capabilities by:
• Providing integrated among EMAP resource
group interpretative assessments by
ecoregion.
1-6
-------
• Demonstrating the associations between
indicators of forest condition and stress
on forested ecosystems.
• Cooperating in the expansion of EMAP to
high priority areas (e.g., tropical forests) to
address international monitoring issues
such as global change.
1.4 ASSESSMENT ENDPOINTS
The EMAP-Forests approach is struc-
tured around a suite of assessment endpoints
defined for indicators of forest condition.
Assessment endpoints are quantitative goals
for these indicators (Suter 1990; Messer 1990).
The assessment endpoints provide an
appropriate basis for structuring a program
because they wilt satisfy the public demand
for relevant environmental information. The
challenge of EMAP-Forests is to identify
indicators that can be combined in a
quantitative manner to make overall
statements of status and trends in
assessment endpoints. Section 3 presents a
more complete discussion of indicators and
assessment endpoints.
1.5 LEGISLATIVE AND AGENCY
MANDATES
The EPA and the FS have received
specific directions from Congress to address
the effects of air pollution on forests. Title IX
of the Clean Air Act amended (1990) states:
"In carrying out subsectbn (a), the
Administrator, in cooperation, where
appropriate, with the Under Secretary of
Commerce for Oceans and Atmosphere, the
Director of the Fish and Wildlife Service, and
the Secretary of Agriculture shall conduct a
research program to improve understanding of
the short-term and long-term causes, effects,
and trends of ecosystems damage from air
pollutants on ecosystems. Such program
shall include the following elements: (1)
Identification of regionally representative and
critical ecosystems for research. (2)
Evaluation of risks to ecosystems exposed to
air pollutants, including characterization of the
causes and effects of chronic and episodic
exposure to air pollutants and determination of
the reversibility of those effects. (3)
Evaluation of the effects of air pollution on
forests, materials, crops, biological diversity,
soils, and other terrestrial and aquatic
systems exposed to air pollutants."
In 1988, the Forest Ecosystems and
Atmospheric Pollution Research Act (Public
Law 100-521) explicitly authorized the FS to
undertake the necessary monitoring to track
long-term trends in the health and productivity
of forest ecosystems in the United States:
"The Secretary, acting through the United
States Forest Service, shall (a)increase the
frequency of forest inventories in matters that
related to atmospheric pollution and conduct
such surveys as are necessary to monitor
long-term trends in the health and productivity
of domestic forest ecosystems..."
In addition to these agencies, other
agencies have specific mandates regarding
the management and protection of forested
ecosystems. Examples include the National
Park Service, the Bureau of Land Management,
and State Forestry agencies. Other agencies
such as the Soil Conservation Service and the
Fish and Wildlife Service have specific
mandates regarding the condition of certain
aspects of forested ecosystems.
A representation of the need for
coordination among these agencies is given in
Figure 1.1. Forest condition at the center of
the diagram is impacted by a combination of
air pollutant, natural, and management
stresses. A multi-agency monitoring program
such as forest health monitoring is being
designed to monitor the cumulative impact of
these stresses on forest ecosystem condition.
When forests are found to be in an adverse
1-7
-------
NATURAL STRESSORS
e.g. climate, pests, fire
POLLUTANT STRESSORS
e.g. acid rain, ozone
FOREST ECOSYSTEM
CONDITION
f
LAND MANAGEMENT
STRESSORS
e.g. land use,
harvesting procedures
ENVIRONMENTAL
PROTECTION
AGENCIES
FOREST
HEALTH
MONITORING
LAND
MANAGEMENT
AGENCIES
POLLUTANT CONTROL
POLICIES
LAND MANAGEMENT
POLICIES
Figure 1.1. Need for coordination among agencies involved in EMAP Forests.
condition, detailed studies will be implemented
to identify the cause. If a linkage to air
pollution or other form of pollution is found to
exist, then this information can be provided to
the EPA policy analysts. If the problem can be
attributed to land management actions, then
this information could influence land
management policies. If natural stress is
identified as the cause, then no applicable
corrective actions may be required.
1.6 HISTORY
Two important FHM projects were
undertaken during the summer of 1990. Prior
to discussing these projects, a short history of
the development of EMAP-Forests and FS-FHM
will be provided.
1.6.1 History of EMAP-Forests
Planning for a program in forests
within EMAP began with the appointment of
Richard Olsen as the EMAP-Forests Technical
Director in July, I988. Prior to this
appointment, Mr. Olsen had served as the
Manager of the Western Conifers Research
Cooperative in the Forest Response Project
(FRP), a research program conducted under
the auspices of the National Acid Precipitation
1-8
-------
Assessment Program (NAPAP) Task Group V-
Terrestrial Effects. "Hie FRP was funded by
the FS, EPA, and the National Council of the
Paper Industry for Air and Stream
Improvement.
Mr. Olsen immediately began an
assertive program to develop a national
monitoring plan. He organized a series of
workshops to identify candidate indicators of
forest health and initiated work on a
monitoring design. Due to his previous
experience in the interagency FRP, he made a
special effort to develop EMAP-Forests in
cooperation with similar efforts in the FS.
In April 1989, Mr. Olsen was re-
assigned to a new program and Craig Palmer
assumed the role of EMAP-Forests Technical
Director while continuing in his position as
Quality Assurance Officer for the FRP. The
EMAP program focus at this time was the
selection of indicators of ecological condition.
EMAP-Forests staff conducted a series of
workshops resulting in a preliminary set of
indicators and prepared a chapter for the
overall EMAP indicators report (Hunsaker and
Carpenter 1990). Throughout this period, the
FRP served as useful tool to foster
communication between the EPA and the FS in
the development of a possible multi-agency
FHM program.
In November 1989, the EMAP program
conducted a competition between the resource
groups for the funding of FY90 field activities.
The EMAP steering committee selected the
EMAP-Forests proposal for funding. The focus
of this proposal was a field evaluation of
forest health indicators in cooperation with
proposed FS monitoring activities in New
England.
1.6.2 History of FS-FHM
The FS began its development of a
forest health monitoring program approx-
imately five years ago. With the development
of the Acid Rain National Early Warning
System (ARNEWS) forest monitoring project in
Canada, momentum grew for the development
of a similar program in the United States. A
series of pilot studies were undertaken as part
of the FRP under a project called the National
Vegetation Survey (NVS) directed by Joe
Barnard of the Southeast Experiment Station.
One objective of the NVS was to
develop techniques to inventory and monitor
symptoms of atmospheric pollution-induced
stress, damage, and/or death of forest stands
and trees. A select group of indicators for
determining forest condition was identified in
workshops in 1987 and implemented during
pilot tests in 1988 and 1989. Visual damage
survey plots were established in mixed
hardwood forests (128 plots), high elevation
spruce fir forests (31 plots), and natural
loblolly pine stands in the piedmont (157 plots)
and coastal plain (222 plots) regions.
During this time, a committee was
formed to develop a long-term forest health
monitoring approach for the FS. A multi-tiered
process was proposed consisting of three
levels with increasingly detailed monitoring.
These will be discussed in section 2. This
proposal was circulated in the FS in the fall of
1989. Top-level managers within the FS also
met with representatives of the New England
states to discuss implementation of a forest
health monitoring program in 1990. On
November 20, 1989, agreement was reached
between the FS and the New England states
to begin such a program. Shortly thereafter, a
regional project manager was appointed (Bob
Brooks) and a national FS-FHM coordinator
was selected (Joe Barnard).
1.6.3 History of 1990 Field Projects
The period after November 20, I989
was filled with planning activities within EMAP-
Forests and the FS-FHM program. Although
1-9
-------
both agencies intended to undertake forest
health monitoring field activities in 1990, no
mechanism existed to ensure that one
program could be developed to meet the
needs of both agencies. Two separate
motivating factors were driving the planning
process in each agency. EMAP-Forest staff
were preparing a defense of proposed
indicators and therefore recommended
indicator evaluations on a regional basis. The
FS staff were attempting to begin the
implementation of a long-term monitoring
program and therefore desired to build upon
indicators already evaluated in the visual
symptom pilot tests of the NVS.
In late March 1990, a meeting was held
in Portsmouth, New Hampshire with
participants from EMAP-Forests, the FS-FHM
national program, and the FS-FHM New
England program. At this meeting it was
decided that two separate projects be
undertaken in 1990. One project would
establish long-term monitoring plots using the
EMAP grid (described in Section 2) and would
include visual symptoms measurements. This
project would be championed by the FS with
the assistance of state field crews. This effort
would be called New England Forest Health
Monitoring (NE-FHM). EMAP-Forests staff
were to provide assistance in QA and IM.
A second project was proposed for a
f ield evaluation of several additional indicators.
EMAP-Forests staff were to take the lead in
planning this project but the field work was to
be implemented under the direction of the FS.
It was suggested that approximately twenty
plots be established in New England in
northern hardwoods and twenty plots be
situated in Loblolly pine stands of Virginia on
sites that would not become FHM plots. As a
result, this second project was named the
20/20 pilot study. Subsequent workshops
resulted in the selection of five indicators for
evaluation in the 20/20 study including growth
efficiency, visual symptoms, soil productivity,
foliar nutrients, and vertical vegetation
structure.
Both projects were successfully
undertaken during the 1990 field season. In
the NE-FHM project, over two hundred plots
were established across New England. The
percentage of plots established in each forest
type closely approximated the relative
abundance of the various forest types found in
New England. The 20/20 project was also
successful in achieving the majority of its
objectives. Final reports are currently in
preparation for these projects. Present
planning efforts are directed towards the
expansion of FHM plots to six new states with
indicator development activities occurring
simultaneously on a subset of these plots.
The success of these projects has
provided a strong basis for cooperative efforts
in the 1991 field season. If a multi-agency
FHM program is developed, it will owe a debt
of gratitude to its origins that were fostered
and cultivated in the FRP.
1.7 COORDINATION
Coordination with components within
and outside of EMAP is fundamental to the
success of the EMAP-Forests strategy. This
subsection identifies target groups with which
EMAP-Forests must establish a working
relationship, explains why coordination with
these groups is important, and describes
efforts to coordinate specific strategies in
FY91.
1.7.1 Coordination Within EMAP Cross-
cutting Activities
To support resource monitoring
activities, EMAP has developed a number of
cross-cutting activities (see Figure 1.2). The
activities include air and deposition, landscape
characterization, information management,
integration and assessment, statistics and
1-10
-------
EMAP CROSS-CUTTING ACTIVITIES^
INTEGRATION ACTIVITIES
Ain Aisd DeposiriON
LANdsCApE
CliAnACiEnizATioN
iNfoHMAfiON
Manaqemcni
InteqoAliON AC«d
Assessment
COORDINATION
ACTIVITIES
SrArisiics And Desiq*
IndiCAfORS
Loqisrics
ToiaI QuAliiy
Manaqemeni
TEchwoloqy TnAissfsn
iNfERNAliONAl
Ac T ivi l its
DEVELOPMENTAL
RESEARCH
EnvironmentaI
StAiisiics
~
EcoloqicAl ImdiCAion
DeveIopmem
LANdscApE Ecology
EcoloqicAl Risk
C^ARACrERiZAliOKI
Figure 1.2. EMAP cross-cutting activities.
design, indicators, logistics, TQM, and
technology transfer. These groups provide
support in areas such as developmental
research in environmental statistics, ecological
indicator development, landscape ecology, and
ecological risk characterization. These cross-
cutting activities, each of which is headed by
a technical coordinator (TC), enhance overall
program uniformity and optimize the use of
planning resources.
Cooperation and coordination will
prove mutually beneficial to EMAP-Forests and
these cross-cutting activities. EMAP-Forests
will gain insight from the specialized tasks of
each cross-cutting activity. On the other hand,
the cross-cutting activities will benefit from the
lessons learned by EMAP-Forests, one of the
first resource groups to implement field
operations. EMAP-Forests can also influence
the development of overall EMAP approaches
1-11
-------
by supporting the development of guidance
documents for these cross-cutting activities.
Currently, EMAP-Information Manage-
ment and EMAP-Quality Assurance have
EMAP-wide committees that consist of
resource group information managers and QA
officers, respectively. EMAP-Information
Management also has an EMAP-wide
committee of geographic information systems
(QIS) analysts that addresses spatial data
management, spatial analysis, and reporting
issues. EMAP-Forests will encourage the
other cross-cutting activities to form similar
committees to conduct workshops, review
planning documents, develop training
procedures, and guide development activities.
Additionally, EMAP-Forests has assigned a
lead for each cross-cutting activity and then,
where possible, located that lead at the same
EPA laboratory as the TC for the EMAP cross-
cutting activity. For example, the EMAP TCs
for design and indicator development are
located at the EPA Environmental Research
Laboratory at Corvallis, Oregon. EMAP-
Forests leads in design and indicator
development are located there also, thus
enhancing communication and coordination
between the groups.
EMAP-Forests personnel will actively
participate in the development of both the
EMAP-Integration and Assessment and EMAP-
Landscape Characterization strategies to
assist in ensuring that all needs are being
adequately addressed and that efforts are
being coordinated. One goal of EMAP is to
use integrated assessments to report on
environmental health at regional levels. This
methodology implies an integrated program
that must be well-coordinated among the
various components. The development of
linkages between various levels of information
is paramount for success in this integration
process. The efforts of EMAP-Integration and
Assessment and EMAP-Landscape Characteri-
zation are central to how these linkages are
established.
During 1991, EMAP-Forests staff will
coordinate a number of specific efforts with
EMAP cross-cutting activities. For example,
EMAP-Forests staff will serve on the Clean Air
Act Committee with EMAP-Air and Deposition
staff. Joint pilot activities will be explored with
EMAP-Landscape Characterization. A pilot
project on the global positioning system will
be conducted with the EMAP-Information
Management staff. EMAP-Forests staff will
also work with the design TC to evaluate
monitoring design alternatives, the indicator
development TC to develop detailed objectives
for regional demonstration studies, and the
logistics TC to explore the possibility of multi-
agency regional logistics centers. In addition,
EMAP-Forests personnel will participate in the
development of QA guidance documents for
terrestrial ecosystems with the EMAP QA
Officer. EMAP-Forests will also participate
in an EMAP-wide user needs analysis
process that will identify informational,
hardware/software, staffing, and data needs
of the various program components.
1.7.2 Coordination among Resource Groups
Coordination among resource groups
is needed to ensure that all ecosystems are
monitored and to prevent duplication of
efforts. EMAP planners recognize that the
current EMAP approach to dividing the
landscape is relatively arbitrary and could be
accomplished by many different approaches.
An advantage of the EMAP divisions is that
they encourage cooperation with other
agencies with specific mandates such as the
FS (forests), the U.S. Fish and Wildlife Service
(wetlands), or the USDA (agricultural lands).
Most importantly, these resource groups are
interrelated in nature; therefore they should be
1-12
-------
interrelated conceptually. For example,
surface water quality is affected by forest
management practices, and wildlife are affect-
ed by activities across the landscape. For this
reason, it is important that monitoring be
developed In such a way that the various
resource groups can be integrated.
Coordination among resource groups
offers other advantages. Similar indicators
are often selected by different resource
groups. For example, soils have been selected
as an indicator by all terrestrial resource
groups. Soil scientists who are chosen as
indicator leads can assist and coordinate soil
indicator development activities across
resource groups. This approach not only
saves money, but it also fosters the
integration of data collected by the various
groups. Lessons learned by one resource
group can be used by another.
EMAP structure also encourages
coordination among some resource groups.
All terrestrial resource groups are under the
direction of one associate director.
Coordination is fostered through workshops,
biweekly conference calls, coordinated peer
reviews, and the sharing of some indicator
leads. Unfortunately, this coordination has not
extended very well to other resource groups.
Recognizing the need for better coordination
between all resource groups, all of the EMAP
TDs meet on a quarterly basis and participate
in biweekly conference calls.
A number of specific coordination
issues need to be resolved this year. For
example, the status of the pinion juniper is in
question because it fits the definition of both
forests and arid lands resource groups.
EMAP-Forests and EMAP-Agroecosystems
need to resolve the status of woodlots. In
addition, surface water nitrate is one of the
proposed indicators for EMAP-Forests, and it
is also one of the measurements made by
EMAP-Surface Waters. However, EMAP-
Surface Waters may not sample a surface
water on the same watershed associated with
forest monitoring even if they sample in the
same hexagon. To require them to do so
would bias their sampling scheme. Forested
wetlands is another unresolved area. This
year, EMAP-Forests staff will need to work in
conjunction with the EMAP-Wetlands group to
develop protocols for sampling wet or
submerged soils. EMAP-Near Coastal is
interested in land uses in watersheds
impacting near coastal ecosystems. The
characterization of landscapes will need to be
coordinated with EMAP-Forests efforts in
landscape characterization.
1.7.3 Coordination With the U.S. Forest
Service
The importance of coordinating with
the FS has been a theme of this plan. This
coordination focuses on the FS-FHM but also
includes coordination with regional groups
responsible for monitoring. Ultimately, the
goal is to develop a truly multi-agency program
with common goals, objectives, and
approaches that meet the needs of all
agencies.
During the past year, efforts have been
undertaken to encourage the development of
a multi-agency program. National planning
meetings have been held on a quarterly basis
to develop a multi-agency monitoring strategy.
At these meetings, special committees have
been organized to address the issues
specifically. A representative of each of these
committees participates in a weekly
conference call to review progress in planning
or field activities. On occasion, these
committees have held special meetings to
resolve issues. For example, the design
committee has held two national planning
meetings and has invited many of the best
statisticians from the FS and the EPA to
attend.
1-13
-------
During FY91, the principal coordination
objective will be to develop a multi-agency
forest health monitoring plan. This EMAP-
Forests strategy document is the first step in
that direction. Additional workshops will be
required and detailed studies will need to be
completed to resolve many of the issues
presented in this plan. An important
component of this year's efforts will be an
outreach program. It will include individuals
from all regions of the country in the national
planning effort.
1.7.4 Coordination Within the EPA
A number of divisions in the EPA have
expressed interest in EMAP-Forests, including
the regional offices, the program offices, and
other research programs such as the global
climate change program and the ozone
program. EMAP-Forests has not adequately
addressed the needs of these clients. During
FY91, EMAP-Forests will renew its efforts to
better inform and coordinate with these
groups. Regional liaison individuals will be
identified and contacted. Program offices will
be contacted, and individuals with interest in
the ecological condition of forests will be
identified. Principal scientists who work at the
EPA on a program that has a forest
component will also be contacted. All of these
individuals will be sent information packets on
EMAP-Forests. Client-need surveys, which will
include written and oral interviews will be
conducted. Invitations will be sent to attend
field tours or to participate in planning
meetings. The EMAP Monitor (a newsletter)
and EMAP task descriptors will be sent to
individuals who express an interest in EMAP-
Forests. A list of EMAP-Forests publications
and a request sheet will be sent to them on a
periodic basis.
1.7.5 Coordination With Other State and
Federal Agencies
In addition to the FS, other government
agencies have a responsibility for the
management of forest land. These agencies
include state forestry agencies, the BLM, and
the NPS. The support of all of these agencies
is needed for a successful nationwide forest
health monitoring effort. The permission of
these agencies will be needed for access to
the lands they manage. Most states,
management agencies, and the NPS have
already initiated forest health monitoring
activities. As a result, forest scientists in
these agencies have developed expertise that
would enhance the planning and
implementation of the FHM effort. These
scientists have pointed out that they can
benefit by ensuring that comparable data are
collected to allow them to evaluate their
results in a regional perspective.
Coordination efforts with state forestry
agencies began in 1990 in New England
through the implementation of forest health
monitoring with state crews. The FS has
worked well with state agencies and has led
the effort to coordinate with the states. The
current national strategy is to implement FHM
on a state-by-state basis.
This year other agencies must be
included in the national planning effort. They
can contribute to the effort and will more
readily cooperate with the implementation of
monitoring on their forest lands if they have
had some input during the planning process.
Forest health monitoring will expand to the
western United States in 1992 and agencies
with responsibility in this region must be
contacted. Representatives of the agencies
will be invited to attend national planning
meetings and participate in workshops to plan
pilot activities for their regions.
1-14
-------
1.7.6 Coordination With Research
Organizations
To be successful, FHM must be
founded on sound scientific principles.
Research scientists at universities and state
and federal research centers can assist in the
development of the program in many ways.
During the planning process, researchers can
assist in the identification and selection of
candidate indicators and in the development of
a conceptual framework for interpreting the
relationships among the indicators. Current
research sites offer many opportunities for
evaluating new indicators. Examples include
the National Science Foundation- funded Long-
term Ecological Research (LTER) sites, the
Man in the Biosphere reserves, and the
National Energy Parks.
Coordination can be accomplished if
scientists are informed of developments in the
planning and implementation of FHM. For
example, a presentation on forest health
monitoring was made at the 1990 fall LTER
meeting and at the International Ecological
Indicator Symposium. A synthesis of the
EMAP-Forests indicator strategy has been
submitted to a scientific journal.
During this coming year, represen-
tatives of the scientific community will
evaluate and peer-review all EMAP-Forests
plans and reports. Indicator development
proposals will be solicited and funded for the
detection level of monitoring. In future years,
scientists will be encouraged to participate in
evaluation and research monitoring activities
as well.
1.7.7 Coordination With Forest Health Moni-
toring Activities in Other Countries
Other countries, including Canada (the
Acid Rain National Early Warning System) and
many European countries, have ongoing forest
health monitoring activities. Coordination with
these countries would enhance the compara-
bility of data and encourage the sharing of
lessons learned during the implementation of
forest health monitoring programs. Other
cooperative efforts could be initiated. These
include sample exchange programs,
development of reference materials, and
international workshops on forest health
monitoring.
This year, EMAP-Forests staff will
participate in forest health monitoring work-
shops in Canada and Europe. EMAP-Forests
staff will also participate in the 4th Inter-
national Quality Assurance Workshop to
address terrestrial OA issues. It is hoped that
international experts will participate as peer
reviewers of EMAP-Forests plans.
1-15
-------
2. APPROACH AND RATIONALE
Section 2 provides an overview of the
EMAP approach with an emphasis on the
conceptual and technical components as
applied to EMAP-Forests. The rest of the
sections fill in the details of the strategy plan.
2.1 RESEARCH DESIGN
2.1.1 Overview of the EMAP-Forests
Research Approach
An understanding of EMAP'S four-
tiered approach is important to the
presentation of the EMAP-Forests strategy for
meeting the program objectives. In the overall
EMAP approach, a tier is a level and type of
activity related to monitoring and assessing
ecological condition. Figure 2.1 outlines the
four-tiered approach to monitoring activities
(Anonymous 1990). Although most of this
document discusses activities at the Tier 1 and
Tier 2 levels, Tiers 3 and 4 are also important
components of long-term monitoring.
Tier 1, the broadest level, focuses on
landscape characterization, the estimation of
a resource's extent and geographical
distribution. This may include the pattern of
use on the landscape. Techniques such as
remote-sensing and the geographic
information system (GIS) are important at this
level. For EMAP-Forests, characterization of
the forest resource includes determination of
the extent and type of forests on a regional
basis.
Tier 2 activities are intended to allow
estimation of status and trends in resource
condition. A suite of chemical, physical, and
biological measurements are obtained from a
subset of Tier 1 sites and this information is
aggregated to make statements about the
status and trends of the resource on a
regional basis. Section 3 presents a more
detailed discussion about the suite of
measurements to be used in EMAP-Forests.
Tier 1 and 2 activities will allow the
successful completion of the primary EMAP
objectives and will be the priority in early
funding. However, additional information
needs will require the activities of Tiers 3 and
4. These complementary levels will provide
information about status, trends, and
diagnostics for more specific subpopulations
of interest and provide the link to ecological
research.
Tier 3 has two primary functions.
First, it is to determine whether or not the
perceived conditions warrant additional
evaluation, and then it is to provide the
diagnosis of a problem and the basis for
deciding what actions should be taken. This
activity may include intensifying the sampling
grid in a specific area, allowing more
concentrated data collection to evaluate a
specific problem.
Tier 4 is essentially the research
component of EMAP which supports the
monitoring activities of Tiers 2 and 3. The first
phase of Tier 4 focuses on the generation of
an ecological conceptual framework which will
be used to implement, test, and refine
monitoring indicators from Tiers 2 and 3. Tier
4's second phase includes the use of
ecosystem process studies to evaluate
experimentally whether or not the monitoring
indicators are adequately reflecting the actual
condition of the ecosystem.
Because EMAP-Forests is currently
working with the USDA Forest Service (FS) to
develop a multi-agency forest health
monitoring (FHM) program, it is important to
outline briefly the FS's approach to forest
health monitoring.
2-1
-------
Figure 2.1. EPA EMAP tier outline.
Tier
i
Landscape
Characterization
i
i
Tier 2
Tier 3
Trends in
Landscape
Change
Diagnostic
Analysis
i
i
Tier 4
Research
Process
Questions
Research
Monitoring Activities
2.1.2 Overview of the Forest Service
Research Approach
Over the past five years, the FS has
been developing a program for forest health
monitoring (FS-FHM) (USDA Forest Service
1989). The FS-FHM program has a three-tiered
structure, with each successive tier
representing an increased level of detail in the
monitoring program.
Detection or "routine" monitoring is the
most extensive tier and will be based on a
network of permanent plots including the
forest inventory and analysis (F1A) plots and
other permanent plots maintained by the FS.
Additional permanent plots will be used to
ensure representation of all forest lands. A
subset of "sentinel plots" will be chosen from
the geographically-based network of
permanent plots and will be used to collect a
greater amount of information than from
regular FIA plots. Information from routine
pest surveys distributed across U.S. forests
will be collected by the FS Forest Pest
Management (FPM) program and by state
agencies. These and other specifically-
focused monitoring activities will be linked
with the sentinel plot network. It is this
2-2
-------
monitoring level that will be the primary linkage
to EMAP-Forests.
The second tier is evaluation or "ad
hoc" monitoring. This level is initiated in
response to detection monitoring results.
When an area or problem of concern is
identified by detection results, specific
evaluation needs will be determined and
activities such as additional surveys, site- or
area-specific evaluations and more detailed
monitoring will be initiated.
Research or "investigative" monitoring
is the third tier. Sites representing key forest
ecosystems throughout the U.S. where both
special and ongoing long-term studies are
conducted will provide detailed information on
all components of the forest ecosystem. This
level of monitoring provides data to better
understand causal relationships and predict
rates of change in forest condition. Examples
of research monitoring are the Coweeta and
Hubbard Brook research sites. Participating in
research monitoring are FS Experiment
Stations and universities.
Fortunately, it is evident that the EMAP
tier structure is very similar to the FS tier
structure. This coincidence will encourage the
development of a multi-agency program. A
proposed tier structure for coordination by
both agencies is presented in Table 2.1.
2.1.3 Overview of EMAP Design
To meet the overall EMAP objectives,
EMAP-Statistics and Design has developed the
following design criteria:
• Estimate, with known uncertainty, the
health and status of any regionally defined
resource.
• Describe baseline data allowing rigorous
description of trends in health and status
of regionally defined resources.
• Identify associations among charac-
teristics both within and among resources
to establish possible causes of changed
condition.
• Quickly respond to new issues and
questions.
Important requirements and features of the
design include:
• Explicit definition of target populations and
their sampling units.
• Explicit definition of a frame for listing or
otherwise representing all the potential
sampling units within each target
population.
• Use of probability samples on well-defined
sampling frames to estimate popu-
lation characteristics rigorously through
randomization and use of probability
methods for sample unit selection.
• Flexibility to accommodate a variety of
resource types and a variety of problems,
some of which have not yet been specified.
• A hierarchical structure permitting
sampling at a coarser or finer level of
resolution than the general grid density,
giving flexibility at global, national, regional
or local scales.
• An ability to focus on subpopulations of
potentially greater interest (e.g. specific
types of trees rather than all trees).
» An ability to quantify statistical uncertainty
and sources of statistical variability for
populations and subpopulations of
interest.
The EMAP-Forests proposed design
strategy (see sections 4 and 5) is based
upon the overall EMAP design. A permanent
2-3
-------
Table 2.1. A comparison of characteristics of the phases of monitoring (Adapted from Figure 3 in Riitters et al. 1988).
Monitoring phase
EPA-EMAP Name ->
FS-FHM Name ->
Characteristic
Persistence
Frequency of
protocol
changes
Tier 1 & 2
Detection
long-term
seldom
Tier 3
Evaluation
short-term
frequent
Tier 4
Research
short- or
long-term
often
Spatial
coverage
Represents
Frequency of
analyses
Focus of
spatial
analyses
Focus of
temporal
analyses
Cause-
effect
inferences
Number of
parameters
Specificity
Auxiliary
data needed
extensive
intensive
forests
continuous
extrapolation
over regions
historical
and curtent
correlation
only
few
integrative,
non-specific
historical
trends
extensive or
intensive
forest issues
as needed
interpolation
within regions
current
possible
a few more
diagnostic
current
extensive or
intensive
mechanisms
as needed
both
current and
future
required
many
highly
specific
future
trends
national sampling framework has been
proposed that consists of a hexagonal plate
containing a triangular grid of approximately
12,600 points placed randomly across the
coterminous United States (see Figure 2.2),
Alaska, and Hawaii. A 40 km2 hexagon around
each point may be characterized at the
landscape level using remote-sensing and
geographic information service (GIS)
capabilities, aerial photography, and existing
landscape information data bases (see Figure
2.3). These data could be updated at
approximately 10-year periods.
The 40 km2 hexagons describe an area
that is one-sixteenth of the area of the United
States. They provide the basis for regional
landscape characterization estimates and the
2-4
-------
Figure 2.2. The baseline V* is9ab0U* 27 ki,ometBrS'
in the conterminous United btates. oh a
2-5
-------
Figure 2.3. The landscape characterization hexagons are 1/16th of the total area and centered on
the sampling points. The randomly positioned sampling grid occupies a common but
randomly selected position in each of the base tessellation hexagons.
changes in resource characterization over
time. A Tier 1 sample consists of the resource
units for any explicitly defined subpopulation
contained within the 40 km2 hexagons. Using
Tier 1 data, classification and further
subpopulation development can be done. Any
ecological resource in the landscape is
sampled according to strict protocols in
proportion to its abundance and frequency of
occurrence so that the resource sample
reflects the true characteristics of the
resource.
The triangular nature of the grid points
also allows the increase or decrease of the
grid density, according to the sampling
requirements of specific resources. Figure 2.4
illustrates a three-, four- and seven-fold
increase in grid density. Note that the
baseline grid (large dots) can be distinguished
within the denser grids. The flexibility of this
design will allow sample selection at a
resolution useable by individual states, if
desired.
Another important aspect of the overall
EMAP approach is the temporal and spatial
interpenetrating design of site characterization
and field sampling. Although the sampling
grid consists of 12,6000 points distributed
across the coterminous United States, only
one-fourth of these points will be considered
each year; a subset of these will actually be
sampled. A four-year cycle will be followed
during which all 12,600 grid points will be
considered. At the beginning of the second
four-year cycle, the points from the first year
of sampling will be revisited. Sections 4 and
5 present more detailed discussions about the
EMAP-Forests and FHM designs.
2.2 EMAP CONCEPTUAL FRAMEWORK
2.2.1 The EMAP Assessment Framework
within EPA's Office of Research and
Development (ORD) is one of six elements of
the Ecological Risk Assessment Program
(Figure 2.5). EMAP focuses attention on
important issues of environmental regulation
2-6
-------
7 tolp
Figure 2.4. Enhancement factors for increasing the base grid density. Enhancement will be made
only in the sample grid.
Ecological
Risk
Management
Ecological
Risk
Assasimtnt
Figure 2.5. EMAP provides a foundation for the ORD's Ecological Risk Assessment Program.
Principal interactions of EMAP with other elements are shaded. (From EPA 1990)
Risk. *^c
Communicfttion
kvv'rBioni
ftaatoretion ^
4 Management
Risk Characterization •« s
Ecological Efteeta
Ecological Eapoaure
Environmental Monllortng & Aateitmeni Program
2-7
-------
and management by characterizing ecological
risk and communicating this information to the
Agency Administrator. It also tracks the
responses of ecosystems to actions taken to
mitigate ecological effects or to reduce risk.
To achieve these goals, EMAP uses a flexible,
multi-tiered, and regional monitoring design
that emphasizes the assessment of indicators
of large-scale and long-term ecological effects
in relation to indicators of anthropogenic and
natural stresses.
In the EMAP risk assessment
framework, three broad categories of
indicators (response, exposure-habitat, and
stressor; see Section 3) are related as shown
by the example in Figure 2.6. EMAP relies on
response indicators to describe ecological
condition. Statistical associations among the
values of response indicators and those of
exposure, habitat, and stressor indicators,
coupled with knowledge about plausible
processes and effects mechanisms, will be
used to identify possible reasons for poor or
changing ecological condition (Messer 1990).
The assessment strategy (see Section
7) and the indicator development strategy (see
Section 3) use indicators to link measurements
to environmental values. A top-down
approach starts by defining the environmental
values of concern to society. These values are
then represented by assessment endpoints
which are "formal expressions of the actual
environmental value that is to be protected"
(Suter 1990). In EMAP, assessment endpoints
are defined, for example, as "proportions of
sites subnominal with respect to particular
response indicators within a region" (Messer
1990). Definition of these assessment end-
points leads to identification of the needed
indicators and thus measurements.
For statistical assessments, EMAP
utilizes a statistically based sampling design
(see Section 4) that provides unbiased
estimates of indicators, with known con-
fidence limits for well defined regional
populations. The extent of populations, the
proportions of sites which have specified
response indicator values, the associations
among indicators, and the trends in extent,
proportions, and associations can be
estimated by this design (see Section 6).
Some trade-offs between ecological
and policy relevance are required 1or
interpretive assessments. Keyed to indicators
and endpoints, the EMAP approach is
somewhere between the extremes of
intensively monitoring the environment (to fully
e^lain forest condition) and intensively
monitoring the society (to fully assess social
perceptions of condition).
Indicators are made to represent key
processes and to relate to key values and
perceptions about the forest. This strategy
does not typically permit causal inferences to
be made at the process level and does not
necessarily address all of the processes that
interact to determine forest status and trends.
For interpretive assessments, moni-
toring data augmented by knowledge of
plausible causes and mechanisms permits
analysis of correlations or coincidences of
indicator values in space and time (Messer
1990). These analyses can satisfy two of the
four criteria suggested by the NRC (1989) for
inferring causality in forest environmental
assessments. That is, a particular cause-
effect hypothesis can be addressed by
measures of correlation (consistency and
strength) and by measures of temporality (of
cause and effect). EMAP cannot address the
NRC criteria of responsiveness, or mechanism
of effect. The weight of evidence obtained by
monitoring data may be strong enough to
implicate or clarify certain causal hypotheses,
but additional data will usually be required to
fully test those hypotheses.
2-8
-------
ON-FRAME INDICATORS
Exposure-Habitat
Indicators (E)
SPATIAL
ASSOCIATIONS
Tli«u* Concentrations
TEMPORAL
ASSOCIATIONS
Ambient Concentration*
Water, Air, Soil, Sediment
Exotic*/Genetically
Engineered Organisms
Habitat
Landscape Pattern
STRESSOR INDICATORS (S)
Ecosystem Process
Rates & Storage
Community Structure
Populstlon Structure
Cross Pathology
Predator-Prey Relations
Successions! Stage
Cllmstlc Fluctuations
Pest-Disease Relations
Natural Process Indicators
Atmosherlc Deposition/Emissions
Demographics
Discharge Estimates
Fertilizer & Pesticide Applications
Land Use
Permits
Pollutant Loadings
Hazard Indicators
Hydrologlc Modification
Landscape Pattern
Pest Control
Dredging/Filling
Fire Management
Harvest Rate
Management Indicators
(OFF-FRAME DATA)
Figure 2.6. The EMAP risk assessment framework.
-------
Some limitations arise as a result of
imperfect scientific understanding of forests
and of regional interactions between forest
condition and environmental stresses. Forests
exhibit complex behavior because they are
subject to a variety of ubiquitous and variable
stresses from natural and anthropogenic
sources (Smith 1981,1984). The complexity is
manifested in, for example, interactions among
stressors, development of diversity, biological
mimicking of symptoms (Treshow 1984), and
compensation for stress-induced effects.
Changes in forests may be slow, rare, and
subtle (Strayer et al. 1986). These difficulties
and a traditional scientific focus upon fine-
scale processes have prevented any
meaningful regional definition of normal spatial
and temporal patterns and trends.
Nevertheless, "...knowledge of the structure
and physiology of forests and trees is now
sufficient to develop a basis for detecting
disruption or disturbance from a variety of
causes" (NRC 1989).
EMAP is designed to provide
information to those concerned with regional
and national environmental quality (Messer
1990). These individuals do not base their
decisions on the actions of individual polluters
or resource managers, but, rather, strive to
target environmental protection efforts in the
most effective way to ensure overall regional
or national environmental quality. Because
geographic and time scales are related (O'Neill
et al. 1986), EMAP is designed to focus on
long-term (i.e., decades to centuries) regional
phenomena.
The focus of analyses on assessment
endpoints means that the apparent condition
of each sampled site has to be classified as
"good" or "poor" in terms of the observed
values of the response indicators. Whether
developed from policy or scientific
considerations, the classifications are one
interpretation and will always be partly
subjective. In EMAP, the inevitable subjectivity
of interpretations is approached by
emphasizing agreement on the data and
indicators, and by presenting monitoring
results in such a way that alternate
interpretations can be made.
Any regional sample is expected to
reflect a wide range of condition at a given
time; even when conditions are "normal"
everywhere, a certain proportion of sites may
be in apparently "poor" condition because of
normal stresses. In this circumstance,
changes in the regional population
distributions of indicators would be
appropriate measures of change. Yet another
issue is that not all indicators will necessarily
give the same signal of condition at the same
site if they gauge different environmental
values or aspects of condition. Normalization
of indicator values, covariance-type analyses,
and analyses of aggregates of indicators are
three possible ways to overcome these
difficulties (see Sections 3 and 7).
The monitoring strategy is multi-tiered
in the sense that it includes both broad-scale,
integrative approaches and finer-scale, specific
approaches (Table 2.1). As noted previously,
EMAP's Tiers 1 and 2 provide for routine
regional-scale monitoring. Tier 1 is concerned
mainly with resource characterization and
estimation of indicators of stresses and con-
dition over landscape-scale (e.g., 40 sq. km.)
sampling units. Tier 2 emphasizes ground-
based measurement of indicators on much
smaller (e.g., 1 ha.) statistically representative
sites. Tier 3 is designed to permit more
intensive sampling of more specific indicators
in response to conditions observed in Tiers 1
and 2. Tier 4 refers mainly to very intensive
site-specific monitoring of a few locations to
answer research-oriented questions about
forest conditions, and how to monitor and
interpret them.
2-10
-------
Tiers 3 and 4 are the primary
opportunities to bring existing long-term
ecological monitoring sites into the picture.
These sites can be used for reference sites for
continuous monitoring of indicators, or as a
framework for doing retrospective analyses of
certain measurements and indicators.
Incorporating existing sites into the
probabilistic Tier 1 and 2 sample designs is
not simple but may be possible (e.g., Overton
1990). A critical issue in how to associate an
existing site with the probabalistic sample is
to determine what portion of the sample frame
that the existing site represents. It is also
important that many improvements to EMAP's
initial design will come from researchers that
study ecological processes at these existing
sites.
Cross-tier linkages are an important
element of the EMAP monitoring design.
Successful linkage requires conceptual
connections across spatial and temporal
scales, and between integrative and specific
models of forest processes. The issues of
linkage can be best resolved by considering all
tiers when planning any one tier, that is, by
explicitly recognizing a hierarchical framework.
A practical concern is the decision criteria to
initiate Tier 3 and Tier 4. Some criteria may be
termed "external", for example specific
legislationor meritorious scientifichypotheses.
"Internal" or data-based criteria are needed to
document the ptanned responses to
information generated by Tier 1 and Tier 2
monitoring.
Another important linkage is between
the ecosystem types recognized by EMAP
(e.g., forests, wetlands, and arid lands).
Resource Groups corresponding to ecosystem
types were set up originally to facilitate
cooperation with other Federal Agencies (e.g.,
USD A FS. USDA ARS, USDIBLM) that address
these different resource types. Common
issues were then addressed by EMAP-wide
groups (e.g., network design, quality
assurance, information management, and
logistics). An overall Integration and
Assessment group was set up to facilitate
cross-ecosystem communication and
coordination, and to prepare multiple-
ecosystem assessments. In an example of
this approach, the forest group is set up as a
module (Figure 2.7) that is linked to other
resource modules thorugh Tier 1 monitoring,
information management, and assessment
activities (Figure Z8). Every module is
supported by EMAP-wide projects in quality
assurance, remote sensing, logistics,
geographic information systems, and
statistics (Figure 2.9).
Other approaches to integration are
certainly possible. In response to early
reviews, EMAP has given increasing attention
to the possibility of achieving closer
integration and more holistic assessments by
erasing some of the definitional or program-
matic distinctions among ecosystems. In
this approach, ecologically-relevant landscape-
level sample units defined at Tier 1 are
sampled in a more coordinated fashion by the
different ecosystem groups at Tier 2. This
requires a more expansive view of ecological
condition by each ecosystem group and pos-
sibly modification of the statistical designs
described in this strategy. To help test this
possibility, EMAP-Forests has been developing
certain indicators in cooperation with EMAP-
Arid Lands and EMAP-Agroecosystems
resource groups and is exploring assessment
approaches through the EMAP-wide
Integration and Assessment Group. This
approach is programmatically more difficult
because it requires coordination among a
number of Agencies that cooperate with
EMAP.
2-11
-------
ADVISORY
COUNCIL
TDEPA
TD USFS
QUAUTYASSUR.
STATISTICS
AND
DESIGN
IMPLEMENTATION
TIER 2
on ground sampling
LOGISTICS
' TER1 >
Landscape
Chujcieniiiion
D*TECRaT!ON AND
ASSESSMENT
INFORMATION
MANAGEMENT
Figure 2.7. Structure of EMAP-Forests resource group as a module.
2-12
-------
Forest lands
TIER 1
Landscape
Characterization
CQ
m
j*
m
-i
INTEGRATION &
ASSESSMENT
CH
o
a.
INFORMATION
MANAGEMENT
Figure 2.8. Resource groups - hub relationship.
2-13
-------
£
it
EMAP
Conceptual Organization Showing
Terrestrial Resource Groups in Context
Director/Deputy
EMAP Manager
Steering Committee
Associate Director*
Technical
Directors
Assuta^^-
— Logistics
Sensing
sensing a^_^-
and DesjQ^
1. Group 1
b. Forest Lands
b. Arid Lands
c. Agricultural Lands
d. Landscape Characterization
e. Information Management
Remote Sensing and GIS
2. Group 2
I. Surface Watera
g. Wetlands
Statistics and Design
3. Group 3
h. Near Coastal
I. Great Lakes
4. Group 4
|. Integration & Assessment
k. Air & Deposition
5. Group 5
I. Headquarters Administration
Logistics
Figure 2.9. EMAP conceptual organization showing terrestrial resource groups in context.
-------
2.2.2 Towards an EMAP-Forests
Assessment Paradigm
The wind was flapping a temple flag,
and two monks were having an
argument about it. One said the flag
was moving, the other that the wind
was moving; and they could come to
no agreement on the matter. They
argued back and forth. Eno the
Patriarch said, "It is not that the wind
is moving; it is not that the flag is
moving; it is that your honorable minds
are moving."
- Platform Sutra
An assessment paradigm is a point of
view for organizing, synthesizing, and
interpreting data. EMAP cannot do everything
for everyone, but an assessment paradigm
implies that It will do something for someone.
A paradigm is a necessary point of departure
for the discussion of the more specific goals
and objectives. Recognition of a particular
viewpoint helps to see how EMAP-Forests
relates to other viewpoints and therefore other
monitoring efforts; this helps to identify a
unique paradigm as a raison d'etre.
Ecclesiastes' statement that "there is nothing
new under the sun" more or less applies to
environmental measurements; the uniqueness
of assessments may be more a matter of
interpreting measurements in unique ways.
The National Research Council
Committee on Forestry Research (NRC 1990b)
advocated an "environmentalism" paradigm
that complements traditional paradigms (e.g.,
utilization, conservation, and preservation) to
guide forestry research into the 1990's.
According to the NRC report, the environmental
paradigm (Leopold 1949) is a global version of
Leopold's classic land ethic which holds that
"... a thing is right when it tends to preserve
the integrity, stability, and beauty of the biotic
community..." Sustainability, maintenance of
diversity, and aesthetics are examples of
concerns that are valued in the environmental
paradigm (NRC 1990b). Humans are viewed
as participants in the environment rather than
as independent observers, users, preservers,
or conservers of the environment.
Adoption of the environmentalism
paradigm may not be a "safe" strategy for
EMAP-Forests assessments. To put it bluntly,
the concepts of environmentalism are
imprecise and the scientific merit is
challenged. Even though public opinion polls
show that environmental values are very
important to society, there is neither social nor
scientific consensus about what the paradigm
means. The danger is that an absence of
consensus can lead to dissipation rather than
focusing of assessment efforts. On the other
hand, embracing a more familiar paradigm
(e.g., utilization) may seem easier, but risks
being incomplete and redundant.
While Leopold's world-view is often
cited by ecological scientists, it is not at all
clear how to best use the paradigm. The
current situation can be compared to airplane
design in the 1920's: any interpretation of the
basic idea is all right so long as the thing flies.
If evolution of airplane design is an analogy,
the most useful and accepted versions of a
"scientific" paradigm may eventually emerge.
The futurist Alvin Toffler (Prigogine and
Stengers 1984) noted that "...science is an
open system embedded in society and linked
to it by very dense feedback loops...its
development is shaped by cultural receptivity
to its dominant ideas," and a U.S. Federal
Court concluded essentially that "science is
what scientists do". These statements
suggest that even though environmentalism is
scientifically ill-defined at present, it is still a
valid world-view within which scientists may
practice their professions.
In any science, axioms and rules set
the stage for theories and hypotheses, and
permit consistency of inferences. The
2-15
-------
observation that the environmentalism
paradigm lacks objectivity is not fatal; no
science can claim true objectivity except in the
context of the accepted axioms and rules
(Penrose 1989). So the real issue is that
scientists have not agreed upon the axioms
and rules of environmental science. Until
consensus is reached, the paradigm will best
be used as a general guide to help define
environmental values of concern. As a
practical matter, the specific questions of
monitoring design and data synthesis will
require recourse to more familiar sciences.
2.3 FORESTS AS AN EVOLVING PROGRAM
WITHIN EMAP
The role of forest monitoring in EMAP
is an evolving one. As EMAP-Forests moves
toward the multi-agency FHM program, it is
vital that the goals and objectives of all
participating agencies are considered.
Currently, EMAP-Forests and FS personnel are
striving to accommodate the objectives of both
agencies. This process will continue to be
vital as additional federal and state agencies
and universities begin to participate in the
FHM program.
A major aspect of the EMAP-Forests
approach to long-term monitoring is
maintaining options. One example is the
design strategy that utilizes a uniform grid,
allows for unbiased samples, allows
enhancement of the grid to meet special
sampling needs, and allows post-stratification
of data. Another example of maintaining
options is the indicator approach (see Section
3). The focus is on a suite of indicators which
can be modified over time. There is an
emphasis on establishing an effective
indicator development process that will allow
indicator refinement. This process is built
around assessment endpoints, encouraging
long-term indicator evolution. Similarly, the
EMAP-Forests approach to assessment is an
effort toward maintaining options for the
future (see Section 7). This approach is
designed to allow a range of data assessment
possibilities from annual statistical summaries
with a rapid turn-around time to more in-depth
studies resulting in interpretive reports.
A second major aspect of the EMAP-
Forests monitoring approach is the effort to
base the multi-agency FHM program on the
strengths of each participating agency. For
example, both the EPA and the FS bring
particular strengths to the FHM program. The
EPA has a national focus with emphasis in
quality assurance (OA), air and deposition
data, climatic programs, and a national pool of
scientific expertise. The FS has a regional
focus with experience in large field programs,
working with state agencies, and a strong
background in forest science research.
As the multi-agency FHM program
evolves, it will be important to continue to
highlight and utilize each agency's areas of
expertise.
2.4 FIVE-YEAR STRATEGY
2.4.1 Multi-Agency Cooperation
At present, the FS and the EPA are the
major agencies participating in the FHM
program. So far, the program has remained in
the east. Since there has not been significant
western representation at planning meeting
and workshops, there may be biases in
design, logistics and indicators towards the
east. In 1991 an effort has been made to
incorporate western representation into FHM.
This includes the state FS, and the National
Forest System as well as FPM and FIA.
Meetings on the concepts of FHM were held in
December 1990 and January 1991 for western
FHM participants. There is a possibility that
the states of California and Colorado will
participate in some indicator testing on a
small set of plots in 1991. In 1992 California
and Colorado will be fully implemented.
2-16
-------
As FHM expands to all states, other
agencies will be included in FHM. The SCS
already is involved with field sampling of soils.
Other agency involvement would include the
National Park Service, the Bureau of Land
Management, and the Fish and Wildlife Service
to name a few. These groups will have input
into the FHM program through formal
processes, as additions to the Advisory
Council (Figure 2.7) and interagency
agreements and informal processes as peer
reviewers and by personal communications.
By 1995, when full implementation is
anticipated, the agencies that play major roles
in the FHM program will be recognized. These
agencies will form the nucleus for the
development of the program towards the
future.
2.4.2 Implementation
The FHM implementation schedule is
currently being planned by the FS (see Table
2.2). EMAP-Forests will follow this schedule.
However, the FS may not implement the full
suite of indicators in a region if some
indicators have not had previous use in the
region and there is a question of whether or
not they will provide comparable results to
other regions. EMAP-Forests will provide
pilots and demonstrations (see section 3)
before full implementation of these indicators
and will follow the implementation schedule as
depicted in Table 2.2.
2-17
-------
Table 2.2. Forest Service-FHM implementation schedule as currently planned and the predicted nunber of FHM plots
EMAP-F Sanple Plots Worksheet
1990
1991
1992
1993
State
For Area
State
For Area
State
For Area
State
For Area
CT
1,826
CT
1,826
CT
1,826
CT
1,826
HE
17,607
ME
17,607
ME
17,607
HE
17,607
HA
3,225
MA
3,225
HA
3,225
HA
3,225
NH
4,987
NH
4,987
NH
4,987
NH
4,987
Rl
405
Rl
405
HI
405
RI
405
VT
4,544
VT
4,544
VT
4,544
VT
4,544
Total
32,594
NJ
2,007
N
-------
3 STRATEGY FOR INDICATOR DEVELOP-
MENT AND IMPLEMENTATION
The EMAP program seeks to: (1)
describe current ecosystem status, (2) identify
long-term changes in ecosystem status, (3)
characterize the components of ecosystem
change, and (4) suggest avenues for
diagnostic research. To complete these
objectives, the program has adopted an
indicator-based approach to the assessment
of ecosystem condition (Knapp et al. 1990).
This approach assumes that: (1) indicators of
specific interrelationships between ecosystem
functions (e.g., rates of nutrient transfer,
capacity for nutrient conservation, level of
redundancy of function, etc.) are known, (2)
indicators can be related within an
assessment framework to specific changes in
ecosystem condition (e.g., growth, morbidity,
mortality), and (3) indicator measurement at a
national survey scale is logistically,
economically and technically feasible. When
the above criteria for indicators are not met,
procedures have been established to evaluate
options for the development of new indicators,
to assess potential utility of these indicators
within the existing assessment framework,
and to evaluate the need to develop new or
additional assessment frameworks. This
section describes the strategy and procedures
that EMAP-Forests will use to complete this
component of the program.
3.1 DEVELOPING A CONCEPTUAL
FRAMEWORK FOR ASSESSMENT
The forest health monitoring (FHM)
program will assess the effects of multiple
stresses on forest ecosystem condition.
Ecosystem processes are linked to spatial and
temporal combinations of environmental
components (climate, soils, topography,
vegetation, trophic structure, etc.). Therefore,
the success of an indicator and of the
corresponding modeling and assessment pro-
gram will depend on the development of an
appropriate diagnostic framework for
identifying major resources of concern,
suggesting research priorities, and defining
attainable conditions of sustainable
ecosystem health. This framework should be
developed around a regional concept,
recognizing that the nature of problems and
solutions vary among definable, ecological
regions. The framework should focus on the
development and application of a suite of
tested indicators and models that accurately
predict risk to specific ecosystem sub-
populations. It should also provide guidelines
for specifying the most reliable models for
determining ecosystem risk for various stress-
management scenarios.
3.1.1 Scale of Ecosystem Response
Characterization
The FHM personnel will develop an
assessment framework within which
indicators may be used to characterize
ecosystem condition and/or be aggregated
into some common index of condition (or
predictor of impending change). It is
necessary to adopt or develop models of
forested ecosystem structure and function
that embody the most current hypotheses
regarding the inter-relationships among
ecosystem functional components.
Ecosystem response to stress is a
function of the magnitude of perturbation
resulting from the stress (acuteness of
damage), the area exposed to the stress
(extent), and the capability of the ecosystem
to eliminate or ameliorate the damage
(resiliency). It is therefore impossible to
extricate the assessment of an areal response
from that of the component ecosystems'
responses. It is also impossible to develop an
assessment of areal effect without first
making assessments of the range of
ecosystem responses expected within the pre-
defined area and the proportion of the
resource in question for which loss is
3-1
-------
politically and/or economically acceptable (i.e.,
relative value judgements).
Careful consideration should be given
to the scale at which assessments among
indicators are made. Application to a regional
estimation of forest condition requires
aggregation of fine scale results, usually
through some statistical estimation of the
regional distribution of the various
environmental components employed in the
conceptual model's representation of
processes. It is not obvious whether or not
the statistical aggregation of fine scale
process simulations on selected sites will be
superior to simple models applied directly at
the regional scale. Consequently it is
desirable that, if possible:
1. Comparisons are made at the spatial scale
at which the models will ultimately be used
to estimate forest condition.
2. Differences in input data should be
minimized so that differences in output
most strongly reflect the effects of model
structure.
3. Differences among model predictions due
to differences in model structure should be
distinguished from differences in
calibration data sets, aggregation error,
etc.
4. Opportunities for combining the
advantages of different modeling
approaches should be explored.
Appropriate scales for ecosystem
response characterization can be addressed
through a set of procedural steps common to
regional ecosystem assessments (Figure 3.1).
While the objectives of each step are the same
for any region or stress, the level of available
analytical data will likely vary for both regions
and pollutants. Several factors cause this
variability: disparate understandings of the key
processes in different ecosystems and the
state of development of data bases for
different regions. Decision criteria relating to
the elements inherent in each step, as
described below, should be considered.
1. Articulate and prioritize regional issues and
stresses and identify indicators of
condition. To incorporate consistent
decisions into regional analyses, a set of
issues, stresses, and assessment criteria
must first be defined and prioritized.
These define the goals of the data
analysis. For example, issues of concern
common to loblolly pine forests in the
southeastern United States might include
compiling information about the potential
for long-term production, the spatial and
temporal periodicities of insect damage,
and/or the most appropriate management
strategies for different locales within the
region. The types of environmental data
that are compiled, analyses and
interpretation techniques that are
performed, and graphic products that are
produced all hinge on these issues.
Assessment criteria selection sets
limitations on the parameters charac-
terizing ecosystem response, thus defining
thresholds of condition. Assessment cri-
teria development requires the selection of
"indicators" that are representative of the
criteria (see Sections 1.4 and 3.1.4). It may
also require the development of
information about the state of the
ecosystem in the absence of
anthropogenic stresses to bound
expectations of attainable ecosystem
quality (Section 3.1.4.4).
2. Data compilation. Available environmental
data that appear to cause or reflect the
issues of concern should be collected. For
a study of the effects of ozone on loblolly
pine forests, for example, this might
include compiling information about air
mass movement and stagnation, monthly
weather patterns and insolation, prevailing
wind directions and speed, landforms
3-2
-------
Prioritize Issues of Concern
Compile Resource Characterization Data
High Data Quality?
No
Recompile or Reclassify
Yea
Assess Coincidence or Common Patterns
in Resource Characterization Data
Characterize Variability Within
and Among Sub-Populations
j Delineate Sub-Popuiations
i !
Figure 3.1. Regional assessment.
3-3
-------
(elevation, slope, and aspect), soils (type,
moisture holding capacity, and nutrient
availability), distribution and background
data on loblolly pine forests (management
practices and tree growth), and sources of
atmospheric pollution (volatile organic
compounds and nitrous oxides). Digital and
non-digital maps, points, and tabular data
can also be included.
3. Data quality assessment. A percentage of
each point or tabular data set should be
examined for concurrence between sample
site locations and descriptions, for the
number and types of methods from which
the data are derived, and for assurance
that data values are scientifically plausible.
Maps should be examined for documen-
tation or indication of the resolution and
reliability of the data used to generate the
maps.
4. Synthesis of new data layers. Data sets
can be reclassified or combined to
generate new data sets. Numeric data can
be sorted into classes to detect spatial
geographic patterns in the data that might
relate to other environmental charac-
teristics. Several data sets can be
combined to create classes that represent
composite characteristics.
5. Regionalization. Regionalization (or
delineation of response sub-populations) is
important because stress exposure can
only be altered on a regional, as opposed
to an individual ecosystem scale. It is also
important because significant sub-
populational responses (e.g., high elevation
spruce fir) are more likely to be detectable
in shorter time frames than are those of
the greater population.
6. Resource characterization. In addition to
mapping or calculating the extent of the
response sub-populations, it is important
to document the range in conditions within
sub-population types and, if possible, any
trends of change.
7. Stressor characterization. Characterization
of exposure to stresses at multiple scales
(temporal and spatial) is important for
correlation and pattern analysis based on
historical and current ecosystem status,
model calibration, and estimations of
ecosystem condition under alternative
exposure scenarios. This information may
come from multiple sources such as
atmospheric models, site-specific wet and
dry deposition measurements, interpolat ion
techniques, or from combinations of these
approaches.
8. Selection of model and response
characterization. Some models have been
used to predict how changes in exposure
are reflected in ecosystem status change.
Many model types (e.g., statistical or
process) may exist to predict response to
a particular stress or combination of
stresses. Each model type has benefits
and disadvantages that limit or favor its
use in any particular application. The
application of multiple models to any
particular issue can provide increased
confidence in the predictions if the models
are properly validated and if the predictions
are convergent. If the predictions are
divergent, multiple models can provide a
basis for identifying additional research
needs.
9. Ecosystem response presentation. An
integrated approach to forest condition
assessment and the presentation of such
information must represent complex,
discontinuous, and spatially-distributed
factors regulating ecosystem responses to
imposed stresses. Maps should be
capable of displaying comparisons
between ecosystem response and stress
load, and of comparing population
characteristics within and among regions.
3.1.2 Defining Forest Health
A major use of indicators in EMAP will
be to assess condition or health of ecological
resources. Rapport (1989) lists three
approaches or criteria commonly used to
3-4
-------
assess ecosystem health: (1) identification of
systematic indicators of ecosystem functional
and structural integrity; (2) measurement of
ecological sustainability or resiliency (i.e., the
ability of the system to handle natural or
anthropogenic stress loadings; and (3) an
absence of detectable symptoms of
ecosystem disease or stress. "Thus, ecological
health is defined as both the occurrence of
certain attributes that are deemed to be
present in a healthy, sustainable resource, and
the absence of conditions that result from
known stresses or problems affecting the
resource.
For any ecosystem or aggregate of
ecosystems, many available options and
methodologies describe the "health" or
condition of the resource of interest. Section
3.2 presents the generic approach that the
EMAP program is proposing for the
development of ecosystem condition and
response indicators prior to their
implementation in the national monitoring
program. Discussions focus on the
identification of decision criteria rather than
the presentation of a suite of indicators of
forest condition.
3.1.3 Use of Assessment Endpoints
The FHM reports on the condition of
forested ecosystems will be based on
indicator(s) response(s). These responses
represent the quantifiable changes occurring in
some components of the forested ecosystem.
The balance of indicator response (net and
relative magnitudes of change in positive or
negative direction) however, reflect societal
values of forested ecosystems. The EMAP-
Forests assessment framework recognizes the
differing uses to which forests are placed.
Societal values can therefore be described as
fitting into one of three broad categories:
• Ecological Integrity - The concept of
ecological integrity recognizes the
importance of maintaining ecosystem
functional capacity and considers both
biological and abiological resources.
• Economic Value - The economic value
represents the capacity for the system to
generate both direct (e.g., sales) and
indirect (e.g., regulation of water
availability for agriculture) sources of
livelihood.
• Sociologic Value - This value incorporates
the intrinsic desires of society to maintain
some parts of the world in a "natural state"
and includes recreational and aesthetic
components.
To provide a structure bridging the gap
between societal concepts of value and the
measurement of quantifiable components of
the ecosystem, a number of assessment end-
points should be identified (see examples in
Table 3.1). Using such a structure, it is
possible (and likely) that any individual
indicator will be interpretable in the context of
several societal values. For example, soil
chemical analysis data will be used in
developing interpretations for the assessment
endpoints of soil productivity, soil weathering
rate, soil contamination, and nutrient cycling
balance.
An example of the relationships in the
assessment framework is presented in Figure
3.2. Reading the figure from right-to-left, the
societal value serves as the focus through
which all assessment endpoints can be
interpreted. The assessment endpoints
encompass broad categories of ecosystem
component characteristics. The aggregation
of these characteristics defines the status of
the ecosystem. The FHM program will provide
information on the condition of the
assessment endpoints (i.e., quantifiable
characteristics). However, policy offices are
responsible for making recommendations that
relate to societal values (e.g., whether mitiga-
3-5
-------
TABLE 3.1. Examples of relationships between societal values, assessment endpoints, and indicators.
Societal Value
Assessment Endpoint
Indicator
Ecological Integrity
Abiotic Resource
Biotic Resource
Soil credibility
Soil productivity
Soil weathering rate
Soil contamination
Soil water retention
Water quantity
Water quality
Air quality
Biodiversity
Soil surface recession
Gully formation/density
Stream sediment load
Chemical analysis
Nutrient ratios
Available nutrients
Microbial biomass
Textural analysis
Caption Exchange Capacity (CEC), Anion
Exchange Capacity (AEC)
Mineralogy
Chemical analysis
Moisture percent
Tension lysimeters
Piezometers
Climate history
Gaging weirs
Water chemistry (in cooperation with
MAP-Aquatics)
Stream physical structure (related to
habitat quality)
Air chemistry
Visibility
Particulate scavenging rate
Floral & faunal species
Landscape distribution
Habitat quality/contiguity
Populational mixing
(continued)
3-6
-------
TABLE 3.1. (continued)
Societal Value Assessment Endpoint Indicator
Biotic Resource
(continued)
Nutrient cycling balance
Soil chemistry
Foliar chemistry
Sec. carbohydrates
Index development
PAR
Contaminants
Foliar chemistry
Histopathology
Fecal analysis
Quality of animal resource
Genetic diversity (populational mixing)
Census taking
Food source quality
Fecundity
Quality of vegetative
resource
Foliar chemistry
Vertical structure
Species diversity
Regeneration rate
Growth rate
Net primary productivity
Damage surveys
Pathological surveys
Symbioses (floral & faunal)
Landscape characterization
Species range
Niche exploitation
Patch dynamics
Regeneration requirements
Economic Value
GNP from forest products
National/state economic reports
Biomass by FP category
Harvest inventories
Mensurational monitoring
Harvest category ratings
(continued)
3-7
-------
TABLE 3.1. (continued)
Societal Value
Assessment Endpoint
Indicator
Economic Value
(continued)
Water export
Habitat provision
Tourism and recreation
Gaging weirs
Agricultural irrigation rates
Municipal water consumption
Hydroelectric water flow rates
Fish hatchery counts
Stream physical structure (related to
habitat quality - joint with EMAP-
Aquatics)
National/state economic reports
USD! accounting reports
Sociologic Value
Designated use/usability
Pristine quality
User fees/counts
Animal counts/sex ratios
Responses to visitor surveys
Landscape characteristics
Management history
Land use history
Acres of wilderness
Road density
Age of forest (successional history)
tive action should be required). These recom-
mendations will not be contained in FHM
reporting. Indicators may be comprised of
individual field measurements or aggregations
of field measurements; they are the technical
base for quantifying the characteristics of the
assessment endpoints.
3.1.4 Application of Indicators
Indicators carry no capacity to assign
a value judgement. They serve as a "tag",
marking a point of condition in time and space
which can be applied to multiple perceptions
of value. Assessment endpoints depict the
distribution of indicators or may be statistical
representations of indicator distributions. A
specific example of the scheme relating values
and indicator measurements by interpretation
through assessment endpoints is presented in
the bottom half of Figure 3.2. In this case,
total soil and foliar carbon and nitrogen are
measurements made as part of the soil and
foliar chemistry indicators. Measurements of
wet and dry nitrogen deposition at off-frame
network sites can be modeled to provide
estimates of deposition to on-frame sites.
The integration of these components provides
information regarding the status of nutrient
transfers among these pools and thus the net
nutrient cycling balance of the soil-biotic
matrix. The addition of an available nitrogen
3-8
-------
INDICATOR
FIELD MEASUREMENT
M INDICATOR
FIELD MEASUREMENT
F IELD MEASUREMENT
ASSESSMENT
ENDPOINT
SOCIETAL UALUE
FOLIAR
CHEMISTRV
OTHERS
SOIL
CHEMISTBV
ATMOSPHERIC
DEPOSITION
(MODELED)
QUfU BIO RESOURCE
NUTRIENT CYCLING
BALANCE
AUAILABLE NITROGEN
TOTAL SOIL AND FOLIAR
CARBON AND NITROGEN
DRV AND WET NITROGEN
DEPOSITION
OTHERS
Figure 3.2 Forest health monitoring (FHM) assessment framework.
-------
measurement offers the opportunity to
aggregate these data into an index of potential
capacity for soil productivity. This index may
provide the capacity for estimating optimal
balances of these two nutrients in differing
forested ecosystems or regions. The
distribution of the nutrient cycling balances
attained by differing sub-populations within a
region could then be used to provide a
quantifiable assessment of regional system
resiliency (i.e., bounding levels describing the
known experience and condition). A shift of
the indicator outside this known distribution
may signal the onset of system dysfunction.
For a more detailed discussion of integration
and assessment, see Section 7.
3.1.4.1 Indicator Selection Criteria
Knapp et al. (1990) identified a number
of criteria for indicator selection. Table 3.2 is
copied from their document. This section
describes the criteria that must be applied to
the selection of indicators. Existing
constraints in meeting these criteria would
lead to the selection of indicators on an
interim basis while additional information is
collected that would enable the use of a more
desirable set of indicators.
• Societal Value - Changes in indicator
status should result in a willingness to
manage stress sources. Though policy-
makers can be advised of the significance
of an array of technically relevant
indicators, the willingness of society to
accept regulation on the basis of indicator
changes must also be considered. The
values that society places on forested
ecosystems can be aggregated into three
categories: ecological integrity, economic
value, and sociological value. These three
categories drive the EMAP-Forests
monitoring program. All indicators selected
for implementation must be interpretable in
an assessment context (see Sections 3.1.3
and 7 for further discussion) that has a
direct relationship to these values.
• Ecological Integrity - The ecological
integrity of a forested ecosystem is a
function of the quality of and interactions
between its component parts (i.e., abiotic
and biotic elements). There is a growing
awareness that the "health and quality" of
the human condition is inextricably linked
to the "health" of the ecosystems people
inhabit and the use to which ecosystems
are placed (e.g., waste disposal).
Humankind is learning that the term
"ecosystem" is a function of multiple
scales. For example, the source of
atmospherically deposited stresses to a
watershed may be thousands of square
kilometers, the affected vegetation in the
watershed may occur in only a few square
kilometers, and the area affected by the
watershed's export (larger streams and
groundwater) may again be thousands of
square kilometers in area (see Section
3.2.3 for a discussion of reporting strategy
incorporating broad spatial resources).
• Economic Value - The economic value of
forested ecosystems lies in marketing vast
quantities of forest products each year,
management of forests for tourism (e.g.,
national park system and private souvenir
vending), and many other services that are
currently treated as external to the goals of
forest management per se.
• Sociological Value - The sociological (or
aesthetic) value placed on an ecosystem is
an intangible quality stemming from a
sense of personal value found in nature.
• Conceptual Model Output - Because the
FHM assessments of forest condition will
be made using conceptual models as
hypotheses of forest structure, function,
and response, indicators included in the
3-10
-------
Table 3.2. Indicator selection criteria.
Critical Criteria
Must reflect changes in ecological condition, pollutant exposure, or habitat condition,
and respond to stressors across mo6t pertinent habitats within a regional resource
class.
Must be related unambiguously to an assessment endpoint or relevant exposure or
habitat variable that forms part of the ecological resource group's overall conceptual
model of ecological structure and function.
Can be quantified by 6ynoptic monitoring or by cost-effective automated monitoring1.
Exhibits low measurement error and stability (low temporal variation) during an index
period.
Must have sufficiently high signal strength (when compared to natural annual or
seasonal variation) to allow detection of ecologically significant changes within a
reasonable time frame.
Sampling must produce minimal environmental impact.
Regionally Responsive
Unambiguously Interpretable
Simple Quantification
Index Period Stability
High Signal-to Noise Ratio
Environmental Impact
Sampling Unit Stable
Available Method
Historical Record
Retrospective
Anticipatory
Cost Effective
New Information
Desirable Criteria
Measurements of an indicator taken at a sampling unit (site) should be stable over the
course of the Index period (to conduct associations).
Should have a generally accepted, standardized measurement method that can be
applied an a regional scale.
Has an historical data base or a historical data base can be generated from accessible
data sources.
Can be related to past conditions by way of retrospective analyses.
Provides an early warning of widespread changes in ecological condition or processes.
Has low incremental cost relative to Its information.
Provides new information: does not merely duplicate data already collected by
cooperating agencies.
1 Most important In selecting core indicators (phase 5).
monitoring plan must be specifically in-
cluded (or amenable to inclusion) in con-
ceptual models of forest condition and
response. See Section 3.1 for a more
complete discussion of the FHM strategy
for conceptual framework development.
• Specificity and Sensitivity - Indicators
adopted by the program must be sensitive
to changes in stress exposure and/or
reflective of the long-term changes in
forest structure. "They must be
operationally definable in terms of some
3-11
-------
measurement or combination of measure-
ments. See, for example, the discussions
of the soil productivity and foliar nutrients
indices in Section 3.1.4.3.
• Application - In addition to the selection of
an indicator, its form of expression must
also be considered. For example, an
indicator such as available N may be
caressed in the following ways: (1) as the
percentage of samples which fall below or
exceed some threshold value; (2) in terms
of changes in the median value; or (3) in
terms of the percentage of map units that
contain ecosystems below some threshold
value. The choice of an indicator and
reporting format reflects the desire of
decision makers as well as the ecological
relevance of the information and the
structure of available data bases.
3.1.4.2 Indicator Categories (from Knapp et
al. 1990)
A key element of the EMAP approach is
the linkage of indicators to assessment
endpoints. Potential indicators are identified
using conceptual models of ecosystems,
followed by systematic evaluation and testing
to ensure their linkages to the assessment
endpoints and their applicability within EMAP.
The models used may be based either on
current understanding of the effects of
stresses on ecosystems, or on the structural,
functional and recuperative features of
"healthy" ecosystems. Important information
about assessment endpoints falls into one of
the following categories: condition of the
ecosystem, exposure of the endpoint to
potential stressors, or availabilityof conditions
necessary to support the desired state of the
endpoint. To provide appropriate linkage
between assessment endpoints and
indicators, indicator development in EMAP will
produce Indicators that fall into one of four
categories (Hunsaker and Carpenter 1990):
1) Response indicators represent character-
istics of the environment measured to
provide evidence of the biological condition
of a resource at the organism, population,
community, or ecosystem levels of organ-
ization.
2) Exposure indicators provide evidence of the
occurrence or magnitude of contact of an
ecological resource with a physical,
chemical, or biological stress.
3) Habitat indicators are physical, chemical,
or biological attributes measured to
characterize conditions necessary to
support an organism, population,
community, or ecosystem (e.g., availability
of snags, substrate of stream bottom,
vegetation type, extent and spatial
pattern).
4) Stress indicators are natural processes,
environmental hazards, or management
actions that effect changes in exposure
and habitat (e.g., climate fluctuations,
pollutant releases, species introductions).
Information on stresses will often be
measured and monitored by non-EMAP
programs.
3.1.4.3 Indices as Aggregate Indicators
One of the objectives of FHM will be
the development of "Indices of Integrity" for
forest ecosystem populations and/or sub-
populations. Although these indices will not
be used as the summary reporting and
assessment framework, they will provide
additional information in the development of
interpretive reports (see Section 11) when used
as independent indicators of status and
change. Such indices are developed by
assigning weights (or importance values) to
individual variables representing (or
integrating) key functionalities of the
ecosystem. The weighing of any specific
variable (e.g., foliar C/N ratio) can vary
regionally or even by species, depending on
the "encountered range" of that variable in the
3-12
-------
baseline population (see Section 3.1.4.4). This
approach has been developed for fish
communities (e.g., Karr 1981; Karr et al. 1986),
has been applied in regional frameworks
(Miller et al. 1988), and has been used in the
development of rapid bioassessment protocols
for streams and rivers (Plafkin et al. 1989).
Figure 3.3 illustrates a generic example of how
an integrity index might be used in an assess-
ment endpoint context. The choice of an index
score that is representative of the subpopula-
tion (placement of vertical dashed line) is likely
to be regionally specific and can be addressed
in a "technical context1 by comparison to the
baseline condition. It can also be addressed
in the FHM interpretive summaries. However,
the "allowable" magnitude of curve shift in a
"sub-nominal direction" (i.e., proportion of the
regional population that society is content to
let become "less healthy") is a policy issue and
is not within the purview of the FHM
interpretive assessments.
(Number of Plots Within Sub-Population)
100
90 —
System-Dependent?
80 —
60 J_
40 —
30
Policy
50
20
40
30
Integrity Index Score
Figure 3.3. Application of integrity index.
3-13
-------
Peer workshops conducted during 1989
recommended the likely riear-term success in
developing soil productivity and foliar nutrients
indices. Pilot field activities were conducted
during the summer of 1990, and data analyses
are examining the feasibility and logistical
issues surrounding these indices. Develop-
ment of the soil productivity index has
advanced the most for several reasons. There
is a large volume of information available from
the agricultural and soil science research
arenas and from the Direct/Delayed Response
Program (Church et al. 1989). Discussion of
the developmental status of each index
follows.
So/7 Productivity Index
Many researchers have written
qualitative or quantitative descriptions of the
relationship between vegetative growth and
soil or foliar chemistry. An overview of soil
rating systems in the United States is
presented in an excellent review by Huddleston
(1984). Though much of the productivity rating
research has been done in an agricultural
setting, the concept has also been applied to
erosion studies (Bruce et al. 1988), and its
application to forestry and forest soils is
generating even greater interest.
Possible soil measurements include
economically and logistically feasible
parameters that are important to determine
and monitor soil productivity. Physical
parameters such as drainage class and
physiographic position have been used
repeatedly in growth response studies. Many
of these same parameters are being measured
as part of the FHM soil monitoring activities.
Though considerable research has
been devoted to the identification of soil
processes that are important in forest
vegetative response, the necessary
parameters have not been linked together in a
summary index or model that is suitable for
application on regional or national monitoring
scales. An intermediate evaluation technique
known as "collapsed classification" (Miah et al.
in preparation) is being developed, however,
and may serve as a precursor to eventual
summary indices. Based on a factorial design
utilizing confounded blocks (Cochran and Cox
1957), this intermediate framework enables
data users to evaluate status and trends
through a multivariate partitioning technique
that aggregates different parameter ranges of
concentration into appropriate response
groups. It is believed that key soil productivity
variables identified in the intermediate
framework could eventually be combined into
an index or subindex that identifies, on a
regional basis, the effects of degradation of
soil quality on vegetative response and other
general indicators of forest condition (Ott
1978).
Significant advances have been made
in the development of indexing systems. To
develop indices of forest condition, EMAP-
Forests personnel are attempting to acquire
data from forest resource scientists. Recent
research suggests that the Diagnosis and
Recommendation Integrated System (DRIS)
indexing system (Beaufils 1973) and Timmer
ratios might establish associations among
various soil and foliar measurements and
forest mensuration data. Both techniques
utilize various configurations of parameter
ratios that attempt to identify optimal ranges
of concentration for specific responses such
as vegetative growth. These approaches will
be evaluated for applicability within the FHM
assessment framework.
Parameters that are expected to serve
as inputs to soil indices are being tested. The
parameters are configured into "parameter
groups" that are defined by similar or
associated types of variables (e.g.,
macronutrients). Data for individual
parameters will be weighted according to the
relative importance (e.g., macro- versus micro-
3-14
-------
nutrient, root toxic, etc.) lor forest response of
each parameter within a parameter group.
Each parameter group will be subjected to a
subsequent weighing function (that may vary
across regions or forest cover types) as part
of the calculation of an index for a given
aggregation. An investigation into this facet of
indicator development has been initiated.
From a sampling design standpoint, an
index should be constructed to represent the
actual status of forest condition with respect
to the endogenous plot composition. An index
should allow trend assessment for a given
plot or region over a given time period. For a
specific index such as soil productivity, the
component parameters of interest should be
limited to those operative properties that are
known to influence productivity (e.g., soil
moisture status, clay content, organic carbon
content, surface horizon thickness, etc.). It is
unlikely that an index could be based strictly
on classification by soil genesis or other non-
operative taxonomic criteria.
Operative properties should be
amenable to aggregation or dispersion to a
level that Is necessary to define appropriate
configurations that can be used to interpret
the assessment endpoints of interest. In the
application of DRIS, for example, the index
initially could be derived by using equations to
determine appropriate ranges and confidence
intervals for the independent variables of the
ORIS norms. Later applications would capture
dependent variables from other FHM indicators
as they are implemented. Ultimately, the index
might be used to express overall relationships
with respect to the assessment endpoints or
the population characteristics of interest.
Because of this, the index must be able to
accommodate and account for differences in
parameters, methods, and procedures that are
used to measure forest status across all
regions of the United States.
Foliar Nutrients Index
Foliar nutrients and chemical
contaminants have been proposed as
'Exposure-Habitat' indicators in EMAP-Forests
(Hunsaker and Carpenter 1990). Nutrient
deficiencies often affect growth or produce
characteristic visual symptoms that are
indicators of disruption of normal physiological
function. The diagnosis of foliar nutritional
status may allow for the description of
deficiencies or excesses in one or more
elements, where significant external evidence
of foliage/tree disturbance is present.
Quantitative measurements of foliar nutrients
may correlate with visual symptoms that are
due to gaseous air pollutants such as the
ozone-induced reductions in N, P, K, Ca, and
Mg (Allen, personal communication), or the
deposition of foliar toxins. In the absence of
visible symptoms of disturbance, nutrient
deficiencies/imbalances of potential toxins,
may serve to explain past and future growth
reductions. The goal of this indicator is to
describe the relative nutrient balance of trees
as opposed to quantifying individual
concentrations of a few 'important' nutritive
elements. The development of a basic
understanding of how nutrient stress and
contaminants limit forest stand productivity is
an essential first step if FHM is to detect
changes in growth and/or to predict whether
or not they will occur due to changes in
climate, atmospheric pollutants, carbon
dioxide, or silvicultural practices.
The proposed variables for analysis
include elements considered to be essential to
plant growth and several potential
contaminants. This indicator and the
associated variables are consistent with other
international forest monitoring programs such
as Canada's Acid Rain National Early Warning
System (ARNEWS), the United Nation's
International Co-operative Programme on
3-15
-------
Assessment and Monitoring of Air Pollution
Effects on Forests, and the Nordic Council of
Minister's Integrated Monitoring in the Nordic
Countries.
Foliar nutrient concentrations are
known to vary in response to a number of
biological, structural, and environmental
factors. These factors need to be accounted
for to reduce excessive variability. Temporal
variation of nutrient concentrations in plant
tissue exists from year to year, within year,
and within season. Minimizing or accounting
for the potentially high between-year variation
is of paramount importance to successful
trend assessment. Temperature and
precipitation patterns have been associated
with annual alterations in foliar nutrient
concentrations (Leaf et al. 1970). Thus,
techniques for quantifying the influence of
weather patterns on foliar elemental
concentrations (Bickelhaupt et al. 1979) will be
evaluated in future assessment activities.
Stability of elemental concentrations
throughout the sampling period is necessary
to ensure interpretability of samples that are
collected during the ten-week field season.
Nitrogen, P, and Kconcentrations in deciduous
foliage generally increase early in the growing
season and decrease during or after periods
of rapid leaf growth. Seasonal changes in
other nutrients vary by species and soil (Alban
1985; Woodwell 1974). Since foliar collections
will be confined to one site visit during the ten-
week field season, the sampling procedures
have been designed to limit nutrient variability
resulting from sampling of tissue.
Variation between trees within plot
may be due to such factors as age of tree,
crown class, and soil characteristics.
Sampling of the previous year's coniferous
foliage will control much of the expected
between-tree variation. In stands of relatively
low site index, the between-tree variation in
nitrogen concentrations was lower in one-year-
old needles as opposed to current year
growth.
Element concentrations are also
known to vary within a tree, by position in the
crown, and age of foliage. Several different
patterns in foliar nutrient distribution within
tree crowns have been reported for various
species. Sampling will be restricted to full sun
foliage in the upper third of the crown to
eliminate this component of variability during
sampling. Concentration with site index and
soil nutrient concentrations are both important
to the proposed use of foliar nutrients/con-
taminants as exposure/habitat indicators.
3.1.4.4 Establishing Baseline Condition
Inherent in the application of the FHM
assessment framework is the realization that
the central issue in today's assessments of
environmental health is not: "How do we return
ecosystems to the undamaged state?" Given
that ecosystems are continually changing and
that humans will never be removed as a major
impact upon the environment, attention must
instead focus on determining what
remediation level can realistically be attained
for a given locale versus the societal concept
of desired condition with respect to
designated area use. Due to the differences in
ecosystem inter-relative functions and human
use levels, this question will be answered
differently at different points on the globe.
The assessment of ecosystem
response to environmental change requires
that baseline conditions be determined which
incorporate historic variation in ecosystem
status. The baseline approach estimates the
state of the ecosystems in the absence of
anthropogenic disturbances. Future
ecosystem status would then be compared to
the baseline state. The baseline approach
addresses natural variability in the attainability
of ecosystem condition and recognizes that a
proportion of the ecosystems within a given
3-16
-------
region may never achieve a predefined level of
"health". Techniques useful in this inference
include dendrochronology, historical record
examination, pedogenic examinations, and
isotopic ratio characterizations. this
information in and of itself does not
define the reference condition, but describes
regional population distributions in a less
anthropogenically-impacted state. Both the
inherent variability of ecosystem response
(spatial and temporal) and the variability in
ecosystem characteristics (e.g., internal acidity
sources) can be identified by this approach.
As a hypothetical example of how
baseline sub-regional/sub-population in-
formation might be used in an ecosystem
characterization framework, consider the
policy-assessment question: "What is the
distribution of regions where a specific forest
species can be expected to grow well?" The
specific questions that a regional
demonstration may answer are:
• What is the range of a particular species
where it is the dominant species and is
vigorous?
• What is the range of the species where it
is dominant but not vigorous?
• What is the range where the species is
vigorous but is not dominant?
• What is the range where the species is
present but neither dominant nor vigorous?
• What is the range where the species is not
and cannot be located?
This information provides a data
overlay that can be used to seek parameters
within each of these ranges which covaries
with the species range (e.g., available soil
nutrients, moisture, temperature, elevation,
etc.). After these parameters are identified,
the groupings of co-occurrences can be
mapped, thus delineating an empirical model
of expected ranges. Given this model, it will
be possible to evaluate actual monitoring data
in the context of scenario projections as to
how the species range will shift in response to
changes in stress(es) distribution (e.g., global
climate change, human population shifts, or
atmospheric pollution). Such estimates can
then be included in interpretive reports
evaluating discrepancies in rates of change
between environmental parameters and
ecosystem components.
3.2 INDICATOR DEVELOPMENT
PROCESS
The proposed approach is designed to
provide information about ecosystem condition
that is relatively free of interpretation bias.
This will provide user flexibility which is vital to
the differing needs and priorities of the large
client base served by EMAP. The framework is
designed in the form of a progressive flow
diagram with specific decision criteria driving
progression from one level to the next (Figure
3.4). The framework guides indicator
development through an assessment process
that considers needs and objectives,
acceptable data uncertainty, appropriateness
of available analytical procedures, data
management procedures, statistical pro-
cedures, and the need for integrative
assessment among multiple indicators. Thus,
assumptions inherent at each phase of
indicator development are formally considered
and presented for peer review. Forcing formal
consideration of assumptions is perceived as
essential to the uniform development of
indicators suitable for a national monitoring
program because program design and
selection of measurement criteria are often
based on the •cumulative learning" and/or
opinions of the participating personnel.
Specific phases of the development process
are discussed in more detail in sections 3.2.1 -
3.2.5.
The strategy for implementing the
indicator development process is presented in
3-17
-------
IDENTIFY ISSUES/ASSESSMENT ENDPOlNTS
OBJECTIVES
Develop (dlloitod o<
andpolut atetu«
Experl Knowledge
Liuraivr*
Coiteepmei Uofltla
EQUATION
Cflterla •«(}
CANDIDATE INDICATORS
Ph«te 3
PHeriilia ei«»i
ciiiiii
Experl Kno*|«#9i
Utaralvd Nirl««
Conceptual Menu
Criteria Welt
Paor Review
RESEARCH INDICATORS (Pilot »6afa Uattng)
Evaluate Paitormanot
Aftalyie aalailno data
Slflinlilloni
f laid tatte
IIKIiUoil •ittii(l»«ila
Canoaptual eiadeie
Criteria M«tf
Pear Ri«li«
DEVELOPMENTAL INDICATORS
Evaluate pafloraaAoa
Aaeaea logiallqa
Coal eiladttvev
'laid taata
tiatlaiioal i
Criteria Mar?
Pear Rev lev
Aqena? Ravlav
CORE INDICATORS
Ifliplamani regional
and national
ifiaaiicring
Data «ftatr*ta
Valua added?
Aaanay and peer
ravle*
Aaaeea new laulaitori
Evaluate aaaaaamam endpoieta
Figure 3.4. Flow diagram representing the indicator development process.
Figure 3.5. Briefly, the strategy recognizes
that:
1. Indicators must be developed and used
within a unifying framework that
represents the best current understandings
of integrated forested ecosystem process
and functionality (i.e., conceptual and/or
dynamic process models).
2. "Health" is a relative term which can best
be described by the range of conditions
distributed throughout a population.
3. Though individual measurements of state
variables are expected to serve well as
indicators of forest ecosystem condition,
all individual components of the system
exist in an interdependent balance. Thus,
the development of indices that aggregate
3-18
-------
+. DEVELOP MULTIPLE CONCEPTUAL FRAMEWORKS (MODELS)
STAND DYNAMICS
PROCESS-ORIENTED
* CHARACTERIZE FOREST POPULATION DISTRIBUTIONS
BASELINE STATUS AND EXTENT
REFERENCE CONDITION
+ DEVELOP AGGREGATED INDICES OF INDICATOR EVALUATION
SOIL PRODUCTIVITY INDEX
FOLIAR NUTRIENT INDEX
-~ PHASED IMPLEMENTATION OF THE DEVELOPMENT PROCESS
SPATIAL
TEMPORAL
INDICATOR DEVELOPMENT PILOTS THROUGH IAG AND CO-OPS
NEW INDICATORS AND/OR INDICES
MODIFICATION OF METHODOLOGIES
ENHANCEMENT OF ASSESSMENT FRAMEWORK
REGIONAL IMPLEMENTATION VIA FS IMPLEMENTATION PRIORITIES
Figure 3.5. Indicator development and implementation strategies.
multiple components' responses (e.g., the
balance of nutrient flows between plants
and soils may reflect differences in state
capacity and actual state) are expected to
provide an "early warning" of the onset of
system dysfunction prior to the
appearance of symptomalogy.
4. The FHM program will not be capable of
implementing a completely integrated
program immediately. For example, some
indicators are better suited for immediate
implementation due to historical research
and development (e.g., examination of crown
condition), and some may be better suited to
implementation in specific regions of the
country (e.g., longer-lived trees provide longer
history for dendrochrono-logical reconstruc-
tions of disturbance). Therefore, indicator
development and testing will be early priorities
of the program, and implementation of full-
scale monitoring will proceed only after
successful testing (criteria and procedures
described more fully in sections 3.1.4.1 -
3.1.4.3).
3-19
-------
5. The technical research community has the
expertise to develop and test indicators.
The primary means of infusing new
indicators into the monitoring process will
be through cooperative and interagency
agreements with universities, research
institutes, national laboratories, and other
federal agencies.
3.2.1 Identification and Evaluation of
Candidate Indicators
The purpose of the identification and
evaluation step is to propose indicators best
suited for characterization and monitoring
purposes. Indicators reflect the nature and
application of assessment endpoints, must
characterize the forest resource, and are the
primary means of reporting ecosystem status.
For any ecosystem, a number of candidate
indicators exist that could be selected.
Because there are a variety of levels at which
assessments may be conducted, the EMAP
indicator development framework is designed
to foster comparability among disparate
assessment approaches by distilling the
process to a common set of steps. Selection
of candidate indicators for research and
developmental testing will be a function of
several interacting factors:
1. Whether or not a linkage can be made with
the assessment endpoints (Table 3.1).
Inclusion for development in the monitoring
program will be tied specifically to how
well the proposed indicator is expected to
feed into and enhance the assessment
framework (Section 3.1.3).
2. The availability of data. Are data available
which were collected in a manner
appropriate for application in a national or
regional context (i.e., represented in
models, representative of regional resource
distribution, indicative of ecosystem
change, etc.)? Large quantities of data are
already in existence that can be analyzed
to characterize ecosystem condition and to
develop response models. In cooperation
with the FS-FHM, FIA, and Forest Pest
Management (FPM) programs and through
the multi-agency information management
agreements that will be developed (see
Section 10), FHM will gain substantial
capacity for adding and/or improving
indicators for the monitoring program. The
level of available analytical data will
probably vary for both regions and
stressors because of disparate percep-
tions of the key operational processes at
differing ecosystem scales and varying
degrees of data base development for
different regions.
3. The consequences of uncertainty. There is
always a component of uncertainty
associated with an environmental
assessment. Because the FHM approach
will require the linkage of multiple
components in the stress-ecosystem
relationship (estimation of stress
exposure, assumption of processes
mitigating or exacerbating ecosystem
response, and variation in genetic
response capabilities of receptor
organisms), additive increases in the
uncertainty accompanying the re-
presentation of system response will
result.
4. The characteristics of the ecosystems
under consideration. This includes the
response characteristics of ecosystems
and their spatial distribution. For example,
differences in characteristics such as soil
depth, physical structure, chemistry,
topography, and hydrology may require the
use of different stand biomass algorithms
to describe the same species. Within any
region, these parameters may vary
substantially. Hypothetically, this would
create a range of response potentials and
diverse baseline conditions within the
same region.
5. The spatial extent, magnitude, and
temporal domains over which stress
exposure occurs. Exposure to a stress
3-20
-------
may only be detrimental to forest condition
during certain times of the year, and
thresholds of critical exposures may differ
both spatially and temporally. Estimation
of ecosystem condition requires an
understanding of how an ecosystem will
respond over time to differing stresses and
stress loads. This estimation must be
based on an understanding of the physical,
chemical, and biological processes involved
in response and will be further complicated
by synergistic effects between stresses
(e.g., acidification effects of nitrogen and
sulfur). In addition, special attention must
be given to the spatial scale of analysis
and to the spatial representation of data
because the geographic distribution of
forest cover types and responses, stressor
deposition estimation, and potential for
stress abatement may differ.
3.2.2 Research Indicator Phase (Pilot
Testing)
Following identification of candidate
indicators that are believed to offer a high
likelihood of utility in a monitoring venue,
research pilot tests will be conducted to
evaluate the suitability of the indicator for
regional field trials. Pilot tests will enable
analysts to determine whether the indicator
will continue to be used in the program, be
dropped from consideration, or be "shelved"
until new technology makes use more feasible.
Figure 3.6 presents a flow diagram that
outlines the specific issues that must be
addressed within indicator evaluation field
pilots.
Pilot studies may be conducted at
multiple scales, but the following questions
are central to the success of any pilot. "What
is the specific objective of the pilot?' "By what
pre-determined criteria does FHM evaluate the
success or failure of the indicator(s) being
tested?1 Research must be designed
appropriately to answer a specific set of
questions with quantifiable certainty.
Examples of acceptable objectives for pilot
tests include:
• Measurement method modification for
regional suitability.
• Hypothesis testing as to whether or not
the indicator is related to system condition.
• Determination of the logistical require-
ments necessary for implementation.
• Comparison of multiple methods to
conduct the same measurement.
• Estimation of components of spatial or
temporal variability in indicator mea-
surements.
Examples of minimal acceptance criteria
warranting advancement to a regional
demonstration include:
• Spatial variability less than n standard
deviations.
• A specified level of regional covariance
with another indicator.
• Measurement error less than n percent.
• Cost effectiveness.
• Field crews must be able to complete
sampling in one field day.
The appropriate plot sampling design
must be developed after the objectives and
acceptance criteria have been established.
Insights into this phase should be forthcoming
from the candidate indicator selection phase.
An objective of a pilot test might also be to
select the plot sampling design most
appropriate to an indicator. The execution of
the pilot study will provide the first-level
evaluation point. It will answer questions
regarding appropriate field crew make-up,
other logistical constraints and needs for
implementation, and whether or not the
indicator can be implemented in a cost-
effective manner.
These questions lead to an analysis of
the data collected and its quality. If the data
3-21
-------
Indicator
Decison
Good
vs. Bad
Criteria
Options
Resource
comparisons
No
Drop
Yes
No
Yes
Regional Testing
Measurement Error
Field Crew Make-up
Cost Effective
Pilot Scale Testing
Sampling Design
Pilot Scale Testing Data
OA/QC Appropriate
Logistical Lessons
Plot Size Requirements
Aggregation Issues
Plot
Regional
Indicator Integration
Spatial
Temporal
Landscape position
Within Population:
Spatial variability
Temporal variability
Any variability
Figure 3.6. Indicator development issues.
meet the required quality criteria, analyses can
be completed to determine the utility of the
indicator. It is necessary to maintain the
program flexibility and delay the progress of
indicator implementation to assure the quality,
utility (i.e., interpretability within the context of
assessment endpoints), and effectiveness of
the indicators in the national monitoring
program. One aspect of this can be
accomplished through modeling evaluations in
hypothetical environments. For example, given
measured spatial and temporal variances, a
mathematical evaluation of the regional
sampling intensity needed to obtain
reasonable results can be estimated. Such
information can be used to provide criteria
limits within which an indicator must perform.
If the criteria are not met, specific indicators
may be dropped from consideration or
retested in another pilot study after
modification of the measurement procedure
and/or plot sampling design. If the criteria are
met, an indicator is ready for advancement to
the developmental testing phase (regional
3-22
-------
test). Discussions regarding the status of
indicators measured In the 1990 Pilot can be
found in Section 3.3.
3.2.3 Developmental Indicator Phase
(Regional Testing)
After the successful research pilot
testing phase, an indicator is advanced to a
regional performance evaluation. In the
developmental phase, the indicator will be
tested by using the sampling frame, methods,
and data analyses intended for the EMAP
network. The specific objective here is to
identify a subset of indicators that are suitable
for full-scale program implementation. These
issues will be tested further by using the
EMAP sampling and data analysis protocols at
regional scales. Regional demonstrations will
be used to test whether or not data collected
for these indicators are regionally interpretable
and to confirm the results of site-specific pilot
studies on regional scales. Data from regional
demonstration projects will be assessed
through peer, agency, and public review of the
raw data, regional statistical summaries, and
associated interpretive reports. The primary
product of this phase is a set of core
indicators for implementation in routine EMAP
monitoring efforts. As with the research pilot
tests, the developmental tests will be focused
to achieve specific sub-objectives (Figure 3.7).
Test Interpenetrating Design Landscape relationships
Resource comparisons Suite Integration
No
Logistical Lessons
QA/OC Appropriate
Core Implementation
Sampling Design
Regional Testing Data
Regional Indicator Testing
Integrity Indei Development
Ongoing
Intra-lndicator Integration
Spallat
Temporal
Inier-lndlcator Integration
Spatial
Temporal
Regional Comparative Analyses
Resource comparisons
Integrative variability
Wllhln/Among Population
Spatial variability
Temporal variability
Any variability
Best Data Analysis Mode
Stratify
Aggregation Issues
Trend Analysis
Figure 3.7. Developmental indicator assessment objectives.
3-23
-------
Regional demonstration tests provide
information such as the most appropriate
reporting framework for the indicator and the
similarities and differences of temporal and
spatial scales of the process and evaluation
scheme. Figure 3.8 illustrates one option for
developing a reporting framework. The issues
of concern are embodied in the assessment
endpoint definitions (sections 1.4 and 3.1.3).
The regional demonstration provides the
resource characterization data necessary to
proceed with framework development.
Subpopulations can then be delineated repre-
senting spatial aggregates of response
characteristics (i.e., indicator response or
status is more similar within a sub-population
than between sub-populations). Information
can then be aggregated for analysis and
reporting according to these response-similar
sub-populations. Such analyses may also
focus indicator developmental efforts by
providing information about expected spatial
variability and hence improving the capacity for
development of the most appropriate sampling
designs.
3.2.4 Core Indicator Phase (Implementation)
The indicator in question may now be
categorized as "core" and is ready for full-scale
implementation after all criteria of the
developmental phase are met. This phase (re-
Prioritize Issues of Concern
Compile Resource Characterization Data
High Data Quality?
No
Recompile or Reclassify)
Yes
Assess Coincidence or Common Patterns
in Resource Characterization Data
Characterize Variability Within
and Among Sub-Populations
Delineate Sub-Populations
Figure 3.8. Indicator reporting framework.
3-24
-------
evaluating and modifying the suite of core
indicators) is an ongoing process that begins
upon initial implementation of core indicator
monitoring at regional and national spatial
scales. This continual process of reinspecting
the indicator suite ensures complete indicator
coverage of important environmental values,
assessment endpoints, and stresses,
incorporation of appropriate advances in
technology and information, and adequate
capability to detect changes and identify
trends in the status of ecological resources.
In this phase, it is Important that EMAP
balance continuity of methods (to maximize
trend detection capability) with procedures to
refine or replace indicators that fail to perform
satisfactorily. This phase is implemented
through procedures that are designed to
critically review the performance of core
indicators through a time-series of regional
frequency distributions, to evaluate alternative
indicators to address emerging issues and
inadequate core indicator performance, to add
new indicators as deemed desirable, and to
substitute superior indicators for inadequate
core indicators.
The EMAP monitoring program will
document the co-occurrence of stressors and
affected populations or sub-populations. Such
information is not adequate for the
assignment of cause-effect relationships
which is not an objective of EMAP.
Nevertheless, full-scale implementation of the
core indicator suite will provide a mechanism
through which other research programs can be
focused to fully develop this relationship
(Figure 3.9).
Examine Sub-Populations
Problem/Response
~~r~
Characterize Temporal/Spatial
Patterns of Response
Presence/Absence of
Known Stressors
I
Spatial/Temporal Coincidence
of Stressor(s)/Eflects
1
Delineate Areas for Diagnostic Research
Figure 3.9. Pointing toward causality.
3-25
-------
3.2.5 Indicator Addition and Replacement
Although reference to indicator
acceptance and failure has been made in each
indicator section, it is important to point out
that the program will not continually add new
indicators to the field program. As a national
monitoring program, EMAP will add and/or
delete indicators depending upon their
capacity to provide necessary information to
interpretation and assessment. However, the
number of indicators to be measured will be
strictly limited and prioritized according to the
value added in characterizing ecosystem
status and trends in condition. Redundancy
among indicators providing the same infor-
mation will be perpetuated only as long as it
takes to evaluate their relative value.
3.3 IMPLEMENTATION PLAN FOR FY91
3.3.1 Lessons From FY90 Activities
Figure 3.10 summarizes the status of
indicator development within EMAP-Forests. A
peer workshop was conducted in the spring of
1989 to identify a number of candidate
indicators expected to have a high potential
for interpretability and applicability with
respect to forest condition. To date, EMAP-
Forests has tested six of these in pilot
studies. These indicators are percent ad-
STATUS
CORE" AND "DEVELOPMENTAL"
INTERPRETABILITY
LOW-
HIGH
FEASIBILITY
AN DEMOGRAPHICS
VEG A BUND/DIVERS
CARB0S/2ND CKEMS
LITTER DYNAMICS
AN MORPHOLOGY
AN BIOMARKERS
MOSSES & LICHENS
SOIL BIO. PROCESSES
NITROGEN EXPORT
STABLE ISOTOPES
"DEVELOPMENTAL"
GROWTH EFFICIENCY
FOLIAR CONTAM.
HIGH
AN SPP/GUILDS
FOLIAR NUTRIENTS
HABITAT LIN CLASS
SOIL PRODUCTIVITY
VISUAL SYMPTOMS
LANDSCAPE PATTERN
"CORE"
Figure 3.10. Status of EMAP-Forests indicator development.
3-26
-------
sorbed photosynthetically-active radiation
(PAR), vertical vegetation abundance and
structure, foliar chemistry, soil characterization
and chemistry (see Section 3.1.4.3 for more on
foliar and soil indices), growth (mensurational
measurements), and visual symptoms. Palmer
et al. (1990) provides a complete description of
indicator sampling methodologies and design.
Samples collected during the summer of 1990
are being analyzed, and statistical summaries
are being prepared. Much logistical
information was gained including the sampling
and data handling and storage and transfer
procedures required for these indicators (see
Sections 9 and 10). Preliminary examination
suggests that the methodologies for all
indicators except PAR appear to be
satisfactory for movement to the devel-
opmental phase. The PAR indicator
experienced a problem related to weather
conditions and measurement variability;
resolutions are being designed. However, final
status decisions will not be made until data
analyses are completed.
3.3.1.1 Vertical Structure of Forests
The overall objective of 1990 summer
activities was to test the utility of forest
vertical vegetation structure as an indicator of
ecological condition and environmental and
anthropogenic stress. Some of the main
reasons that vertical structure of forests is
proposed as an ecological indicator include:
1) Vertical structure or forest profile has
biological relevance as an element of the
diversity of plant communities.
2) Animal and plant species richness and
diversity is positively correlated with the
degree of forest stratification (e.g.,
MacArthur and MacArthur 1961; Willson
1974; Harper 1977; Dueser and Shugart
1978; August 1983).
3) Conservation and maintenance of animal
and plant species diversity has been
identified as an important public value.
4) Forest vertical structure is susceptible to
anthropogenic stress. For example, forest
profile is routinely manipulated by forestry
practices such as logging, plantation
establishment, and thinning. In some
situations forest profile may be greatly
simplified (e.g., when mixed-species
forests are cleared and replaced by pine
plantations) and the associated animal
and plant life impoverished (Atkeson and
Johnson 1979; Repenning and Labisky
1985; Childers et al. 1986; Felix et al. 1986).
Furthermore, structurally simple forests are
less tolerant of biotic stresses such as
disease and insect attacks (Schmidt 1978;
Knight and Heikkenen 1980).
5) Measurement of forest vertical structure is
more applicable to detection monitoring
than direct measurement of bird, mammal,
or insect populations. Measurement of
vertical structure is a component of habitat
quality. Changes in habitat quality due to
forestry or urbanization indicate changes in
the quality and quantity of animals (Figure
3.11). Measurement of vertical structure is
relatively easy (i.e., sessile organisms) and
inexpensive (i.e., one visit by a one- or two-
person crew is probably sufficient).
Past and Ongoing Work
While there is sufficient reason for
considering vertical structure as an ecological
indicator, additional questions need to be
answered to justify monitoring forest profile to
characterize forest condition and to detect
environmental and anthropogenic stress.
These questions include:
1) How responsive is forest profile to more
subtle forms of disturbance such as
thinning or to environmental gradients of
moisture and nutrient availability?
2) How well can the effects of anthropogenic
stress and stress due to natural factors be
separated?
3-27
-------
Measurement
End points/Response
Number and area,
fragmentation,
and connectivity of
Maridscapo units
Species Composition,
vertical structure,
and paichlness of
forest stands
Landscape-level habr.at
condition of forests
Stand-level hattfat
condition
Assessment Endpoint
Changes in Haoitat
Quantify and Quality
of Forests
Societal Value
Conservation and
maintenance of plant
and animal abundance
and diversity
Figure 3.11. Relationships of measurements, indicators, and assessment endpoints for
maintenance of plant and animal diversity.
3) Does forest profile differ by forest type?
4) Does vegetation profile change predictably
as forests age?
5) What is the spatial variation in forest
profile within a region?
Some of these questions have been or
are being addressed through two pilot
projects: 1) the 1990 20/20 Study and 2) an
exploratory analysis of an existing geo-
referenced data base of more than 3100 forest
inventory plots on the Georgia Piedmont. This
analysis is designed to test hypotheses
regarding relationships of forest profile with
tree species composition, environment, and
disturbance. Preliminary analyses of the 20/20
study can be summarized as follows:
1) The ocular method is 2 to 3 times faster
(ca. 25 min) than the point method as
implemented in the pilot study.
2) The ocular method provides estimates of
foliage occupancy that are 2 to 3 times
higher (ca. 45%) than the point method.
3) The ocular method gives a less specific
profile that has only two dimensions, as
compared to the point method.
4) The ocular method and the point method
estimate the same number of broad
species classes.
3.3.1.2 Soils (1990 20/20 Study)
The overall optimization goal of soil
sampling in the 1990 20/20 study was to
construct a design that would reduce within-
plot spatial heterogeneity within some bounds
that would be acceptable to the data users.
The two primary objectives of soil sampling in
the study were to (1) estimate within-plot and
withirvsubplot spatial variability in soil
characteristics, and (2) test the feasibility and
logistics involved in implementing soil
productivity monitoring on a nationwide scale,
The resulting data are expected to be used to
optimize the soil sampling design in the 1991
field season and beyond.
The issue of temporal heterogeneity in
soil chemistry was less critical to the
optimization effort because it is envisioned
that the logistics staff will design a plot
sampling sequence for each region that will
enable each designated plot to be resampled
at about the same interval of the index period
over the course of the project.
Specifically, the soils staff of the 20/20
Study is investigating:
3-28
-------
• Uncertainty stemming from single-hole
sampling as opposed to multiple-hole
sampling at each plot.
• The utility of compositing master horizon
samples while in the field versus
compositing the data from replicate master
horizon samples.
• Whether the provision of destructive-
sampling zones between the fixed-radius
subplots will allow collection of soil data
that are representative with respect to the
forest vegetative data.
• The number of soil holes within a plot that
must be sampled.
• Required sampling depths and the types of
horizons that should be sampled.
• A determination of whether samples will be
composited and at what stage.
• Logistical considerations relating to the
resources required to characterize a plot
and collect the necessary samples.
• The utility of laboratory analytical methods
that were selected.
• Reporting units for the many different soil
parameters.
• Identification of any ancillary data that
may be needed to link the component
variables of the soil productivity indicator.
Logistics information related to the
seventh bullet above has already been
evaluated. Upon completion of the soil
analysis activities in April 1991, relevant
questions related to the other nine specific
objectives can be answered.
3.3.1.3 Foliar Nutrients (1990 20/20 Study)
The primary objective of this study was
to determine the within-tree, within-plot, and
between-plot variance components for foliar
nutrient and chemical contaminants for the
tested sampling design. A secondary objective
was to determine the analysis effects of
compositing samples within trees. If a
primary contributor to the total within-plot
variability can be determined, then plot
sampling can be designed to reduce the
overall variability. Similarly, the study of
composition effects is needed to determine
where compositing is most effective in terms
of cost and variance reduction.
In the Southeast and Northeast, 10
plots were selected for each region from the
40 plots selected for the overall study plan. At
each plot, two branches were sampled from
each of six dominant/co-dominant trees of the
selected species. The visual damage indicator
was implemented concurrently on the same
trees. The same plots were also selected by
the soil nutrient/contaminants indicator for
intensified soil characterization.
Foliar samples have been dried,
ground, weighted, and placed in storage.
Chemical analysis will commence in January,
1991 and data analysis is expected to be
completed by April, 1991.
3.3.2 FY91 Pilot Field Objectives
Figure 3.12 summarizes the outcomes
of a joint planning meeting between the EMAP-
Forests and the FS-FHM planning groups.
Briefly, the FS plans to conduct full-scale
implementation of the visual symptoms and
growth/mensurationalmeasurements within all
hexagons in six New England states, Georgia,
and Alabama. Additionally, the FHM program
will collect soil characterization data at 1/4 of
the hexagons in these states (EMAP
interpenetrating sampling design).
The EMAP-Forests program will
implement a regional pilot study in which it will
add soil and foliar chemical analyses and
measurement of the vertical vegetation
structure and distribution to the FS's list at 1/4
of the hexagons. The objectives of this pilot
study are twofold: (1) to characterize the
spatial variability in the individual indicators
within and among populations in these states
(regions), and (2) to test the hypothesis that
3-29
-------
OUTCOMES OF JOINT PLANNING MEETING WITH FOREST SERVICE
DEC. 3-6; DENVER. CO
0 FS WILL IMPLEMENT IN NORTHEAST, GA, AL. Ml:
/ VISUAL SYMPTOMS
ALL HEXES ^ QROWTH/MENSURATIONAL MEASUREMENTS
1/4 OF HEXES — SOIL DESCRIPTIVE CHARACTERIZATIONS
O EPA/EMAP-FORESTS WILL CONDUCT REGIONAL PILOT IN NORTHEAST, GA, AL, Ml
/ SOIL CHEMISTRY
1/4 OF HEXES
-------
structure to meet the objectives developed in
activity 1 (including defining strata, estimating
patchiness, varying number and area of
sample units for species counts) and
improving real-time remeasuremerrts.
3) Continuing with the analyses of existing
data with the objective of conducting
"monitoring on paper" for detection and
assessment of status and trends.
3.3.2.2 Soils
The primary objectives of soil sampling
in the 1991 field season are to:
• Demonstrate that the soil productivity
indicator, optimized for available funding
and personnel, can be successfully
implemented in two large sub-regional
forested areas of the eastern United
States utilizing a cooperative effort among
multiple agencies.
• Continue to develop key components of the
soil productivity indicator and evaluate its
utility in synthesis and integration with
other ecological indicators.
• Begin to construct regional estimates of
the concentration ranges for critical soil
parameters used in the interpretation of
soil condition with respect to the overall
assessment endpoints.
• Develop draft data quality objectives for
the various levels of soil data collection
identified in Section 8 of this document.
3.3.2.3 Foliar Nutrients
The primary objective for 1991 will be a
literature review of foliar nutrients/ contami-
nants. Research will be directed toward the
development and refinement of techniques to
diagnose nutrient limitation and imbalances in
forest stands.
A secondary objective will be to utilize
the data from the literature review and the
20/20 Study to perform standard components
of variance analysis in the presence of
measurement error. System measurement
error will be estimated from the 20/20 Study
and will be used to begin separating
measurement error from variability estimates
for within-tree, between-tree, within-plot, and
between-piot. Future analysis will address the
use of other indicators (e.g., soil nutrients,
foliar area index) as covariates.
A third objective is the acquisition of
historic data bases that contain data on foliar
nutrients for large numbers of trees over at
least a three-year period. Analysis of these
data bases will address questions of spatial
and temporal variability.
Data from the 1990 field pilot studies
is being used to examine particular
components of variability in foliar chemistry.
Other such components will be examined in
future field seasons. Although an Important
first step, it is not enough to use research
information to modify design to reduce
specific variability components. We must also
quantify all components of variability to
determine whether or not we will be able to
use the data in the EMAP assessment
framework. A fourth objective is the design of
a field pilot study for implementation in at
least one ecoregion for the 1991 field season.
The pilot study would serve as a logistics test
of a large scale implementation and it would
also provide estimates of spatial variability
across an ecoregion for non-plantation plots.
3.3.2.4 Visual Damage Survey (Symptoms)
Visual symptoms refers to a suite of
pathological and entomological measurements
to assist in the assessment of forest health
and status and trends of disease. Disease is
defined as any deviation in the normal
functioning of a plant caused by some type of
persistent agent. In the case of decline
diseases, a complex of agents including biotic
and abiotic components may lead to the
3-31
-------
diseased state. Visual symptoms
measurements are intended to detect any
condition falling outside the generally accepted
norm for a species (i.e., the baseline). Specific
components included are listed in Table 3.3.
The measurements proposed have been used
in some form in other research projects and
established monitoring programs (Anderson
and Belanger 1986; Alexander and Carlson
1989; Magasi 1988; Millers and Lachance 1989;
Anonymous 1987).
Various methods of estimating foliage
amount have been investigated and used to
classify tree condition in monitoring and
survey projects. Crown density (Alexander and
Carlson 1989; Anderson and Belanger 1986)
was developed for southern pines. Crown
transparency (Millers and Lachance 1989) was
developed for sugar maple (Acer sacharrum
Marsh.). The European crown rating method
(Alexander and Carlson 1989; Anonymous
1987) was developed for use on both conifer
and deciduous trees and has been the method
of choice by the United Nations Economic
Commission for Europe for the past six years.
Hie performance of these three methods will
be compared during the FY91 field season.
Additional information will be gathered
symptoms indicating air pollution exposure
documenting the incidence of plants with
symptoms (Anderson et. al. 1989).
TABLE 3.3. Visual Damage Survey Variables.
Plot and Sample Trees
Sample Trees Only
Elevation
Tree height
Slope
Height to live crown
Aspect
Density and diameter
Stand disturbance
Increment cores
Air pollution indicator species
Mainstem injury - Type
Tree species
- Location
DBH
Crown - Needle retention (binoculars)
Crown • Ratio (estimate)
- Dieback
- Class
- Dwarf foliage
- Discoloration
- Epicormic branching
• Defoliation
Branch - Needle retention (observed)
- Needle length
• Twig symptoms
- Leaf Symptoms
- Damage class
- Discoloration class
- Discoloration type
Root signs and symptom
3-32
-------
The objectives of the 1991 field
program will be to estimate components of
variance in visual symptoms (spatial, among
trees within plots, and among plots and
sub-plots). Comparisons will be made of
different methods of air pollution indicator
plants (Alexander and Carlson 1989; Anderson
et al. 1989) and crown foliage measurements
(Alexander and Carlson 1969; Anderson and
Belanger 1986; Millers and Lachance 1989;
Anonymous 1987). Results will be used to aid
in the development of a standard set of
methods for the national monitoring program.
The visual damage survey is an uncontrolled
field survey. Experimental units include both
tenth acre and 10-point plots (w493) and
sample trees (w1972).
3.3.2.5 Percent Adsorbed Photosynthetically
Active Radiation
The Percent Adsorbed Photosyn-
thetically Active Radiation (PAR) indicator is
expected to provide information on the use
efficiency of photosynthetically active radiation
incident to the forest canopy. As a potential
surrogate for the more difficult to measure
indicator of leaf area index, PAR is planned for
use as an indicator related to net stand
production and canopy condition, a marker of
canopy closure (thus related to expectations in
basal increment area growth), and as a
component of ground-truth for remote sensing
measurements. EMAP-Forests staff have met
with various groups by way of background
research into the feasibility, utility, and
appropriate plot design for the PAR indicator.
It is anticipated that the ultimate design will
be determined by March.
The basic objectives currently planned
for FY91 activities are:
1. Methods development - test continuous
ambient sensors to complement
under-canopy measurements (proposed to
resolve problems associated with variable
cloudiness).
2. Plot design - test larger plot sizes, possibly
supplemented by pre-stratification from
aerial photos, depending on opportunities,
and test various plot protocols to deter-
mine structure.
3. Assessment - continue making PAR
measurements as part of a suite of
measurements, and try to link PAR more
closely with measures of habitat and
vegetation structure. Attempt linkage with
those who have satellite data for the
locations.
3.3.2.6 Wildlife Condition, Habitat, and
Distribution
The status of wildlife is one
component of forest ecology that is of mutual
concern to EMAP-Forests and to the U.S.
Department of Interior Fish and Wildlife
Service (FWS). EMAP-Forests is currently
exploring an opportunity to develop an
Interagency Agreement to improve
EMAP-Forests' monitoring design and
assessments as they pertain to wildlife
ecology. Specifically, the EMAP-Forests
program is seeking to increase efficiency in
the specification, development, and testing of
indicators of wildlife condition and/or habitat,
and to improve analysis and interpretation
capabilities regarding the status and trends of
wildlife components of forests.
3.3.2.7 Landscape Characterization
Because the FHM program is being
designed as a multi-agency cooperative
endeavor, it is desirable that the systematic
EMAP grid sampling design be linked within
some type of framework to existing forest
health and management monitoring programs
such as the FS FIA and FPM programs.
Linkages between these existing sampling
frameworks can be facilitated through the
3-33
-------
application of multi-level landscape
characterization monitoring.
The first level of the multi-level sample
would be designed to permit stratification on
permanent landscape features such as
landform and forest/nonforest. Several strata
could occur In any one 40 km® EMAP hexagon,
landform/forest cover delineations would then
be used to select a sample framework for
high-resolution, second-level photo-plots. For
example, nonforested strata might be sampled
at a lower intensity to monitor afforestation, or
deal with errors in detecting forest cover on
low-resolution aerial images. Habitat, forest
type, or other criteria that are expensive to
apply to entire hexagons might be used
to provide a framework for developing
extent estimates from plot-level indicator
measurement data.
The second level would be designed for
inexpensive remeasurements of a few basic
indicators of forest health. For example, tree
mortality and defoliation may be measured
using high-resolution aerial photography
and/or videography. Because high-resolution
imagery has a narrow field of view, approx-
imately 1 km2 (250 acres), high-resolution
imagery is impractical for complete coverage
of each 40 km2 primary sampling unit. A
second-level sample plot is proposed using 3
to 10 second-level photo-plots in each 40 km2
first-level sample unit to accurately estimate
tree mortality and tree defoliation. These
conditions are often rare and not spatially
contiguous (although there are many
exceptions), and large photo-plots would more
efficiently quantify mortality and defoliation
than smaller field plots. The least expensive
indicator would be the number of dead or
defoliated trees per unit area (status and
extent). However, to estimate the rate of
change in mortality and defoliation extent, the
number of trees in each second-level
photo-plot might have to be estimated from
the high-resolution imagery, perhaps via
subsampling the imagery. Rate estimation
requires that each individual sample tree must
be found on two dates of imagery taken 12
months apart, possibly requiring a reduction in
the size of the second-level photo-plots to
save interpretation time. Detection error may
be significant, especially for large plot sizes,
and methods should be adopted to estimate
the proportion of dead or defoliated trees that
are not detected with interpretation of aerial
imagery. It might be desirable to use aerial
photography once every 5 to 10 years for
estimating forest type, tree heights, tree
species, regeneration, fuel loading, habitat
type, stocking density, and stand development,
and to use aerial videography for the same
plots in intermediate years for less expensive
measurements of tree mortality and
defoliation. An interpenetrating rotation
between aerial photography and aerial
videography is also possible.
FHM plots would be nested within the
framework of the 1 km2 second-level plots to
take advantage of the annual monitoring for
tree mortality and defoliation at the
second-level, disturbance history for each plot
interpreted from remote sensing, the need to
quantify the error in detecting tree mortality
and defoliation with remote sensing at the
second-level, and would permit extrapolation
of FHM indicator data (sections 3.3.2.1 -
3.3.2.6) to the more extensive spatial
framework. This integration within the
extensive framework would also provide a
mechanism for comparative evaluation of FIA,
FPM, and FHM data.
Concerns
Efficiencies and precision are gained
by emphasizing remote sensing, but there is
limited infrastructure in place to acquire,
coordinate, interpret, and archive this source
of data. To assure consistency and quality,
the remote-sensing activities would have to be
institutionalized. Ideally, there would be a
3-34
-------
small number (maybe one) of units that have
direct responsibility for this function. The
unit(s) might be branches of existing units
with related missions, such as FIA, FPM, or
state forester agencies.
Synergistic benefits
FPM currently produces annual
assessment reports for insects and diseases
in the west. It might be possible to produce
these same reports using annual defoliation
estimates from high-resolution aerial
photography and less frequent field
examinations of FHM plots. FPM might be
able to make minor adjustments to its current
program to contribute to FHM, while meeting
its current objectives in a perhaps more
efficient and rigorous manner. Similarly, there
are several new monitoring initiatives in the
west: detection of possible effects from global
climate change, and changes in condition of
Wilderness Areas. It might be possible to
design one or two compatible sampling
frames that more efficiently serve several
different sets of objectives.
The use of PROGNOSIS as the
baseline for growth and mortality can also be
used to validate and improve this model.
PROGNOSIS is commonly used by the FS
National Forest System (NFS) for their
strategic planning (e.g., FORPLAN), and
improvement of planning models will directly
improve NFS management. As part of forest
plan monitoring, assumptions used in the
planning process must be verified. Models
such as PROGNOSIS are regional in nature,
and are collections of numerous assumptions
on growth and mortality rates that directly
affect the land management planning process.
Likewise, the use of fuel loading and forest
insect and disease risk models as forest
health indicators will lead to improvements in
those models, with a potential to improve very
expensive management actions for fuels,
insects, and diseases.
High-resolution aerial photography
could be used to reliably interpret forest type,
crown closure, and stand development on a
sample of FHM photo-plots. A subsample of
FHM plots could be very useful for labeling or
training digital classifiers of satellite data (e.g.,
Landsat), and for quality control in the
production of vegetation-cover maps. Another
subsample of FHM plots could be used to
estimate statistical calibration models that
correct for misclassification bias in areal
estimates. This would be valuable to National
Forests and other agencies for reliable
mapping of wildland resources in the west,
and unbiased areal estimates used in local
land-management strategic planning.
High-resolution aerial photography may
be suitable for estimating leaf area index or
photosynthetic efficiency, which are
measurements related to other potential
indicators of forest health. This might be
tested in future research studies.
3.3.2.8 Indicator Evaluation Field Study Plan
Timeline
EMAP-Forests, in cooperation with the
FS-FHM, has an opportunity to conduct field
studies this year. Arrangements are being
made to provide a soil scientist and a forester
or plant taxonomist to collect data on one
fourth of the FHM plots in New England,
Georgia and Alabama in the sampling frame of
the EMAP interpenetrating grid. Another
individual will also be available on a subset of
plots, probably 20 in each of two FS Regions,
for smaller scale studies. These
measurements will be made in conjunction
with the FS-FHM measurements selected by
the FS (visual symptoms, soil type
characterizations, and growth/mensuration
measurements). EMAP-Forests will coordinate
closely with the FS and develop an approved
field study plan. The following timeline and
information has been distributed to staff
authoring sections of the FY91 Field Indicator
3-35
-------
Measurement Plan and is proposed to
accomplish the study plan requirements.
Feb. 1 Letter requesting Demo Proposals in
Annotated Outline form and Commit-
ment to Implementation, Analysis and
Reporting.
Feb. 8 Section Annotated Outlines Due to
Kucera and Strickland. The annotated
outlines should address the following
FY91 Indicator Evaluation Field Study
Plan
Components:
1. Introduction; should include rationale for
inclusion of indicator in EMAP-Forests.
2. Objectives; should include statement of
specific objectives and anticipated study
outputs.
3. Justification; should include literature
and/or data analyses which support
decision to conduct field work at proposed
level (Demo vs. Pilot) and lend confidence
that stated objectives and outputs will be
met.
4. Approach; should provide all information
necessary to serve as a methods manual
for field crews. Components of Approach
should include:
• Sample collection procedures: Specific
•cook-book" descriptions of sampling
protocols.
• Logistics: What is the anticipated
level of the sampling and analysis
effort? What specific personnel,
qualifications, training, and debriefing
requirements are necessary for field
crew staffing? Estimate hours per
plot required for measurements.
Transportation, equipment and
consumable supply procurement,
communications, preparatory and
analytical laboratory, safety, inventory
and storage considerations.
• Information Management: What are
the anticipated sizes of the data files
that will be transmitting to the central
data management group?
Design: What within-plot sampling
design will be necessary for adequate
sampling coverage?
• QA/QC protocols from sample
collection through sample analysis and
data entry.
5. Reporting; should provide a description of
the anticipated structure for information
reporting. What data analysis procedures
are appropriate for your indicator? What
reporting format will you use in
communicating your results? Suggest
deliverables.
6. Timeline; Should provide a timeline for:
completion of analytical, QA/QC, and data
analysis and for delivery of reports.
Feb. 8 Field Study Section Annotated
Outlines sent to Support Leads
(Logistics, Information Manage-
ment, QA, Statistics, and Design,
Indicator Development, Integration
and Assessment, Reporting) and
FS-FHM Program Manager and
Regional Implementation and
Indicator Leads.
Feb. 12 Conference Call 3:00 - 5:00 p.m.
EST. FTS 245-4230. Subject FY91
Field Study. Selection of
measurement projects. Coor-
dination of field study.
Feb. 28 First draft of Sections sent to
editors.
Mar. 13 Edited draft sent out for internal
review.
3-36
-------
Mar. 22 Reviewed drafts returned to
authors.
Mar. 25-27 Authors workshop; internal review
reconciliation and possibly meeting
with FS counterparts.
Mar. 28 Editing.
Apr. 4-15 Word processing.
Apr. 19 Send plan out for peer review.
May 3 Receive review comments-copy to
editor and author.
May 10 Reconciliation sent to editor from
author.
May 10-13 Editing and word processing.
May 14 Document sent to laboratory for
approval.
May 31 Laboratory approval.
June 3 Pretraining and training.
3-37
-------
4 STRATEGY FOR MONITORING
NETWORK DESIGN
4.1 GENERAL STATISTICAL REQUIRE-
MENTS
The design of the Forest Health
Monitoring (FHM) program must permit
statistical estimates of condition and trends
with corresponding precision estimates. To
meet these objectives, the statistical design
must:
• Provide explicit definitions of the target
populations and sampling units.
• Provide an explicit definition of the
sampling frame for the selection of
sampling units.
• Use probability samples on the sampling
frame.
• Permit analyses of a variety of possible
subsets of the data.
• Adapt to a variety of questions, some of
which cannot be specified in advance.
• Have a structure that permits sampling at
coarser or finer levels of resolution, as
required.
This section will discuss how the
EMAP design (Overton et al. 1990) will be used
in the EMAP-Forests program and how the
above criteria are being addressed in the
EMAP-Forests design strategy. The Forest
Service (FS) inventories and monitoring
programs have been discussed in Section 1.
That discussion serves as introduction for the
discussion of the statistical designs of the FIA
and their relation to the EMAP-Forests
sampling frame in this section.
4.2 DEFINITIONS OF POPULATIONS AND
SAMPLING UNITS
4.2.1 Populations
To answer questions about the
condition and trend of forest ecosystems,
target populations and subpopulations must
be defined. The FHM target population is
defined as the areal extent of forested
ecosystem about which estimates of
conditions will be made. Target populations
can be defined by a region or an attribute. For
example, the population of interest might be
the forests of the Northeast as defined by FIA
units, only high elevation spruce/fir forests, or
all stands of sugar maple in the New England
area. At the broadest level, the target
population for FHM is all forest ecosystems in
the United States.
The development of a sampling frame
to address forest condition for all forest areas
necessitates an exact definition of a forest
ecosystem (Section 1.3). This definition is still
not sufficient to distinguish forests from some
of the other EMAP ecosystems. As an
example, forested wetlands could fall into
either forest or wetlands ecosystems.
Therefore, a cooperative effort between the
EMAP-Forests and EMAP-Wetlands resource
groups may provide a better coverage of
forested wetlands. Similarly, some people
might consider areas of chaparral as forest
ecosystems; others might consider them to
be arid lands and therefore within the EMAP-
Arid Lands Resource Group. In addition, areas
such as thick, extensive hedgerows by
agricultural lands are not clearly the
responsibility of either the EMAP-Forests or
the EMAP-Agroecosystems resource groups.
These lines of division or cooperation must be
drawn before the sampling frame and Tier 2
sampling methods can be fully developed.
Resolution of these issues is discussed in
Section 1.
4.2.2 Subpopulations
Within EMAP, subpopulations are
defined as the classes of resource types
about which statements of condition and trend
are made. In addition, certain subpopulations
will delineate any stratification that an EMAP
4-1
-------
ecosystem group decides to use. Thus,
subpopulations serve two major purposes.
They increase the precision of condition and
trend estimates by reducing extraneous
variation and target specific sets of resources
for reporting and assessment.
EMAP-Forests has identified 21
particular forest types that can be delineated
regionally. In a given region, only two to seven
of these 21 constitute major forest types.
Other possible variables by which to classify
subpopulations include stand size and class,
site index, geographical region or ecoreglon,
elevation or elevation and forest type in
combination, and landscape characteristics.
Most of these will be used for reporting or
post-stratification.
4.2.3 Sample Units
For purposes of sample selection, the
population should be divisible into what may
be called sample units. The set of all possible
sample units constitutes the population as a
whole. Identification of each sample unit is
necessary to prevent ambiguities in sample
selection. There is ongoing discussion with
EMAP-Design on the technical details of
sample units in the EMAP-Forests context.
Since forested ecosystems do not
have simple boundaries, an EMAP-Forests
sample unit is currently defined as a
contiguous area of forested ecosystem that
meets the FHM definition of forest. The
monitoring network design will specify a single
element in a sample frame as the sample unit,
and this will represent an extent of resource.
The actual plot size and geometry for Tier 2
purposes is discussed in Section 5.
4.3 EXISTING FOREST SERVICE INVEN-
TORIES AND MONITORING PRO-
GRAMS
There are a number of extant USDA FS
inventories and monitoring programs. A
number of these have been reviewed by
Hazard and Law (1989). Much of this
subsection is taken from that document. In
the past twenty years, there has been an
increasing need for forest resource inventory
data. These data have contributed to
assessment and management objectives of
various agencies and organizations. Of the 16
USDA survey units (see Section 4.4), seven are
FIA units and nine are National Forest System
(NFS) units. The NFS regions do not always
coincide with FIA regions.
The FIA has seven geographic units
responsible for surveys. The FIA inventories
provide a comprehensive inventory and
analysis of the renewable forest resources for
Resource Planning Act (RPA) assessments.
They provide information about renewable
forest resources which is used by resource
managers, including state and regional
agencies, industrial firms and associations,
colleges and universities, and state legislative
and congressional stafls. With certain
exceptions, the FIA conducts inventories on
federal, state, county, and private timber
lands. For example, most do not inventory
national forest lands or administratively
reserved areas such as national and state
parks.
The NFS inventories produce resource
information for developing, implementing, and
monitoring National Forest Management Plans.
They also produce resource information for
4-2
-------
RPA assessments and survey reports.
Resource inventories which are conducted on
each national forest may cover a wide range
of resources, including timber, range, soils and
geology, plantlife, fish and wildlife, natural
water occurrences, and quantitative data on
species and community diversity. Most
national forests exclude wilderness areas and
research natural areas from their timber
inventories. The National Forest Management
Act of 1976 mandates that managers of
federal land monitor the impacts of
management activities on all resources. Other
national forest sampling efforts include the
timber sale cruises, regeneration surveys, and
soil condition surveys.
The Forest Pest Management Program
(FPM) supplements the tree mortality
information gathered during the forest
resource inventory surveys done by the FS.
Their sampling efforts are directed toward
forest insect and disease conditions in the
United States.
4.4 OVERVIEW OF THE FOREST SERVICE
FIA DESIGNS
The FIA projects (Section 1) have
partitioned their respective regions into survey
units which are geographical areas inventoried
as separate, statistical populations. These
units are usually defined by enumerating all
counties, states, or geographical regions
within a well-defined boundary. Exclusions
such as wilderness areas are delineated so
that the exact acreage is known for each
survey unit prior to sampling. However, field
measurements are not taken on most areas of
exclusion.
4.4.1 FIA Photo Points
Data collection is usually based on
double sampling for stratification. This
procedure calls for the interpretation of
sample points on aerial photographs as the
first sampling phase. The aerial photo points
are laid out on a systematic grid over the
survey unit. Classification of points on this
grid provides estimates of forest area,
although these estimates may be augmented
by the FIA. Classification of the photo points
also provides the stratification information to
be used in the second stage of the FIA double
sample. Classification for this purpose
depends on ownership categories, land-use
classes, volume classes, and/or major land
classes (forest versus nonforest). Several FIA
units only stratify by major land class (i.e.,
whether or not the photo point is forest land).
Sampling intensities vary among FIA
regions and among survey units. The photo
paints range from one point per 190 acres in
the North Central region to 1 point per 1,400
acres in parts of the Pacific Northwest region.
The frequencies of ground plots generally
occur in proportion to the acres in different
strata.
4.4.2 FIA Sampling Methodologies
The selection strategies in the second
phase of FIA sampling vary among units. The
classification from photointerpretation is used
to select a stratified probability sample. The
ground sample plots commonly consist of a
cluster of points located over a one-to-five
acre area, although the ground sample plots in
Alaska have covered up to twenty acres.
4.4.3 FIA Measurements
The ground plots are usually
permanent, although some FIA units use
partial replacement or complete replacement
of plots over time. The plots are generally
remeasured on a ten-year cycle, except in
areas of relatively slow or fast change in
volume. For example, in Alaska cycles may
extend to twenty years.
4-3
-------
Measurements on the ground plots
provide estimates of stand and individual tree
attributes. The measurements fall into four
categories: area data, plot data, tree data, and
other vegetation data. Area data include
aspects of land use, landscape, stand
dynamics, and wildlife habitat values. Plot
data include plot age, location and history, site
index, and soil taxonomic data. Tree data
include species, diameter at breast height
(dbh), height, cull, tree quality values, tree
history, regeneration, and wildlife values as
related to merchantability, species, and size.
Other vegetation data cover foliage structure,
condition, and regeneration information.
Sections 1 and 2 discuss the reasons these
programs do not meet the needs of EMAP.
4.5 SAMPLING FRAME
In addition to identifying the population
and subpopulations of interest and the sample
units, it is necessary to develop a sample
frame. The sample frame consists of a
representation of sample units comprising the
population. One way of constructing a sample
frame is a list frame, a list of all possible
sample units, such as the one used in the
National Lake Survey. Since EMAP-Forests is
dealing with areas of forest which do not have
simple boundaries, a list frame is not
sufficient.
An alternative to a list frame often
used in a case like forests is the use of a map
as representation of the list frame. As
discussed in Section 4.4, this is the approach
taken by the various FIA units.
Tlie EMAP sampling design divides the
conterminous United States into approximately
12,600 hexagons, each of which has an area of
640 square kilometers. Within these hexagons
are smaller hexagons, the 40 km2 landscape
characterization hexagons. The hexagons
represent a systematic grid with a random
start (i.e. the hexagon centers are randomly
located by selecting a single random point in
space and moving the entire pattern so that
some hexagon center is on that point).
In the following discussion, as well as
in Section 5, the current design concept of
EMAP-Forests will be discussed, along with
methods to evaluate that structure.
4.5.1 Tier 1
For the purposes of EMAP-Forests, the
Tier 1 resource is the forest resource within
the 40 km2 landscape characterization hexagon
of the overall EMAP program. These hexagons
represent a probability sample from the larger
hexagons, with a sampling fraction of 1 in 16.
In the current EMAP-Forests design concept,
the entire forested area of each 40 km2
hexagon is considered to be the potential Tier
1 resource, although this could be modified as
a result of pilot studies. The most efficient
landscape scale for EMAP investigation is not
known at this time. EMAP-Forests personnel
will evaluate whether the scale to be used for
EMAP-Forests Tier 1 should be at the scale of
the field plot, the scale of the watershed
containing the field plot, or the scale of the 40
km2 hexagon.
The importance of landscape-level and
watershed-level processes on local or regional
forest condition is not fully understood.
Furthermore, the importance may vary by scale
and by geographic region or ecoregion. There
are concerns about the capability of EMAP Tier
1 plots to measure forest conditions
appropriately at the landscape level. At the
EMAP Tier 1 level, there is a need for more
analysis on the optimal (or even minimally
acceptable) plot size, spatial sampling
intensity, temporal sampling frequency, detail
of the forest cover classification system, and
the accuracy of the cover classifications. For
EMAP-Forests, the grid density or hexagon
size may be insufficient, the temporal
variability may force the use of more
4-4
-------
temporally intensive evaluations, or the
classification system may be inadequate. The
FS has suggested a pilot study to evaluate
most of these concerns.
EMAP-Forests is funding an FS pilot
study to investigate these concerns. The scale
of the proposed pilot study is limited to areas
that can be efficiently imaged with high-
resolution aerial photography. This
establishes a plot size of approximately four
square kilometers. The FS interpreted 1985
imagery for a 1.2 percent sample of the state
of North Carolina at this scale. This process
used 441 plots, each of which is 4 km2 in size.
North Carolina was originally chosen for this
study because it encompasses diverse
physiographic regions representative of land
cover conditions in the eastern half of the
continental United States.
The planned pilot study will acquire
1991 imagery of the same 441 plots. Major
forest disturbances will be photointerpreted
using replicate images taken six years apart.
These disturbances will include forest
harvesting, regeneration, and land use
changes. The ability to detect other
disturbances such as defoliation, fire, and
windthrow will be investigated. The
association of observed disturbances with
landscape structure measurements will then
be investigated. To test the current EMAP
proposal for Tier 1 work, 1:40,000 scale aerial
photography will be used. For comparison,
much higher-resolution (1:12,000 scale) aerial
photography will be used to evaluate
classification error and classification detail of
the proposed EMAP photography.
4.5.2 Use of FIA Photo Points
Under the current EMAP-Forests design
concept, the most viable sampling frame is the
FIA photo point grid (Section 4.4.1). This
frame permits linkages with the FIA units. This
logistical linkage is crucial to the success of
the project and may outweigh all other
considerations. This sampling frame atso
provides an immediate probability sample for
selecting Tier 2 sites and allows inclusion
probabilities for the probability sample to be
determined directly. EMAP-Forests is planning
to select one Tier 2 site in each hexagon. The
FIA photo point grids are not of consistent
density across all regions. This does not
create any problems within regions, but it
means that inclusion probabilities for Tier 2
sites will be unequal when information from
different regions is pooled (see Section 6).
The current Tier 2 site selection
method is equivalent to the selection of a
single FIA photo point. The FIA photo point
grid for a region is overlaid on a landscape
characterization hexagon. This gives a direct
evaluation of the number of possible photo
points that could be selected. If the Tier 2 site
is selected at random from the photo points in
the hexagon, then the inverse of this number
becomes the inclusion probability for the
selected Tier 2 site at this sampling level.
4.5.3 Tier 2 Sampling
Tier 2 sampling consists of gathering
field measurements for indicators on selected
sampling units. Under the current EMAP-
Forests design concept, the closest FIA photo
point to the center point of the hexagon is
selected as the location for further field plot
selection. The details of site selection and
location based on the choice of FIA photo
point are discussed in Section 5.
4.5.3.1 Association Rules
The current EMAP-Forests design
requires the selection of one FIA photo point
within each hexagon. If stratification methods
are found to be appropriate (see Section
4.4.3.2), then the following discussion would
be appropriate for each stratum of interest. If
multiple photo points are selected in some or
4-5
-------
all hexagons, with or without stratification,
then the following discussion would still apply,
subject to some modifications.
Once the FIA photo point grid is laid
over a landscape characterization hexagon,
there are a fixed number of photo points
eligible for selection as Tier 2 sites. These
sites could be thought of as a list frame, and
association rules would be used to select the
Tier 2 site from the frame. One possible
association rule is to pick the point nearest
the hexagon center. The alternative is to treat
the set of eligible photo points as a list frame
and use a selection method such as Madow's
method (Madow 1949).
Under the current design concept,
there is not an appreciable difference between
these two strategies. The list frame is
equivalent to an equal-area selection method
because all photo points within one hexagon
have the same probability of selection. All
points can be associated with equal areas of
coverage because the photo grids do not have
significant curvature. Furthermore, in either
approach the full set of potential sample
points will be available for future modifications
of the design. The list approach has the
advantage that it conforms to standard
sampling techniques and is well studied.
However, the closest-neighbor approach has
the advantage of data confidentiality. Only the
site selected would appear in data bases
outside of the FIA units, thus protecting the
confidentiality of the other FIA sites. This
feature currently makes the nearest-neighbor
approach the preferred method.
4.5.3.2 Stratification Options
Under the current design concept, the
resource is not pre-stratified. Appropriate
stratification methods for the effective
evaluation of condition and trends in forests
have been discussed, but no clear choice of
stratification method has emerged. It may be
that stratification by current resources would
provide non-optimal sample designs for future
study of the dynamic forest resource (i.e.,
sufficient historical data would be unavailable
to perform future stratifications at the
different levels).
EMAP-Forests is funding the FS to do
a retroactive simulation study to address
some of the questions of stratification and
sampling intensity. In this study, alternative
sampling designs would be appropriately
applied to select stratified subsamples of
existing FIA plots, simulating the selection of
FIA phase 1 photointerpreted plots for EMAP-
Forests plots. This study could also simulate
the use of only EMAP landscape characteri-
zation data for sample frame development,
allowing evaluation of alternative sample
frame approaches. Though complete results
would not be available until early 1993,
sufficient work would be done by fall of 1991
for incorporation in the national plan for the
FHM program.
There are several approaches to pre-
stratification that this study would examine.
One approach is the current approach in
EMAP-Forests that no pre-stratification of
resources will be done, with detailed Tier 2
measurements made only on forested plots.
Alternatives include the pre-stratification of the
sample to enhance the design efficiency,
allowing the possibility for different sample
sizes in different strata. Any pre-stratification
from EMAP or FIA data would be done using
imperfect, remotely-sensed classifications.
Another alternative would be pre-stratification
by landform, which might be advantageous in
the western FIA regions. Landform
information could be taken from available
topographic maps, but there would be
classification error due to the map resolution
as well as problems with GIS registration of
the map information with the EMAP and FIA
grids.
4-6
-------
Stratification could be done based on
current nonforest/forest status. This is the
only stratification currently possible in many
regions using FIA Phase 1 photointerpreted
plots and may be the only stratification
feasible for EMAP landscape characterization.
Stratification could be done by current
nonforest/forest status and evaluation of
hardwood/conifer/mixed forest. This should be
feasible if EMAP landscape characterization is
done on the Tier 1 resource prior to Tier 2 site
selection. Stratification could be done by
current nonforest/forest status, along with the
current 21 Society of American Foresters (SAF)
forest types. An underlying assumption is that
EMAP-Landscape Characterization would be
able to provide more detailed forest
delineations than are typically extractable from
high-elevation aerial photography. Stratifi-
cation could also be done using Omernik's
ecoregions (Omernik 1987) or ecoregions
specifically developed for forest resources.
Another facet of this study is the
potential evaluation of minimum-travel-time
designs, in which clusters of Tier 2 plots
would be selected within a subsample of Tier
1 hexagons. Different levels of classification
accuracy would be evaluated in concert with
the various stratification methods. Since this
approach would involve major changes in
statistical design, significant statistical
advantages would have to result for it to be
used.
4.5.3.3 Schedule of Sampling
The current design concept for EMAP-
Forests is the exclusive use of the
interpenetrating design, with a cycle time of
four years (see Section 2). The grid would be
divided into four, disjoint, systematic samples
in which all sites would be completely
measured in one field season. At the end of
four field seasons, the complete set of plots
would have been visited exactly once, with
complete data collection at that time.
This approach is being considered for
the FS FHM program. In 1990, the New
England states began monitoring visual
symptoms using the EMAP grid design.
However, these states plan on annual
revisitation of sites with an altered sampling
schedule. Under this plan FIA data would be
collected at each site every year, and each
year one indicator's measurements would be
collected across all sites. Scientific reasons
exist for checking the new growth and new
foliage every growing season. This has
resulted in a different approach to sampling
and measurement in EMAP-Forests. There are
concerns about the efficiency of this design
relative to the EMAP design. It is important
that this be evaluated as soon as possible, so
that the problems of sampling schedule can
be resolved for national implementation.
This annual remeasurement plan has
several disadvantages. Since the indicators
are collected in different years, it may not be
feasible to integrate indicator data due to
seasonal variability from climatic and
meteorological conditions. Revisiting sites
every year increases the anthropogenic
damage to the site, increases the likelihood
that the confidentiality of the site could be
compromised, and may increase the difficulty
in getting land owner permission to visit the
site. In addition, simulations on artificial data
done by EMAP and Oregon State University
suggest that it would be less efficient to visit
all sites every year than to visit the same
number of different sites every year for a cycle
of four years, using the interpenetrating
design.
The FS retroactive simulation study
(see Section 4.5.3.2) will also address this
concern. One component of the study will
examine various schedules of spatial
remeasurement. One option is the annual
remeasurement of all plots, with annual
rotation of the measurement of some
indicators. Another option is the
4-7
-------
interpenetrating design with cycles of 4, 7, 9,
and 12 years. Some indicators may not need
to be remeasured every four years due to
extremely low temporal variability. A third
option to be evaluated is the use of the
interpenetrating design with a mixture of time
intervals. This option would incorporate
several remeasurement schedules, including
measuring some plots annually or remeasuring
some indicators less frequently. A fourth
option suggested for evaluation is the cyclic
remeasurement of all plots within contiguous
regions at specific time intervals (e.g., every
four years).
The question of annual remeasurement
on a fraction of the sites needs to be
addressed in more detail. Simulations on
artificial data done by EMAP and Oregon State
University suggest that the inclusion of annual
remeasurement sites in the interpenetrating
design may improve trend detection during the
first one to three cycles, after which the
improvement becomes negligible. These
simulations suggested that annual
remeasurement of approximately ten percent
of the plots for the first couple of
interpenetrating cycles would be a good
choice. This could be achieved by selecting
one ninth or one twelfth of the plots for annual
remeasurement for the first two or three
cycles of the survey. If one twelfth of the
sites were remeasured annually, then 31.25
percent of the sites would be measured each
year, instead of 25 percent. If it is decided to
make these annual remeasurement sites a
permanent component of the design, then we
may try to develop a procedure for periodic
partial replacement of these plots, so that they
do not become overly impacted by repeated
visits.
In addition, a joint simulation study
between EMAP-Forests and the New England
Forest Health Monitoring program is being
designed to resolve the issue of annual
remeasurement of all sites versus the use of
the interpenetrating design. If this study is
done, it would use the visual symptoms data
collected during the 1990 field season in New
England. These data would be decomposed
into the four systematic grids of the
interpenetrating design, and temporal
variability and measurement error would be
assessed from external data sets. Then
simulation studies would be performed to
determine the relative efficiency of monitoring
only a quarter of the sites each year. Various
statistical estimators would be evaluated as
part of this study, including the estimators
specifically discussed in Section 6. Since a
major component of site measurement cost
for the FIA units is the cost associated with
visiting the site, a reasonable relative
efficiency associated with the interpenetrating
design might be acceptable. An alternative
study proposed by EMAP-Statistics and
Design would compare these schedules given
equal effort, so that the trade-off between
these schedules could be assessed.
4.6 HIGHER GRID DENSITIES
Increasing the grid density is the
method that EMAP-Statistics and Design
recommends for increasing the sample size for
any subpopulation. For specific subpopula-
tions, it appears that grid densities higher than
the standard EMAP grid will be needed. For
example, high-elevation spruce/fir forest
occurs only in specific elevation contours.
Based on the standard EMAP grid, it is
expected that less than five high-elevation
spruce/fir forest sites would be obtained in the
Tier 2 sample. Another special case applicable
to EMAP-Forests is the special interests of
particular states. For the 1990 field season for
the FS-FHM program in New England, Rhode
Island asked for a more dense grid to be
overlaid on their state so that they could
obtain a sufficient number of sites for
statistical evaluations. The standard Tier 2
sample will not address anything less than a
resource class.
4-8
-------
In each case, it will be necessary to
use an iterative process to select such a
sample. "The first step will be to determine the
desired Tier 2 sample size. It is then
straightforward to determine if the standard
EMAP grid will produce sufficient sites where
the sample size is a monotonic function of
indicator variability. If it does not, then the
density of the grid can be iteratively increased
until the desired sample size is obtained. The
EMAP design is flexible enough to meet this
need.
-------
5 STRATEGY FOR FIELD SAMPLING DESIGN
5.1 INTRODUCTION
Field sampling refers to the collection
of Tier 2 measurements that will be used in
the calculation of the indicators (see Section
3). This section discusses the plot design
approach used in the 1990 pilot studies and
the analyses to be used in assessing that plot
design.
5.2 PLOT SELECTION RULES
A crucial step in calculating the indi-
cator information is selecting field plots to
ensure that the resulting data represent a
probability sample. The Tier 2 site selection
process using the current design concept was
discussed in Section 4.
5.2.1 Connection with Current FIA Plots
In the current design concept, a field
plot is chosen using an association rule to
select a Forest Inventory and Assessment
(FIA) photo point within the 40 km2 landscape
characterization hexagon. After the photo
point is selected, the decision of which field
plot to use must be made.
Current FIA plots present one
possibility. The FHM plot would be overlaid on
the site of the FIA plot. The other alternative
is the creation of a completely new FHM plot.
In the 1990 New England field season, the field
plots were laid on top of existing FIA plots,
and these plots were considered by the FIA to
no longer be FIA plots. There are advantages
and disadvantages to each alternative.
The advantage of using a subset of the
FIA plots as the Tier 2 sample is the direct
linkage with the FIA ground plot system. One
disadvantage, however, is that the density of
FIA plots varies greatly by state and region, a
problem that could result in a more
complicated set of inclusion probabilities for
the plots and more complicated analyses of
the data (see Section 6). Another drawback of
this plan is the potential loss of FIA plots from
the FIA sample when those plots are used for
FHM sampling. The FIA might not want to
include plots that are undergoing any
destructive sampling, a factor that could lead
to biased estimates of some variables.
FIA plots were used in the Northeast
region during the 1990 field season. However,
because of the Northeast sampling design,
this is the only region that can periodically
replace plots. Other FIA projects do not want
to use FIA plots for FHM sites.
The advantages of creating completely
new FHM plots are twofold. This scenario
would simplify analysis by providing an
unbiased probability sample with equal
inclusion probabilities within each stratum in a
region. It would also provide an opportunity
to correlate FHM data with other data
collected at the same location. One
disadvantage would be problems of statistical
linkage to the existing FIA plot system, a
system that presents an enormous source of
potentially useful historical data. This problem
could be partially overcome in several ways.
The FIA photo points would allow links to FIA
areal estimates. It should be possible to use
FIA definitions of subpopulations to provide
links between the FIA and FHM statistical
frameworks. Composite estimators could also
be used to combine information from the
independent samples statistically. A standard
method in forest sampling involves combining
the estimators linearly using the inverses of
the respective variance estimates as weights
(Ray Czaplewski, personal communication).
The current recommendation from the
FIA units across the nation is to use the
existing FIA grid of photo points (see Section
5-1
-------
4.5.2), but riot to use the actual FIA plots in
the future. One photo point per hexagon
would be chosen and if there is already an FIA
plot in place, the ground plot would then be
offset from the FIA plot by a fixed bearing and
distance. The FHM plot could also be located
at a random bearing or a random distance
from the photo point. However, this would not
be necessary to achieve the desired probability
sample. Careful protocols are being
established to ensure that field crews select
sites properly.
In areas where FIA photo point grids
do not exist, the FIA has shown a willingness
to extend the systematic photo point grids to
cover all forested lands. The details of this
have not been established (see Section 1).
5.2.2 Plot Selection Protocols
Inevitably, there will be problems with
the sampling of plots. Plots may be
inaccessible or unsafe to sample, lost, or
destroyed. Criteria for plot selection must be
designed carefully so that a probability sample
is maintained.
Plots that are inaccessible for safety
reasons (i.e., excessively steep slopes) could
be relegated to a stratum of unsampled sites
about which no inferences can be drawn, as
currently suggested by the Forest Service (FS).
Alternatively, these plots could be treated as
missing data in their original stratum and left
as missing data permanently. Plots which are
inaccessible due to landowner denial of
access could remain in the monitoring system
and be marked as missing data. Permission
to gain access would be sought at each
scheduled measurement time. Rather than
using this strategy, the Northeast region in the
1990 pilot chose to select alternate plots when
denied access by landowners. Such plots
must be flagged in the data base and the
inclusion probabilities for these plots must be
altered.
Lost plots are plots that cannot be
relocated from the ground. The FS suggests
reestablishment of these plots in their correct
locations as determined by photo points. The
disadvantage of this procedure is that
historical data from that plot might no longer
be relevant. The alternative would be to
declare that all the data for that plot are
missing and try to relocate the plot in the
future. This must be decided before a
measurement cycle is completed and there is
an opportunity to lose an established plot.
Established plots may be destroyed in
various ways. They may be clearcut, in which
case the FS defines them as non-forested
plots within the original forest type. In this
case, the plot would remain in the monitoring
system and be monitored on the same
schedule. Plots that are converted to
agricultural land would be flagged in the data
base as no longer forested land. These plots
would be followed until they return to forested
land and are monitored on the previous
schedule.
5.2.3 Boundary Case Protocol
Some FIA projects "rotate" points in
their field plots. In other words, if the plots
straddle two or more distinct forest types, the
plot is reconfigured so that the subplots fall
into the same forest type as point number 1 of
the plot. Point number 1 could be the first
sampled point in a 10-point cluster, or it could
be the center point of a fixed-area plot. Plots
that straddle two or more distinct forest types
are a concern to the FIA. It has been argued
that when such overlap is permitted,
unrealistic forest type combinations that do
not actually exist are "created".
Points are rotated to facilitate descrip-
tion and simplify analysis; however, the
introduction of bias is a possible result. Some
FIA projects have adopted specific rotation
techniques; other FIA projects do not rotate
5-2
-------
points. The F1A recognizes that point rotation
may bias volume estimation, but it also
recognizes that not rotating points may yield
biased estimates of area and volume by forest
type. Furthermore, the decision to shift points
is made by the field crew; therefore, it remains
subjective.
A primary concern of FHM is that the
bias introduced into the plot selection
procedure will invalidate the monitoring
network design. The current consensus is the
FHM plots should not be rotated, and methods
for sampling multiple strata or forest types
within a single plot must be developed. An
alternative approach is to use subplot
information (with appropriately adjusted
inclusion probabilities) in estimation
procedures. This would allow the use of non-
rotated plots that span different forest types
without unduly complicating post-stratification
or reporting procedures.
5.3 PLOT DESIGN IN THE 1990 PILOT
STUDIES
In the 1990 pilot studies in the
Northeast and Southeast, the plots, subplots,
and measurement locations were configured
with specific objectives not necessarily
germane to plot design for full implementation
of FHM. One objective was assessment of
variability components for different indicators.
This will not be needed in full implementation.
Therefore, the plot design used in the 1990
pilot studies is not necessarily the optimal plot
design for future studies.
5.3.1 Plot Geometry
The field plots used in the 1990 pilot
studies were fixed-radius plots (i.e. each plot
described a circle about the plot center). The
FIA projects also have extensive experience
with alternative plot designs such as variable
-radius plots (i.e., plots in which trees are
selected with probabilities proportional to their
basal area) and the 10-point cluster (Hazard
and Law 1989). Due to the many quantities to
be measured, there is ongoing discussion
about the plot geometry. As discussed in
Section 4, a single, contiguous extent of forest
ecosystem was desired as an experimental
unit so that a single fixed-radius plot
containing a four-point cluster worked well
(see Figure 5.1).
Most scientists generally agree that
variable-radius plots are more efficient for
measuring current status based on tree
characteristics such as basal area and
volume, but that fixed area plots are easier to
use for measuring change over time.
Furthermore, variable-radius plots do not have
an advantage when assessing current status
of indicators not directly related to tree size.
For example, the vertical vegetation structure
measurement used in the 1990 pilots, although
related to forest stand size and density is not
more efficient when using variable-radius
rather than fixed-radius plots. The use of
fixed-radius plots is easily defended on the
basis of the variety of measurements to be
taken. It is harder to establish fixed-radius
plots, but it is claimed that they are easier to
remeasure than variable-radius plots. In
addition, fixed-radius plots should be easier to
use in assessing changes and trends over
time because the plot delineations will remain
stable.
Thus the fixed-radius plot was selected
for the 1990 pilots. Although there are
advantages to having a single plot design for
the entire FHM program, no plot design can be
optimal for all criteria simultaneously.
However, at a FS conference on FHM design,
it was decided that the current plot design
should be used for the national FHM program,
with special exceptions only when the design
can be shown to be inadequate for particular
cases (i.e., sequoia forests).
5-3
-------
(* I/24-acre each
L I /6-acre total
Area - ' nectare
I 20
Distance between points is -20 f:
Azimuth : 2 = 360
Azimuth ' -3 = 1 20
Azimuth ' -4 = 240
Figure 5.1. Plot design for the FY90 field season.
5.3.2 Plot Size
The size of the experimental unit in the
1990 pilots was originally defined as one acre.
The one acre area was chosen because it has
been the traditional size used by the FIA units
in the past. An important criterion for the 1990
pilots was a plot protocol with which the field
crews would be familiar so that protocol de-
velopment and training requirements would be
minimized.
The FIA defines the minimum area for
classification of forest land as one acre (0.40
hectares). Since the pilot project was
originally designed for the Northeast FIA unit,
the one-acre plot size was considered
reasonable. At the time, there was no way of
assessing whether one acre was efficient for
all indicators. In the actual implementation,
the circle covering the subplots extended over
approximately 2.5 acres (1 hectare).
5-4
-------
5.3.3 Number and Size of Subplots
The design used in the 1990 pilot
consisted of four 1/24m-acre circular subplots
within the experimental unit - one in the center
of the plot and the other three arranged to the
north, southwest, and southeast at 120-degree
angles from one another (see Figure 5.1).
Each subplot had a radius of 24 feet.
Nondestructive measurements were to be
made on the subplots, and all destructive
measurements were to be made off the
subplots. This subplot arrangement for the
pilot study was acceptable to FHM for
assessing plot and subplot variability.
5.3.4 Areas for Destructive Measurements
In the pilot study, destructive measure-
ments on trees were performed in a 12-foot
ring around each subplot (see Figure 5.1). It
has been suggested that all destructive
measurements be done over one crown width
from the subplots. This was not done in the
1990 pilot studies. Some FIA units felt that
one crown width distance might not be
enough to protect the plots from potential
confounding effects of increased pest activity
associated with destructively sampled trees.
Implementation of this rule may necessitate
the use of larger plots. Two branches were
taken from each sampled tree, and multiple
trees were sampled on each subplot. This
was done to assess components of variability
in foliar nutrients and contaminants rather
than as an official method for future foliar
sampling.
Destructive soil sampling was done at
three points. Each point was midway between
the center of the center subplot and the center
of one of the other subplots. Since three pits
were considered affordable and reasonable
based on prior analyses by the soil indicator
group, data from the pits were combined to
provide data relevant to all the subplot
information. In addition to the typical plots,
special plots were established with these
three pits and nine additional pits that were
located in three triangles in the destructive
sampling zones of the exterior subplots. This
design was selected to provide information on
soil spatial variability across a typical field
plot. It is not expected that twelve pits on a
plot will provide an optimal allocation of
resources. Analysis from the pilot data will
help determine reasonable choices for the
number of samples and the locations of the
soil sampling pits.
5.3.5 Linkages Between Indicators
The subplots and destructive sampling
zones were laid out so that linkages between
indicators might be established. For
convenience and investigation of relationships
between the measurements, the vertical
vegetation structure and photosynthetic active
radiation (PAR) measurements were done at
the same 16 points on each subplot. The tree
samples were taken close to the subplots so
that foliar data and destructive visual
symptoms data could be related to non-
destructive measurements on the subplots.
The soil samples were taken to provide plot-
level average soil chemistry on typical plots.
Special plots provided extra information about
soil chemistry within the destructive sampling
zones for potential relationships with foliar
chemistry.
5.4 PLOT DESIGN DEVELOPMENT
Now that the 1990 field season is
complete, it is essential to analyze the data
from the pilot studies and evaluate the
lessons learned. The primary purposes for the
1990 pilot were logistics studies and
assessment of variation components. With
cost and time estimates from the pilots, it is
feasible to begin assessing optimal ways to
sample specific indicators. Many of the
samples will not be analyzed by analytical
laboratories until March. Thus, information
5-5
-------
from these data could not be used in this
document, but may still be available in time to
modify FY91 field activities.
5.4.1 Assessment of Variability Components
For any variable of interest measured
on a forest plot, variability in population
estimates may arise from a variety of factors.
These factors will introduce uncertainty in the
estimation of all the population statistics
including means, totals, medians, and
quantiles. The sources of variability include:
• Real differences in plot-level means across
the region.
• Spatial variation between measured values
across plots.
• Spatial variation within plots or within
subplots.
• Temporal variation between years.
¦ Temporal variation over the sampling
period within the field season.
• System measurement error attributable to
the total variation in all the facets of
sample extraction, collection, handling,
preparation, and analysis.
In the 1990 pilot studies, most
indicator variables were sampled so that
between-plot variability, between-subplot
variability, within subplot variability, and
system measurement error could be
estimated. Measurement error will be
assessed using the quality assurance (QA)
remeasurements that were performed as a
part of the QA program for the pilots (see
Section 8). The other components are
determined using analysis of variance
techniques.
The standard nested analysis of
variance based on plots and subplots within
plots will provide three mean squares
(Cochran 1977). Let the mean square for plots
be s,2, the mean square for subplots within
plots be s22, and the mean square for
observations within subplots be s32. Then one
can use the ejected mean squares to obtain
estimates of the variance components using
the method of moments. Furthermore,
measurement error can be incorporated and
removed from the calculation at the same
time.
If n is the number of plots, m is the
number of subplots, k is the number of
observations within each subplot, f2 is the
sampling fraction for subplots, f3 is the
sampling fraction for observations within each
subplot, and sm2 is the externally estimated
measurement error, then the variance
component estimates can be calculated as:
S,2 - s,2 - (1 - fj sa2 /m - (1 - fj s32 /(km),
S22 = s22 - (1 - fj s,2 /m, and
s 2 = s 2 - s 2
•53 - 5j - Sm
5.4.2 Cost Versus Efficiency
The above variance component esti-
mates can be placed into formulas from
Cochran (1977) to yield optimal sampling
strategy to balance cost and efficiency. Note
that this does not address developing
sufficient precision to detect specific trends
with stated confidence. This only allows one
to assess the optimal arrangement of
resources within a plot for a specific cost or a
given limit on the size of the variance of the
grand mean. The application of Cochran's
formulas in the EMAP context requires that
inclusion probabilities be approximately equal
(see Section 6).
5.4.3 Trend Detection
One of the primary goals of EMAP is to
detect trends of ecologically significant size in
a specified number of years; therefore, it is
important to determine as soon as possible
whether or not a specific indicator will be able
5-6
-------
to meet its data objective. The assessment of
this criterion has two requirements. The
components of variability must be known well
enough to estimate the performance of the
indicator in detecting a trend, and the size of
the ecologically relevant trend must be
specified so the statistician can determine if
this trend can be detected.
The first requirement will be assessed
for several of the indicators as described
above so that variability components can be
used to evaluate performance. Some indicator
groups are developing variance estimates
through historical data, rather than going
through the expense and time of a pilot study.
Some indicators may have problems with
remeasurement or temporal variability that
need to be included in the assessment. Some
indicator groups are evaluating these points to
determine what variance component estimates
are most important to determine in the FY91
field season.
The second requirement is also in
development. Indicator groups are evaluating
their indicators and reviewing their previous
work in indicator development to determine
ecologically significant trends against which
their indicators can be compared.
By the start of the FY91 season, most
of the first requirement and all of the second
should be completed. By fall of 1991, all of the
variance components for the indicators should
either be estimated or designated for future
study. This will permit the full evaluation of
current and future indicators in the EMAP
context.
5.4.4 Evaluation of Subplot Size
The evaluation of optimal numbers of
plots, subplots, and observations within
subplots using the modifications of formulas
from Cochran (1977) generates a certain
number of subplots as optimal. However,
since the subplots can be altered in size, plots
could be enlarged instead of altering the plot
geometry and adding more subplots. This
possibility is relevant because some indicator
groups are concerned that the subplots are
too small. Also, these numbers must be
evaluated carefully since multiple indicators
are being studied, and the formulas are
designed to generate optimal numbers for
computing mean values.
The vertical vegetation structure and
PAR indicator groups have expressed concerns
that they need to cover a larger area to
capture the spatial heterogeneity of the
sampling unit and to achieve spatial stability.
Since both indicators were sampled on a grid
in each subplot, it is anticipated that spatial
heterogeneity will be examined as part of this
year's analyses. This may allow FHM planners
to determine whether or not larger study areas
are necessary for these indicators.
Furthermore, the Tier 1 landscape
characterization analyses may also allow
examination of this question. One important
consideration is that Tier 2 sampling not
measure anything that can be ascertained at
the Tier 1 level for less money. If it turns out
that the spatial scales of interest for these
indicators are large enough to analyze at Tier
1, then these indicators need not be concerned
with collecting that component of the indicator
at the field plot level. Instead, it can be
determined using remote-sensing and GIS
techniques during Tier 1 landscape
characterization.
5-7
-------
6 STRATEGY FOR STATISTICAL
ESTIMATION AND ANALYSIS
6.1 INTRODUCTION
The strategy for the development of
the statistical structure for data collection, as
well as the strategy for indicator development,
has been discussed in the preceding sections.
Measurements taken in the field or analyzed in
a laboratory can be translated into indicators
of aspects of forest condition. But to assess
the current status and extent and the observed
changes or trends in the data properly, one
must be able to estimate these indicators with
known confidence. This section will discuss
statistical procedures envisioned for these
analyses.
The Forest Health Monitoring (FHM)
staff will not conduct all research, either
ongoing or proposed, for the following topics.
EMAP has a cross-cutting program in
statistics and design, and the staff of the
EMAP-Statistics and Design Coordination
group is taking the lead in researching and
addressing many of these topics.
Furthermore, EMAP-Statistics and Design has
cooperative agreements with university
statisticians who are developing key pieces of
this work. Other resource groups in EMAP and
the USFS Rocky Mountain Forest Experiment
Station are also working on these problems.
Areas in which FHM will concentrate have
been discussed in Sections 4 and 5. The
incorporation of biological models is
discussed in Section 7.
6.2 STATUS AND EXTENT
Graphical displays and descriptive
statistics will be used to represent status and
extent of current resources. It has been
shown that GIS maps displaying extent and
spatial pattern can be created using the kind
of data that will be collected (Church et al.
1989). Estimates of proportions of the
population occurring in various categories will
utilize the cumulative distribution function
(cdf). The cdf is an important tool in the
examination of regional data (Linthurst et al.
1986; Church et al. 1989) and will be used by
FHM.
In addition, parametric (model-based)
estimation techniques will be evaluated to
determine if alternative approaches might be
useful. Spatial behavior on a regional scale
needs to be addressed, and approaches to
this problem will be studied. Special analyses
may be appropriate for specific subpopula-
tions. Methods must be developed to deal
with measurement error, deconvolution of
extraneous variability, and response error.
An important point to address is
overlap with current estimates of extent
performed by the Forest Inventory and
Analysis (FIA) program. The FIA uses
probability-based sampling methods with
appropriate sampling theory estimators.
Furthermore, the FIA has a more dense
network of sample sites than EMAP-Forests.
Hence, it is unnecessary to duplicate FIA work
and make estimates of the same variables
because the FIA estimates will have better
precision. However, FHM will need to collect
the same information as the FIA does to
integrate more effectively with the FIA and to
calculate any indicators that require some of
the FIA's measurements. For example, FHM
may use composite estimators to combine the
two estimates.
6.2.1 Sampling Theory Estimators
The cdf for a set of univariate data will
be generated by using the Horvitz-Thompson
(HT) estimation formulas (Horvitz and
Thompson 1952). Descriptive statistics (e.g.,
totals, means, medians, standard deviations,
and quantiles) will also be generated by using
these formulas. The HT formulas allow
estimation of descriptive statistics from any
6-1
-------
statistical design that is probability-based.
The only requirement for estimation is the
specification of the inclusion probabilities for
the sample data. These inclusion probabilities
are obtainable for any probability sample on a
statistical framework. The EMAP statistical
designs will be restricted to probability-based
designs so that the HT formulas can always
be used.
In the HT estimation of a total and the
variance of the total, two kinds of inclusion
probabilities are needed. First and second
order inclusion probabilities must be
generated. First order inclusion probabilities
are the probabilities with which the individual
sampling units are included in the sample.
These first order inclusion probabilities must
be known for each sampling unit included in
the actual sample, and these should be
generated at the time of sample selection and
archived into the data bases (see Section 10).
Since these inclusion probabilities will be
needed for all analyses that use estimators
from sampling theory, they must be available,
along with the sampling-unit-level data. These
first order inclusion probabilities will be
designated by the symbol tt,, referring to the
inclusion probability for the im sampling unit.
Second order inclusion probabilities are
pairwise inclusion probabilities. They are the
probabilities with which two different sampling
units are simultaneously included in the same
sample. These are typically denoted as it,,,
referring to the probability of including both
sampling units i and j in the sample. The
design features specific to the ecosystem are
needed to determine the rr,,. The design
features include elements such as sample size
and information on the specific strata or
clusters in which sampling units i and j fall.
That information must be carried along with
the data so that use of the data is
uncomplicated. The statisticians will work
with the information management staff to
decide how best to store these data for
optimal data utility and storage.
Estimation formulas are simplified by
the use of weights rather than inclusion
probabilities (Overton 1987), using the
notations:
w,=1 III, and
W|-1 / II|
In practice it has been found to be more
convenient to store the weights with the data
rather than the inclusion probabilities. Using
weights instead of inclusion probabilities, the
HT formulas may be written as:
V^vw,
V (ty) - E,yfw, (wr1) ~ EE^yj (wiwrwy)
where y is any measured characteristic,
and fy is the true total of that characteristic
over the population or any specified
subpopulation. Estimates over a specified
subpoputation are generated by restricting the
above sums to the set of sampling units that
represent that subpopulation.
The total is estimated as above. If the
mean is desired instead, it can be estimated
by dividing Ty by N, the number of units (or
total areal extent, depending on the variable) in
the population or subpopulation. If N is
unknown, it may be estimated by setting y, =
1 in the above equations to obtain:
ft = E,w,
6-2
-------
-------
1.0
0.8
0.6
0.4
0.2
0.0
•100.0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 1,000.0
G(X)
0.2-
i i i I i I i i r
•100.0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 900.0 1,000.0
Figure 6.1. Example of estimated distribution plots with upper confidence bounds, generated both
as numbers (upper plot) and area (lower plot) (Linthurst et al. 1986).
6.2.3 Model-Based Estimation Techniques
Alternatives to the HT estimators
should be studied to ascertain their relative
efficiency and power. Methods using auxiliary
data to strengthen the estimate of status or
extent might prove more effective in some
cases. Specifically, if there is not a detectible
trend in the variable, it may be feasible to
utilize past data from the same sites to
develop a more powerful estimator of status.
For example, the James-Stein estimator uses
related auxiliary data to produce an estimate
with improved precision. The FS has started
looking at the James-Stein estimator with
regard to certain F1A data, and FHM personnel
will work with the FS to examine the utility of
such an estimator in the EMAP setting. A
possibility that has been suggested for trend
detection is an estimator based on the
exponentially-weighted moving average.
6-4
-------
6.2.4 Spatial Patterns
Data collected in the FHM program will
be associated with a particular spatial
location. Although the confidentiality of the
data is crucial, general information about the
location of the plots can be used in analyses
without compromising that confidentiality. In
particular, under the current design concept of
one plot per Tier 1 characterization hexagon,
the hexagon identification number would
provide sufficient information for spatial
pattern analysis without jeopardizing the
confidentiality of the site location. Given that
the plot is located within a 40 km2 hexagon,
stand dynamics and topographical data may
also be usable without compromising the plot
location.
Another consideration is the scale of
the phenomena to be measured. This cannot
be determined by the statistician but must be
determined by the scientists studying the
forest ecosystems. Determination of key
spatial scale features for monitoring the
condition of forest ecosystems would then
provide valuable guidance to the developers of
the design. After the design is fixed, spatial
characterizations may be limited by features of
the design. Choice of density and size of the
Tier 1 landscape characterization hexagons
should be tied to the scale of the phenomena
that are to be measured.
Visual presentation is perhaps the
most important method of examining spatial
patterns. More sophisticated methods cannot
be easily applied without using some
exploratory graphical methods first. A number
of cartographic techniques are available for
the graphical presentation of data. Spatially
continuous or nearly continuous data (e.g.,
synoptic data) can be represented either as a
contour surface in three dimensions or drawn
using contour lines in two dimensions.
Shading, color, and isopleths are all popular
ways of representing a third dimension on a
two-dimensional printout. Another alternative
is the use of a symbol or color to illustrate the
magnitude of a variable at a specific point on
a map.
Visual displays encounter two
problems. First, these techniques do not
usually take into account the differing inclusion
probabilities that different data points may
have. The FHM data represented in this way
must use techniques that will display the data
using the population weights. Second, visual
presentations may require a smoothing
procedure to obscure the sampling unit
locations and protect the confidentiality of the
data. This is not an issue if the scale of the
map is coarse enough.
There are other ways of examining
spatial pattern. Ageostatistical approach that
employs some version of kriging (Ripley 1981)
may be useful when examining regional
gradients or changes in regional gradients.
Adaptive splines (Wahba 1990) or random field
theory (Ripley 1981) may be useful alternatives
to kriging.
6.2.5 Subpopulation Analyses
Another method of examining spatial
patterns is through subpopulation analyses.
The EMAP design allows for inference on any
subset of data that can be described through
the attributes of the sample or population.
Inferences on any subpopulation are obtained
by applying the HT estimation formulas to the
appropriate subsample of the data. The same
analyses can be generated for a
subpopulation as for the population as a
whole. Subpopulation analyses can also be
used to examine spatial pattern. Regional
differences can be expressed by subsetting
the population into regions and examining the
population distributions of the selected
subsets. EMAP-Statistics and Design is
working on several ways to analyze and
6-5
-------
compare distributions. A number of these
methods use the cdf.
6.2.6 Deconvolution
If the measurements used to calculate
an indicator have appreciable variability due to
spatial or temporal variation, then the
estimator of the cdf may have bias in the tails
(Overton 1987). Furthermore, other descriptive
statistics may have large bias. The apparent
effect is to increase the system variability.
The cdf of the observed data is not the
desired cdf but the convolution of the
population distribution and the distribution of
the extraneous source of variation.
The FHM personnel must recognize
extraneous variation, identify this component
of variability, and account for it either in the
sampling design or in the analyses. In some
cases it can be accounted for in confidence
limits. 'Deconvolution" is defined as the
process of actually removing the excess
variation from the data (Church et al. 1989).
This process is undertaken to eliminate bias in
the cdf and to correct the statistics in other
analyses. Methodology for deconvolution in
the general case is under study by EMAP-
Statistics and Design.
Measurement error, which is the
accumulation of error in the data collection
and analysis process, can also produce biases
and variance inflation. Proper quality
assurance (QA) can lower measurement error
and quantify the level of existing measurement
error (see Section 8). Given appropriate QA,
the size or relative size of the measurement
error can be estimated and incorporated into
statistical analyses by using measurement
error models.
Another important source of
extraneous error is sampling error. In survey
sampling, sampling error can be described as
a general term for errors in the planning,
collection, and processing of data. The
problems that are germane to a sampling
design such as EMAP are specification error,
coverage error, response error, non-response
error, and processing error. For the purposes
of the FHM program, the issues of response
error and processing error fall under the QA
issues of instrument error or analytical error.
Specification error and coverage error
should be handled before crews go out into
the field. Specification error occurs at the
planning stage when user needs change or are
misinterpreted, the populations of interest are
not specified correctly, or the concepts of the
program are ill-defined. Coverage error occurs
when some units of the population are
erroneously omitted, or inappropriate units are
inadvertently included. These errors must be
eliminated at the planning stages of the
project.
Non-response error occurs when some
selected sites cannot be sampled (see Section
5.1). If sites that are not sampled differ in
some measured way from sites that are
sampled, then a bias will be introduced into
the data in that measurement. Some of the
problems with non-response error can be
eliminated by working closely with the FIA.
The FIA has worked with private landowners
for many years and has built up a relationship
of trust that will facilitate site visits. If
appreciable non-response error is found or
even suspected, sample survey techniques can
be called into play to evaluate the size and
effect of the non-response bias.
6.3 CHANGE AND TREND
A number of standard statistical
techniques are available for the study of
change and trend. Methods are also being
developed by EMAP-Statistics and Design.
Linear models methodology, such as analysis
of variance, analysis of covariance, and
multiple regression will be used. In order to
6-6
-------
use these, various assumptions will be
formulated and tested regarding spatial and
temporal variance components, statistical
independence, explanatory variables, the
nature of trends, and the pattern of changes.
Nonparametric alternatives such as methods
that are based on signs of differences or
ranks of the data will also be evaluated (Loftis
et al. 1989).
EMAP-Statistics and Design has been
developing techniques that do not focus on a
change in central tendency, but use the cdf to
look for other types of distributional behavior.
Also, trend detection methods that are specific
to the interpenetrating design structure of
EMAP are being developed. Specific research
includes several methods that account for the
effect of temporal correlation to alter the
statistical power of trend tests.
EMAP's interpenetrating design will
achieve increased power to detect change only
after two full cycles of visits have been
completed (i.e., after all sites have been visited
twice). Repeat visits permit a paired analysis
which eliminates one of the components of
population variation (Overton et al. 1990).
Similarly, the power to detect a continuing
trend will increase as more years of data are
collected.
6.4 ASSOCIATIONS
Analysis of associations will play a key
rote in the development of the FHM program
and is crucial to the EMAP objectives. Due to
the complicated nature of the forest
ecosystem and the varying levels of
measurement error in the measurement of
different variables, it will take time and
extensive statistical analysis to develop some
of the indicators for the program. The
traditional statistical techniques have primarily
looked at linear components of behavior.
However, recent developments in statistics
such as projection pursuit analysis (Friedman
1987) and sliced inverse regression (Li 1989)
may permit the evaluation of nonlinear
relationships as well.
In addition, the analysis of
associations in the EMAP context has two
problems that distinguish it from analysis of
associations in the standard statistical
context. Unequal inclusion probabilities can
complicate data analyses, and the
observational nature of the data puts
limitations on the inferences to be drawn from
the statistical analyses.
6.4.1 Consequences of Unequal Inclusion
Probabilities
When the data collected are
associated with different inclusion
probabilities, this must be accounted for in the
statistical analyses. This is typically done by
weighing the observations using the inverses
of the inclusion probabilities. This is one
reason why the EMAP design attempts to
eliminate variable weights, except among
resource strata.
If the weights are all the same within
an individual stratum, simple unweighted
analyses are all that are needed within that
stratum. Furthermore, if the weights within a
stratum are only approximately equal, then it
has been shown (DuMouchel and Duncan
1983) that simple unweighted analyses are
generally sufficient. The same authors also
developed a methodology for assessing when
unweighted analyses would be preferable to
weighted analyses.
A more complicated problem arises if
different strata have different functional forms
for the regression. In such a case it can be
argued that the analysis of the combined data
is meaningless. A regression equation
developed in such a situation may have
applicability to none of the actual population.
This problem of course does not arise if the
6-7
-------
strata are not to be analyzed together. On the
other hand, if there is an important reason to
combine such strata, the analysis can still be
performed. But the weights and the results of
the analysis must be carefully considered.
An alternative in such a case may be
the use of meta-analysis methodologies
(Hedges and Olkin 1985) to examine the overall
behavior of the entire set of strata for general
hypothesis evaluation. Meta-analysis
techniques have not been applied to forest
ecosystem problems as yet. However, FHM
will have extensive datasets for the evaluation
of these methods.
An important problem caused by
unequal weights is the graphical presentation
of the data. Scatterplots are a basic tool of
exploratory data analysis, but a way to make
the number of points in each of the stratum
samples proportional to the population
numbers is required. One proposed FHM
method is to plot circles and make the radius
of each circle proportional to the square root
of the weight. In this way, the area presented
to the viewer will effectively represent the
appropriate relative size of the population.
6.4.2 Consequences of Using Observational
Data
Hypothesis testing in the context of the
EMAP program also presents problems.
Schreuder and McClure (1991) discuss this in
the context of the FIA program. Data
collected will be observational data, not
experimental data. Hypothesis tests
performed on observational data are the same
as those on data from planned experiments,
but the problem is seen in the inferences that
can be drawn from observational data.
Causality is difficult to establish from
observational data due to the large number of
uncontrolled factors. Causality is proven by
formal experiments, while observational
studies can only establish associations. On
the other hand, associations seen on regional
or national scales can provide powerful
associations that might be difficult to see in
individual study sites. The associations
developed from FHM data will need to be
evaluated and then tested or validated in other
settings such as controlled experiments.
6.5 METHODS FOR INTEGRATION OF
INFORMATION
The FHM program will require data
other than those measurements collected at
the Tier 2 sites. And the FHM program can
benefit from work with other programs that
have studied related areas and ecosystems.
6.5.1 Other EMAP Resources
In addition to data collected at the Tier
2 sites, the FHM program will need information
on auxiliary data. The photointerpretation
work by EMAP-Landscape Characterization will
provide essential Tier 1 data on landscape
processes. Under the current design concept
of one Tier 2 site within each Tier 1 hexagon, it
is statistically straightforward to associate
landscape process data with the Tier 2 data.
Decisions on level and scope of landscape
characterization must be made by the
appropriate people. But whether it is decided
to use landscape information for the area
around the plot, the watershed containing the
plot, or the entire hexagon, those data can be
directly associated with Tier 2 data that have
been aggregated to the plot level.
The FHM personnel will also need data
from EMAP-Air and Deposition. Data on wet
and dry deposition, climate, and weather will
be required for many analyses. These data
will be used as essential covariates in
regressions and analysis of variance in order
to remove variation in the measurements and
indicators that is associated with climatic and
meteorological data. Problems with
6-8
-------
association in this case are related to the
measurement error of the estimates.
It is prohibitively expensive to monitor
every site for everything of interest; therefore,
data from EMAP-Air and Deposition will be
estimated from monitoring networks. The
sizes of the system measurement errors for
their variables will be crucial in determining
how to use the variables. If these data have
small measurement error, the variables will be
useful covariates. If the measurement error is
not negligible but reasonably small,
measurement error models may be required to
correct for the variability in these data. If the
measurement error is too large, these data will
not be usable in the analyses.
Data from other resource groups in
EMAP will also be used. Data on nearby
wetlands condition or the management
practices of nearby agroecosystems may
elucidate ecosystem behavior across EMAP
boundaries. Forest soil data may assist in the
quantification of wetlands or surface water
conditions. To use such data, each EMAP
resource group must develop a statistical
design that is based on the overall EMAP
design concept so that links among resource
groups may be made for statistical analyses.
Areas such as forested wetlands,
which may be classified into one of two
different ecosystems, may be of interest to
more than one resource group. In such a
case, it may be feasible to build a cooperative
effort among EMAP resource groups and
sample the features of interest to each group.
6.5.2 FIA Data
Under the current design concept, the
EMAP-Forests program will be based on a
random subset of the FIA forest inventory
network. This will allow EMAP-Forests to link
into the FIA network. Mechanisms are being
developed to extrapolate EMAP-Forests results
to the larger FIA network and to use FIA
estimates to help the EMAP-Forests analyses.
Methods for combining estimates are
mentioned in Section 5.
6.5.3 "Encountered" Data
The most difficult area is the
integration of "encountered" data, data that
are acquired without using a statistical design
(Overton 1990). Large quantities of non-
random or haphazard data are available which
will be difficult to integrate into the FHM
framework due to the lack of a statistical
framework for their collection. This problem is
being examined by EMAP-Statistics and
Design. Several possible approaches involve
subsetting or clustering the non-random
samples into compartments that would
parallel the FHM statistical framework.
6-9
-------
7 STRATEGY FOR ASSESSMENTS
7.1 INTRODUCTION TO ASSESSMENT
Assessment is a process by which
data are converted into useful information. An
assessment strategy describes how know-
ledgeable analysts will organize, synthesize,
and interpret data in order to simplify data,
test for change and differences, generate
hypotheses, determine the consequences of
observations, and evaluate the uncertainty of
conclusions (NRC 1990a). As the prime points
of contact with society, it is the assessors'
role to ascertain social needs and to translate
them into guidance for reporting. Since
assessment requires organization and
synthesis of data, assessors have a role in
linking pieces of the program together. Finally,
assessors have a role in exploring techniques
and developing knowledge that will improve
interpretations.
There are many challenges to
formulating an assessment strategy.
Environmental problems are becoming
increasingly complex and scientists have
limited understanding of them. The
assessment strategy has to address many
public, regulatory, and management concerns
about forests: atmospheric deposition, non-
point pollution, climate change, deforestation,
and biological diversity. Data and models will
be used in different ways to define and
interpret ecological condition and to evaluate
the results of programs and predictive models
(Linthurst 1990). Coordination is needed to
achieve these assessment objectives within
the multi-tiered, multi-regional, and multi-
agency framework of forest health monitoring
(FHM) and within the EPA risk assessment
framework (see Section 2).
The primary, short-term objectives of
EMAP-Forests assessments are to produce
periodic statistical summaries, interpretive
reports, and integrated assessments that
address the regional status and trends of the
nation's forests in relation to human-induced
stresses (see Section 11). The long-term FHM
assessment strategy will have to evolve to
maintain consistency with the overall EMAP
program (see Section 1). The FHM personnel
can help to determine overall, long-term goals
by taking an active role in client identification,
question definition, and evaluation of user
responses.
7.2 STATUS OF FOREST ASSESSMENTS
This section is a first attempt to
describe an assessment strategy for FHM. To
date, the general EMAP assessment strategy
(see Section 2) has been the point of
departure for design decisions. This section
reviews what has been done to implement the
general strategy in the particular case of
forests.
7.2.1 The Assessment Paradigm
Environmental concerns identified by
EMAP-Forests include sustainability, pro-
ductivity, aesthetics, diversity, extent,
utilization, contamination, and quality. These
concerns relate to the environmentalist
paradigm described in Section 2. These eight
concerns will be addressed by assessment
endpoints for the abiotic (soil, water, and air)
and biotic (vegetation, animals) components
of forest ecosystems.
A peer review of the indicator strategy
in May 1990 by the EPA Science Advisory
Board endorsed the general approach to
forest assessment, commended the progress
made, and indicated that the necessary
linkages between environmental concerns and
measurements are possible to define. Since
that time, most of EMAP-Forests' assessment
resources have been devoted to a field test of
measurement systems and to an example
statistical summary. In December 1990,
7-1
-------
attention was again shifted back to the
assessment process.
Figure 7.1 summarizes the current
status of the assessment framework for
forests, considering what is in an ecosystem
(components) and how one may view an
ecosystem (concerns). Despite apparent
holes in the framework, the progress to date
is encouraging. Traditional concerns such as
contamination, utilization, extent, and
productivity can be reliably assessed in most
situations. But concerns for sustainability,
aesthetics, diversity, and quality are not
currently amenable to reliable assessments.
7.22. Other Assessment Activities
An example statistical summary
(Riitters et al. 1990b) was prepared to
demonstrate the EMAP statistical assessment
framework in the context of forests. The
example report did not consider many
indicators that could be included in such a
report, nor did it report the possibility of
testing statistical correlations among
indicators.
Data collected during the 1990 field
season by the Forest Service (FS) in the
six New England states (Miller-Weeks and
ENVIRONMENTAL
VALUE
Soil
Water
Air
Vegetation
Animals
Sustainability
~
~
—
~
~
Productivity
—
—
—
¦
¦
Aesthetics
~
~
¦
~
~
Diversity
—
—
—
¦
¦
Extent
¦
¦
¦
B
Utilization
—
—
—
¦
m
Contamination
¦
—
¦
¦
m
Quality
~
—
B
~
~
High
ASSESSMENT RELIABILITY
| Moderate Qlow —Not Applicable
Figure 7.1. Current status of assessment framework for forests.
-------
Gagnon 1990) are being analyzed with a view
towards producing a statistical summary.
Details of the interagency summary are
currently being decided. EMAP-Forests is
contributing statistical summaries of
meteorological, air quality, and pollution
deposition data to this effort.
Data collected by the EPA and the FS
in New England and Virginia during the 1990
field season are being analyzed to explore
statistical relationships among some
indicators of forest condition (Palmer et. al.
1990). Ecosystem models haws been proposed
to assist interpretation, but their possible
applications have not been specified in detail.
Individual analysts are developing procedures
to summarize and interpret various subsets of
forest indicator data (e.g., soil measurements
and observations of visual symptoms).
Linkages between environmental concerns and
indicators have not been specified in detail.
Prototypes of certain auxiliary (off-
frame) data bases have been acquired to
evaluate their potential utility. These include
portions of the Soil Conservation Service (SCS)
STATSGO data base, the National Oceanic and
Atmospheric Administration (NOAA) TD-3220
data base, and the Forest Service-Forest
Inventory and Analysis (FIA) data base.
7.3 A STRATEGY FOR FHM ASSESS-
MENTS
A forest ecosystem may be defined
(after Waring and Schlesinger 1985) to include
living organisms and non-living substrates
from the top of the canopy to the lowest soil
layers affected by biotic processes. They are
open systems that exchange energy and
materials with other systems. Systems theory
suggests that forests can be modeled as
collections of compartments and fluxes of
materials and energy. Hierarchy theory
suggests that the model scales can be linked
and that models to interpret indicators should
be consistent with the scales of
measurements. These considerations will
have much to do with a modeling strategy for
FHM assessments.
A conceptual model of the forest
ecosystem (Figure 7.2) suggests the potential
biological scope of inquiry of forest
assessments. Scientists can choose to
represent and model these components and
processes at different scales for different
purposes. Ideally, models for biological
interpretation of monitoring data would be
specified at the same scales as the various
indicators that are the main focus of
assessment (O'Neill 1988). Consideration of
linkages among scales is also necessary
because some measurements may be made at
finer scales and because the context for
assessment is always given by a higher level
in the hierarchy. To implement the
environmentalist paradigm, ecosystem models
such as these will have to be augmented to
reflect human interactions with the
environment.
Levins (1966) has suggested that any
single model can emphasize only two of the
three characteristics of generality, precision,
and realism. The trade-off for monitoring is
that models for descriptive monitoring
emphasize generality while those for
interpretive assessment s emphasize precision.
Modeling to meet EMAP's primary Tier 1 and
Tier 2 assessment goals has these objectives:
• Summarize current status, extent, and
trends of forest condition.
• Detect unusual situations of forest
condition.
• Summarize correlative evidence linking
those situations with man-induced
stresses.
• Relate forest condition to environmental
values. Several types of models are
defined here:
7-3
-------
ATMOSPHERE
H2Q. Carbon, Chemicals
VEGETATION
Nutrlsntt. H20
Aquatic
Ecosystem
ANIMALS | i
AQUATIC
OUT
SOIL
GROUND WATER
SURFACE WATER
TERRESTRIAL
VERTEBRATES
TERRESTRIAL
INVERTEBRATES
Figure 7.2. Conceptual forest ecosystem model (Anonymous 1988).
Indicator - Indicator models define
relationships between measurements
and indicators, and hence endpoints.
Classification - Classification models
define relationships between indica-
tors and axes of the classification
schemes.
Index ~ Multiple indicator models
define index values from sets of
indicators.
Interpretive - Interpretive models
define relationships between indicators
and measurements or auxiliary data
not included in the above model types.
Valuation - Valuation models are
objective functions that define
relationships between indicators or
indices and environmental values.
For a given scale of monitoring, link-
ages to finer scale patterns and processes are
defined by indicator models, and linkages to
higher levels are defined by valuation models.
Classification and index models are only
defined at the chosen scale for monitoring and
interpretive models are preferably defined at
that scale also. In this scheme, descriptions
and summaries of status and trends of condi-
tion are based on indicator, classification, and
index models. Linkages to environmental con-
cerns utilize valuation models to define policy-
relevance of status and trends. Exploratory
7-4
-------
analyses and correlations utilize interpretive
models.
Models defined at scales finer than the
monitoring scale are not made explicit in the
general assessment strategy because they
consider specifics that cannot be resolved by
monitoring indicators. Rather, these models
are in the realm of Tier 3 and Tier 4 monitoring
and research. But as part of the indicator
development strategy, Tier 4 models may
identify indicators for monitoring at Tiers 1 and
2. They may also identify more detailed
measurements to track particular cause and
effect relationships for Tier 3 monitoring.
Assessments will utilize these types of
models to organize, synthesize, and interpret
the data. In a flexible and evolving system,
there can be several models of each type, and
not all models need be present or in a
comparable stage of development. To avoid a
chaotic evolution of conceptual models,
measurements, and assessments, it is
desirable to emphasize model refinement
rather than model replacement.
7.3.1 Statistical Assessment Models
Emphasis will be placed on statistical
models developed with knowledge of
biological processes. Initial applications of
these models require simplifications and
assumptions that can be modified later (see
Section 7.3.2). The description of forest status
and trends starts by reducing forest
measurements to a set of values utilizing
indicator and/or index models. This is done for
each indicator or index for each measurement
at each site. In the simplest case, statistical
estimation formulas (Section 6) are then
applied to these values to provide a regional
description of the status and trends of forest
condition. In most cases, it will be possible to
develop more meaningful regional descriptions
by utilizing a classification model to stratify
the sample for analysis.
Assuming that a nominal-marginal-
subnominal scheme (a valuation model) has
been decided for each indicator or index, the
spatial and temporal status and trend
descriptions can be given in terms of
environmental concerns rather than response
indicators. The statistical estimation formulas
of Section 6 also apply here.
7.3.2 Interpretive Assessment Models
Interpretive modeling may improve
upon statistical descriptions by finding and
increasing the accuracy of indicator,
classification, and index models, by
introducing new interpretive models, and by
refining valuation models. Interpretive
assessments will usually require a changing
array of models over time as different
environmental concerns and biological
phenomena become important. It is expected
that interpretive models which prove useful
would be incorporated into statistical
summaries.
Mechanistic and heuristic models are
important modeling approaches. The
mechanistic approach would be needed, for
example, to estimate quantitatively the specific
effects of a specific stress on a specific
environmental value. Mechanistic models
would also be needed to account for
interactions among indicators, and among
indicators and space-time, that are not
accountable using statistical models alone.
The heuristic approach would be needed, for
example, to define the best way of
representing system behavior in a mechanistic
model.
7.3.2.1 Modeling Themes
Model development is needed to:
• Improve statistical descriptions by finding
and increasing the accuracy of indicator,
classification, and index models.
7-5
-------
• Improve interpretations by introducing
interpretive models.
• Improve the relevance of all assessments
by finding and refining valuation models.
7.3.2.1.1 Indicator, Classification, and Index
Models
Statistical assessments can be
improved by enhancing the apparent signal of
condition. This process incorporates
additional data and understanding to produce
new indicators or values with more
information content than before. The
increased information content is evidenced by
an increased robustness, perhaps to particular
stresses.
A simplified example approach for
isolating a signal due to air pollution will be
described. This general approach has been
used in dendroecological studies (e.g., Graybill
1982; Cook 1987; Kincaid and Nash 1988;
Zahner et al. 1989) and in assessments of soil
(e.g., Bouma 1989) and water quality (e.g.,
Radford and West 1986). More complex
formulations are possible.
Let I = f(F) = g(A) + h(P) + e [7.1]
where I = a response indicator of forest
condition,
F = a set of forest state variables
used to construct I,
A ¦= a set of environmental variables
affecting the state variables in F,
P = a set of pollution variables
affecting the state variables in F,
f, g, and h are functions, and
e = remaining unexplained variation.
Without signal enhancement, associa-
tions between indicators would be estimated
by the relationships between f (F) and h(P) and
tested by reference to e. Signal enhancement
is designed to extract g(A) from e, and thereby
reduce the "noise" in the association. This is
essentially a covarlance-type analysis where
the covariate is taken to be a generalized
function of the environment that partitions the
"normal" variation of an indicator.
Distinctions between "normal" and
"abnormal" values of an indicator are
contextual. For example, an analysis of
pollution effects could focus on pollution
signals after adjustments for stand dynamics
or weather in the covariance function, whereas
an analysis of natural versus man-Induced
stresses might utilize a different formulation.
Much more complicated formulations will be
required to deal with confounding and
correlations among the various explanatory
factors that are explored in any analysis.
The "signal enhancement" model is a
reason for understanding normal patterns and
trends in forest condition, as opposed to
developing understanding of mechanisms of
abnormal patterns and trends. This model is
also a reason for collecting certain "ancillary"
data on monitoring sites, that is, data that are
used to estimate "normality" rather than the
indicators of abnormal response, habitat,
exposure, or stress.
The classification schemes become
very important when considering these types
of models. That a given function I may not
have the same meaning for different
classifications was alluded to earlier.
Stratification of the population to obtain
comparable meanings for indicators implies
concomitant subsetting for the definitions of
the submodels in equation 7.1. This is a
problem because it will require more models,
but the models for each case should be
simpler. In fact, subsetting by classification
variables is perhaps a more viable option than
developing a single mega-model applicable to
all situations.
7-6
-------
The classification axes may be thought
of as covariates of the type g(A). Variables
deemed mandatory from first principles may in
fact be handled better via classification. If
classifications are based on vegetation
composition and soils, they will likely reflect
forest and soil development and succession
which in turn depend upon biophysical
variables. One would expect that soil parent
material, and long-term moisture and
temperature regimes will be the most
important biophysical variables that determine
the classification of a given site. With
classification, the functions g(A) probably need
not consider global geology or climate, only
local fluctuations. The importance of local
fluctuations should be explored with a good
understanding of global trends.
When indicators are refined, the
question arises whether or not the indicator
still takes on the same meaning at all sites.
For example, the same quantitative value of I
may imply different conditions in two different
resource classes. One way to account for
these differences is to consider an analysis of
deviations from expectations ("normal
condition") as an alternative to an analysis of
the indicator values themselves. This would
complicate the setting of assessment
endpoints but may offer a more realistic
regional picture of status and trends.
In the simplest case, using the
nomenclature from above, replace the indicator
I by the new indicator I*, where I* is a function
(e.g., a scaled difference or ratio) of the
observed indicator and its expectation under
"normal" environmental conditions. In other
words, I* is a deviation of I from its expected
value that is not explained by ancillary
variables and that is scaled in some fashion to
make it more comparable among resource
classes. The expected value is dynamic
because g(A) is dynamic. In this way,
normality need not imply an unchanging
condition.
With this formulation, it becomes
easier to see how statistical expectations
from historical trends can be utilized. While
expectations for the response indicator I were
defined above with reference to process
models (i.e., g(A)), statistical expectations
based on past experience or spatial pattern
can also be used for this purpose. In general
application, the current best estimate of
unexpected value can be used to adjust
response indicator values. New understanding
of environmental processes can be introduced
into the analysis of change by g(A) which may
be a dynamic function over time and may be
freely modified for analyses of different types
of environmental stresses.
7.3.2.1.2 Interpretive Models
Forest monitoring produces data that
can be used to study the forest conditions in
an epidemiological framework. The heuristic
approach is based on observational data that
arise from a cross-sectional sample rather
than on experimental data that arise from
controlled comparative trials (Fleiss 1973). The
absence of randomization severely restricts
the testing of mechanisms but does not
prevent identification of possible cause and
effect relationships (Mosteller and Tukey 1977).
An early discussion of the possibilities and
approaches is given by Wallace (1978). A
good and more recent discussion of
possibilities in forest monitoring is given in
Schreuder and McClure (1991).
It was mentioned in Section 2 that
forest monitoring data can be used to satisfy
two of the four NRC (1989) criteria for inferring
causality. Measures of correlation and of
temporality are typical tools in epidemiology
and have already been discussed as part of
statistical assessments. This section
describes two possible refinements that could
be used in interpretive studies, namely
"gradient studies" and "fingerprinting". In
practice, these techniques are likely to be
7-7
-------
utilized on an ad hoc basis, lor the appropriate
epidemiological tool will depend on the
particular circumstances.
In situations where a known
environmental gradient exists, observational
data taken along the gradient can potentially
be used to test the association of forest
condition with that gradient. For example,
Ohmann and Grigal (1990) were able to
associate concentrations of sulfur in wood
with sulfate deposition by sampling woody
tissue along a known sulfate gradient. In
other examples, dendroecologistscan typically
associate ring widths with distance from
smelters, ecophyiologistscommonlyassociate
tree distribution to temperature and moisture
gradients. Gradients can exist in time as well
as in space which opens up the possibility of
time-series techniques such as intervention
analysis, and combined space-time analyses.
"Fingerprinting" (MacCracken and
Moses 1982) refers to testing a particular set
of observations against the sets of
observations that would be expected under
various types of stresses. Waring (1990)
advocates and describes this approach for
diagnosing causes of change in forest
ecosystems. Johnson (1988) puts the
discussion into a statistical framework that
suggests an approach to developing
multivariate indicators of condition. Simmleit
and Schulten (1989) provide an application of
statistical "pattern recognition" for finger-
printing damage symptoms In forest trees.
Fingerprinting is a general tool that a know-
ledgeable analyst will use to build a case for
or against a particular cause and effect
hypothesis, and many variations on the basic
theme are possible.
Mechanistic models to interpret Tier 1
and 2 monitoring data are not readily available.
These large-scale and long-term models can
be based partly on existing theory but
additional work is needed to conceptualize a
hierarchical structure that operates at the
appropriate scales and that includes linkages
to higher- and lower-level processes and
scales (O'Neill et. al. 1986). The results of
heuristic modeling currently underway can be
applied to this conceptual development.
7.3.2.1.3 Valuation Models
An important assessment function is
to relate observed changes or possible future
scenarios to impacts on society. The first
priority is to identify valuation models so that
the boundaries of "good" and "bad" condition
are identified for any response indicator. But
these models should go beyond simple
classification of condition. Valuation models
and objective functions are needed to
quantitatively assess the relationships
between a certain value or change of value of
a response indicator, and an impact on
society. This implies that societal values must
be quantified much more specifically than they
are now.
This effort can build upon past
research in resource economics, especially for
environmental values such as "utilization" that
are typically measured by a monetary scale.
For other environmental values such as
"aesthetics" or "sustainability" the metrics are
much less clear.
7.3.3 Statistical Regionalization Using Off-
frame Data
Regionalization refers to the process
of aggregating site-specific data at regional
scales using various classification such as
political or administrative boundaries and
forest or soil types. A focus on classification
combined with the use of off-frame data
distinguishes statistical regionalization from
landscape ecology, although both approaches
yield regional answers. These techniques may
complement the statistical techniques based
on the sample frame (see S
7-8
-------
ection 4.5). Preliminary reviews of available
techniques suggest that post-stratification is
the most viable alternative. Other techniques
of regionalization such as spatial statistical
methods or the extrapolation of intensive
research site data have been suggested, but
their applications require further development.
EMAP is currently evaluating schemes
for combining on- and off-frame data (Overton
1990). The FHM program has additional
concerns such as the compatibility of forest
area estimates made by different agencies.
7.3.4 Auxiliary Data Bases
Table 7.1 lists some of the data bases
that are of primary interest to FHM
assessments. They will be used to aid the
analysis and interpretation of forest
monitoring data, providing unique or
supplemental data for:
• Better estimates of the extent of monitored
conditions (i.e., extrapolation and
interpolation).
• Better interpretation of the status and
trends of forest condition (i.e., correlation
with environmental stresses).
Preliminary plans for the air quality and
atmospheric deposition components of FHM
list the atmospheric constituents of interest
(Table 7.2). Preliminary plans for the
meteorology component of FHM consider
extreme events (tornadoes, high wind, hail),
drought, freeze, growing season measures,
and possibly lightning events.
7.3.5 Uncertainty Estimation
Uncertainty estimation is an integral
feature of assessments (Walters and Holling
1990). It is useful to know when a change in
a given indicator is small in relation to the
uncertainty about the components of the
indicator because conclusions on the basis of
the available information could be erroneous.
Thus, the quality and uncertainty of the data
which are collected and reported will be
documented as part of assessment reports.
Uncertainty is partly due to imprecision
or bias in the measurement system arising
from, for example, inconsistent field
instrument readings, missing data, unstable
analytical stock solutions, or detectability
limits. Uncertainty also results from
extrapolating sample data to regional
populations. Measurement uncertainty can
usually be controlled at an acceptable level
through the application of a rigorous QA
program during all phases of measurements.
Uncertainty of extrapolation, on the other
hand, is controlled and estimated through
application of statistical principles for
sampling and aggregating data to describe
populations.
Each additional level of sample
aggregation adds a degree of uncertainty to
the results which is dependent upon the
appropriateness of the aggregation scheme.
Some of this uncertainty can be reduced by
covariance analyses, but the development of
an effective and straightforward sample
aggregation scheme for each indicator
deserves high priority. Such schemes will help
to guide the development of the various
indicators, the associated uncertainty models,
and the interpretive models that will be applied
in later analyses.
The statistics to be reported for each
variable could include the following:
• Measurement precision, sampling error,
and standard errors for population-level
estimates, which could be reported as
measurement uncertainty, sampling
uncertainty, and extrapolation uncertainty.
7-9
-------
Table 7.1.
Auxiliary Data Bases: Uses, Sources, and Acquisition Intervals.
Data type
Uses
Source Periodicity
Forest
Regionalization
USFS (RPA) 5-10 yr
inventory
Extrapolation
Soil
Regionalization
SCS (STATSGO) 10+ yr
inventory
Extrapolation
Air Quality
Correlation
EPA (EMAP-Air As needed
& deposition
Interpretation
& Deposition)
Weather
Correlation
NOAA As needed
Interpretation
Pest outbreak
Correlation
USFS (FPM) Event-related
surveys
Interpretation
Topography &
Classification
USGS 10 + yr
hydrology
NASA
Wildlife
Monitoring
FWS 5-10 yr
Table 7.2. Atmospheric Constituents of Interest®.
Constituent
Sample Frequency
> Data Use
Ozone
Continuous
regional status
(hourly Averages)
and trends
Total deposition:
Wet deposition6
Weekly
regional status
and trends
Dry deposition*
Weekly
regional status
and trends
Ctoud/fog
Episode
regional status
deposition0
and trends
'From Baumgardner, Shadwick, and Smith. Draft plan for air quality and atmospheric deposition
in EMAP-Forests. EPA, Research Triangle Park NC.
"Parameters for wet deposition are H+, NO"31 SO NH\, Ca+S, AT3, Mg+a, and CI".
'Parameters for dry deposition are total sulfur flux and total nitrogen flux (S02l HN03, HN02, N0'3,
SO'24l and NH\.
^Parameters for cloud/fog deposition are the same as for wet deposition. Cloud/fog deposition
is limited to areas about cloud base and along some coastal areas.
7-10
-------
• Percent of samples above the
mea- surement system detection
limit.
• Percent of planned sample size actually
obtained (a measurement of completeness
of the sample).
• System detection limits and inherent
precision at different magnitudes of
measurement.
• Ratios of various components of
uncertainty.
7.4 STRATEGY ELEMENTS AND GOALS
Report production, infrastructure, and
planning are elements of a strategy. For each
element, short- and long-term goals are
suggested in this section.
7.4.1 Report Production
The long-term goals are to produce a
multi-agency interpretive report on forest
condition in 1993, and to participate in an
EMAP integrated assessment in 1994 or 1995.
In the short-term, reports can be produced as
data become available, determined mainly by
progress in implementation and interagency
cooperation. It is expected that EMAP-Forests
will develop regional statistical assessments
with the FS starting in 1991 or 1992, and multi-
agency, national statistical assessments by
1995 when implementation is completed
nationally. In 1991, EMAP-Forests will assist
the EMAP-Integration and Assessment Project
in preparing an example interpretive report
based on data collected in the 1990 field
season.
7.4.2 Infrastructure
National leaders will be responsible for
each combination of environmental concern
and resource type (see Table 7.1). These
leaders are scientists and are primarily
responsible for defining response indicators
and measurements. They also assist in
defining assessment endpoints for their
response indicators.
A national coordination group will also
be identified. The national group will be a
small interagency group, including the EPA the
FS, and others. This group will be responsible
for defining environmental concerns, resource
issues, and priorities of concern. In
cooperation with scientists, they will also
define assessment endpoints from response
indicators. Finally, the national coordination
group will be concerned with the procedures
and scheduling of analyses and reports.
Short-term goals include identification
of regional and national leaders and their
responsibilities. Longer-term goals include
closer integration of EMAP-Forests
assessment teams with the larger EMAP
assessment team. EMAP-Forests teams could
include personnel from other resource groups
to facilitate the later merging of efforts into
one organizational structure.
A multi-agency outreach program will
be emphasized rather than a strategy of
accumulating a large number of in-house
analysts. This strategy will require that data
bases be readily available and easily
accessible to many analysts around the
country. The Forest Information Center (FIC)
described in section 10 is an important
component of the assessment infrastructure.
7.4.3 Planning
In FY91, much of the assessment
effort will be devoted to preparing parts of the
multi-agency monitoring plan scheduled for
February 1992. As part of preparing the
research plan, EMAP-Forests will:
• Identify the various pieces of the
infrastructure described earlier.
7-11
-------
• Lead an effort to reach out to other
agencies for participation in assessments.
• Conduct studies, workshops, and reviews
aimed at better definition of environmental
concerns and assessment endpoints and
of analytical methods for organizing,
summarizing, and interpreting data.
• Develop plans for integrating with the EPA
risk assessment process.
7-12
-------
8. QUALITY ASSURANCE PROGRAM
8.1 INTRODUCTION
The underlying reason for quality
assurance (QA) is to provide confidence in the
environmental data and statistics generated
by the Forest Health Monitoring (FHM)
participants. Hence, the mission of QA in FHM
is to ensure that all FHM data and statistical
products are of documented and sufficient
quality to satisfy the needs of data users,
policy makers, and the public.
EMAP will operate within the guidelines
of the EPA's Quality Assurance Management
Staff (QAMS). Comprehensive QA techniques
will be employed to ensure the quality and
usefulness of the data. The overall policies,
organization objectives, and functional
responsibilities designed to achieve data
quality goals for FHM activities are described
in this section. Other topics discussed in this
section include the process of establishing
data quality objectives (DQOs), total quality
management, documentation, and reporting.
8.2 QUALITY ASSURANCE POLICY
EPA Order 5360.1, "Policy and Program
Requirements to Implement the Quality
Assurance Program" was issued in April 1984,
to help ensure that all decisions made by the
EPA are supported by a valid data base. This
goal necessitates the integration of QA into all
data collection activities.
EMAP policy, which requires integral
QA components for all data collection and
processing activities, will follow this approach.
A QA program to ensure that all data are of
known and documented .quality will be
established. Resources commensurate with
the goals and objectives of the EMAP program
will be made available to the QA staff to
accomplish these goals.
EMAP is a major environmental data
collection effort and an emerging program.
The overall QA program for EMAP is described
in the Quality Assurance Program Plan (QAPP)
(Graves 1990; EPA 1987). As the processes in
EMAP are refined and optimized, the QAPP will
be modified to reflect these improvements.
The specific FHM QA program will be detailed
in a Quality Assurance Project Plan (QAPjP)
that must be prepared prior to full
implementation of the monitoring program.
8.3 TOTAL QUALITY MANAGEMENT
8.3.1 Introduction
Total quality management (TQM) is a
process of continuous improvement and
innovation led by the program directors in
which management philosophy, planning, and
operating methodology are fully integrated.
EMAP program directors must be committed
to quality improvement in all aspects of the
program. This is a relatively new operational
framework within the EPA which concentrates
on providing a service to the internal or
external client (customer) by improving the
systems within which all work is performed.
This primary tenet is directed toward obtaining
customer satisfaction, a tenet to which QA is
directly and intrinsically linked (Figure 8.1).
8.3.2 TQM Philosophy
The TQM philosophy is participatory in
nature; it is aimed at achieving total employee
commitment to quality. This commitment
must span all levels of EMAP, from the
program directors to the scientists and
technicians conducting basic implementation
and support activities. The EMAP program
directors must be committed to the TQM
approach and its integration into their day-to-
day management activities. This means that
EMAP personnel at every level must be
committed to the management of all aspects
8-1
-------
QA
¦^TQM
QAPP/TQM
ESTABLISH PRINCIPLES AND
GUIDANCE FOR EMAP PERSONNEL
DQO
ESTABLISH CUSTOMER NEEDS
QAPjP
^ ESTABLISH THE PROCESS FOR
SATISFYING CUSTOMER NEEDS
AUDITS
EVALUATE THE PROCESS TO
IMPROVE IT AND TO JUDGE
ITS ABILITY TO SATISFY
CUSTOMER NEEDS
REPORTS
w PRODUCT TO MEET CUSTOMER
NEEDS
Figure 8.1. Relationships between quality assurance and total quality management's
primary tenet "customer satisfaction."
of the quality of the output, product, or service
(Figure 8.2).
Data collection phases, from sample
collection to analysis, can be viewed as a
series of inputs and outputs where quality is
directly affected by TQM applications.
All EMAP participants must understand
the goals of the program and make
contributions to the decision-making
processes that are pertinent to their roles. In
summary, TQM focuses on:
• Client identification - All "clients" must be
identified and brought into the process to
articulate their requirements at each
program level in terms of operations,
resource needs, and fund tons. Effective
and continuous communications of client
requirements must be maintained among
the large network of participants.
• Standards and Performance - Arbitrary
quotas and goals must be replaced by
standards and measures of performance
which are proactive rather than reactive.
• Commitment by Program Directors - TQM
requires commitment, engagement,
direction, and support from program
directors to succeed. This commitment is
exemplified during training by establishing
procedures and policies that foster a TQM
"culture."
• Employee Recognition - Recognition of the
importance of all EMAP participants is a
key ingredient to the success of TQM.
Criteria and mechanisms for employee
recognition are essential. They emphasize
human aspects such as effort, creativity,
and achievement. These criteria serve as
8-2
-------
KEYS TO TOTAL QUALITY MANAGEMENT
CULTURE
EMAP PARTICIPANTS
OUTPUT
VISION
VALUES
GOALS
OBJECTIVES
CUSTOMER FOCUS
TRAINING
ACCOUNTABILITY
RESOURCES
PRIORITIES
PEOPLE SENSITIVE
TEAMWORK
QUALITY DESIGN
REWARDS & RECOGNITION
LEADERSHIP
EMPLOYEE EMPOWERMENT
INVOLVEMENT
QUALITY DATA
QUALITY REPORTS
QUALITY STUDIES
Figure 8.2. EMAP commitment to total quality management.
benchmarks for evaluating the effectiveness of
TQM training and implementation practices.
8.3.3 Training
Training increases employee ejqaertise
and enhances employee decision-making
capabilities. Given the complexity of EMAP,
training should focus not only on technical
proficiency, but also on the development of
organizational and interpersonal skills.
A framework for TQM training and
implementation at the agency level is the
President's Award for Quality and Productivity
Improvement. The award recognizes agencies
that have implemented TQM in a manner that
results in high quality products and services
and effective use of taxpayer dollars. It also
promotes quality and productivity awareness
in all federal programs. A similar but more
focused process within FHM could be
developed for TQM recognition at the Resource
Group, indicator, laboratory, field, or
participant level. Methods for assessing the
effectiveness of TQM at these levels in the
FHM program have not yet been developed.
8-3
-------
Tangible outputs from this focus on TQ could
be in the form of an annual award, for
example, with additional appropriate
recognition in a newsletter format.
8.4
ORGANIZATIONAL STRUCTURE
Figure 8.3 represents the proposed
organizational structure for the operations of
the FHM QA Staff.
EMAP
QA Coordinator
EMAP-F
Technical Director
FHM Program
Manager
EMAP-F
National QA Officer
(EMSL-LV)
QA Officers
Pacific Northwest
Regional QA
Officer
(e.g., ERL-C)
Intermountain
Regional QA
Officer
(e.g., EMSL-LV)
Northeast
Regional QA
Officer
(e.g., Radnor*)
Southeast
Regional QA
Officer
(e.g., AREAL-RTP)
Auditing
Evaluation
Verification
~Headquarters for NE Forest Experiment Station.
Figure 8.3. Proposed organizational structure for FHM quality assurance staff.
8-4
-------
8.4.1 Quality Assurance Coordinator
The quality assurance coordinator
(QAC) has responsibility for the EMAP QA
program and its implementation. The QAC
serves as an advisor and interacts with the QA
representative for each resource group (i.e.,
national QA officer for FHM) to ensure that
each group projects the DQOs of the QAPP.
The QAC also oversees the development of
DQOs and documentation standards.
8.4.2 Technical Director
The FHM Technical Director (TD)
coordinates several activities, including direct
interaction with FHM participants, especially
the National QA Officer and the indicator
leads. Such QA functions include:
• Providing adequate resources.
• Coordinating development of DQOs.
• Overseeing development of QA project
plans.
• Implementing TQM.
• Ensuring adequate training of personnel.
• Ensuring that audits are conducted.
• Supporting the QA program.
8.4.3 National Quality Assurance Officer
The national QA officer (NQAO) is
responsible for QA in all FHM activities and
reports directly to the FHM program manager
and TD. Located at EMSL-LV, the NQAO will
work closely with the indicator leads, regional
QA officers, and other FHM participants,
including the Forest Service (FS) on QA
matters.
Specific responsibilities of the NQAO
include:
• Providing input to the development of the
EMAP QAPP.
• Developing of the FHM QAPjP.
• Guiding and overseeing activities of
regional QA officers.
• Assisting in the development of the FHM
information management system for data
tracking, generation, and processing
activities.
• Facilitating DQO development and methods
selection within FHM.
• Developing guidance documents.
• Assisting the TD in implementing the QAPjP
and QAMS documents and guidelines for
the mandatory EPA QA program.
• Preparing the QA Annual Report and Work
Plan, a document that summarizes the
accomplishments of FHM and recommends
improvements.
• Providing the laboratory communication
link between the QAC, QA representative,
and QAMS.
• Approving all contractual QA supporting
documentation.
• Serving as a representative of FHM QA
during EMAP resource group QA meetings.
8.4.4 Laboratory QA Officers
The laboratory QA officer (LQAO) is
responsible for ensuring that each project
within an EPA laboratory satisfies the
laboratory's requirements for QA programs.
The LQAO evaluates QA plans, coordinates
and supervises systems audits, and
disseminates information. The LQAO at EMSL-
LV will work very closely with the NQAO to
ensure that an appropriate QA program is
developed. The issue of different EPA
laboratory QA requirements will have to be
addressed.
8.4.5 Regional QA Officers
A four-person core team of regional QA
officers (RQAOs) will serve as a basis for the
FHM QA structure. The NQAO will guide and
oversee the QA activities of the regional QA
coordinators. The RQAOs could be head-
quartered in close proximity to the center of
8-5
-------
regional FHM activities but not necessarily co-
located there. Within the FHM QA group, the
development of a QAPjP would include the
cooperative efforts of the NQAO and RQAOs
to ensure that the QAPjP guidelines were
acceptable to all participants. For each
region, the RQAO would be responsible for:
• Ensuring data quality following the
guidance from the national QAPjP.
• Preparing a regional QAPjP.
• Selecting audit teams and conducting
audits for the various in-house projects
including follow-up audits.
• Preparing the QA Annual Report and Work
Plan.
• Providing a communications link between
the NQAO and regional personnel.
8.4.6 Indicator Leads
Each indicator lead within FHM has a
QA responsibility for the data quality of each
within-indicator measurement. To ensure that
the resulting data are of acceptable quality,
the indicator lead must be intimately familiar
with components of the indicator such as
sampling protocols, preparation, analysis, and
logistics. Each indicator lead will be ultimately
responsible for activities such as auditing,
performance evaluation, and methods of
verification (which includes accuracy and
precision checks). The indicator lead will work
closely with QA personnel in planning and
conducting these activities. The RQAO will
eventually be very active in actually conducting
the audits. The indicator lead can develop
protocols and provide training, but cannot
physically audit 3Q to 40 national field crews
each year. However, for the FY91 sampling
season, it is probable that the indicator leads
will have to plan for their field audits. "This
could be done by training a non-crew field
person. The NQAO and RQAO will serve as an
advisor and resource person to the indicator
lead. Cooperation between indicator leads
and NQAOs is important in the development of
each indicator.
8.4.7 Matrix Activities in EMAP
Data from other sources such as
landscape characterization, air and deposition,
and the other resource groups will be used.
The FHM NQAO, working with the TD, will be
responsible for coordinating QA activities with
other cross-cutting groups and resource
groups.
Liaison programs are designed to
inform cooperating agencies and organiza-
tions. Several types of liaisons have been
designated: intra-agency, congressional, and
interagency. To date, the program directors
(Associate Directors and above) have those
liaison responsibilities.
8.4.8 Communications
The establishment of mechanisms and
protocols for the exchange of information is
essential to the success of any scientific
endeavor. This is especially true for a
program of such large scope and complexity
as EMAP. EMAP QA communications must be
timely and responsive.
Figure 8.3 can also be used to
designate lines of communications. RQAOs
will contact the NQAO to report progress or
issues. The NQAO will be responsible for
conveying information to TDs and the EMAP
QAC. The RQAOs will be in communication
with regional implementation/logistics
personnel.
Conference calls will be established at
regular intervals for the FHM QA staff to
discuss QA issues. Conference calls between
the NQAO and the EMAP QAC will also be
established.
8-6
-------
8.5 QUALITY ASSURANCE OBJECTIVES
8.5.1 Specific Objectives
Specific QA objectives that are being
defined for FHM include:
• Compatibility of data - Common types of
data generated by different methods and
by different resource groups must be
compatible. Data that describe the same
measurement activity must also be
compatible over time. Emphasis must be
placed on data base flexibility because of
issues, unknown now, which may arise in
future years of this multi-year program.
• Satisfying the DQOs - Acceptance criteria
established during the DQO-development
process serve as benchmarks for
satisfying data user requirements. The
FHM DQOs may be established for several
levels of data collection (e.g., sample
measurement system, measurement
parameter, or indicator level).
• Documentation of data collection -
Effective documentation will ensure that
information on data quality, statistical
design, algorithms, protocols, and
analytical procedures are available for
scrutiny. Documentation is also invaluable
for any technical defense requirements.
The scope and duration of EMAP will
certainly require reexamination of data,
methods, and conclusions as the program
progresses. Thorough documentation is
essential in establishing the flow of
information, alternatives, decisions, and
conclusions which form the basis for new
activities or require the revision of previous
conclusions based on new methods or
information.
• Data verification - A systematic approach
to data verification will ensure that all data
are subjected to some basic standards of
accuracy that verify the authenticity, but
not necessarily the validity, of the data.
Verification is accomplished by comparing
data at each level of processing to
established data quality criteria.
• Data validation - A systematic approach
to data validation will confirm that
environmental measurements, processes,
and products are within acceptable bounds
of accuracy and validity with respect to the
population of interest.
8.5.2 Data Quality Objectives
The DQOs are considered to be
specific statements of the level of uncertainty
a data user is willing to accept in a body of
environmental data, with respect to the kind of
scientific or policy question that motivated the
data collection activity. DQOs are definitive,
quantitative, or qualitative statements
developed by data users (e.g., scientists,
policy makers, interest groups) in conjunction
with the QA staff. The DQO process uses an
iterative approach that balances costs versus
uncertainty to achieve a desired or acceptable
level of quality. This information can also be
used to allocate or redirect resources to
specific monitoring phases in order to
generate data of sufficient quality to support
management decisions or answer specific
scientific questions. The FHM QA program
could modify the QAMs process for
establishing DQOs.
8.5.2.1 Hierarchy of DQOs
Data quality and DQOs can be defined
for many levels of FHM data collection
(Figure 8.4). Possible levels might include:
• Measurement-level DQOs (MQOs) for
specific measurement parameters are
estimated by using existing or initial
baseline data. The MQOs may define
acceptance criteria for detectability,
precision, accuracy, representativeness,
completeness, and comparability in field
and laboratory measurement data (Byers
et al. 1990). Another criterion may be to
8-7
-------
Societal/Environmental
Endpoints
-¦ Cross-Ecosystem Assessment
Ecosystem QO
— Aggregate Indicators
Indicator QO
- Indicators oI Ecological Interest
Measurement QO
-¦ Baseline Oala
Figure 8.4. Hierarchy of data quality objectives.
optimize measurement uncertainty with
respect to non-measurement sources of
uncertainty (e.g., due to sampling design
constraints or naturally-occurring spatial
and temporal variability that often is
confounded within environmental data).
• Indicator-level DQOs (IQOs) are derived
from aggregated parameter data for
ecological indicators. The IQOs might
focus on the uncertainty associated with
the data aggregation procedures that are
used to assimilate measurement-level data.
As the indicators are expected to provide
specific information about resource
condition, component factors such as
measurement quality, sampling design, and
statistical analysis will likely be important
in developing these IQOs. The relative size
of trends to be detected based on the
specified indicators will provide important
input in the development of the IQOs.
• Resource-level DQOs (RQOs) are derived
from aggregated indicator data for the
FHM Resource Group. At this level,
indicator data may be aggregated to
provide an overall FHM assessment of
resource condition. Uncertainty associated
with each indicator is an additional factor
in the total uncertainty confounding the
interpretation of resource condition.
• Ecosystem-level DQOs (EQOs) are derived
from aggregated resource data for overall
ecosystem assessments. Integrated data
from all resource groups may be used by
EMAP policy makers to make regional-
scale assessments of overall ecological
condition in different regions of the U.S.
The uncertainty associated with each
resource-level assessment must be
included in the uncertainty estimate for the
regional-scale assessment. This level of
aggregation might not be feasible in the
8-8
-------
FHM context. "There might be many conflicting
concerns/endpoints which EMAP managers do
not want to try to aggregate.
8.5.2.2 The DQO-Setting Process
The development of DQOs will begin at
the planning phase (i.e., research plan) and
continue through intermediate phases (pilot
studies and demonstration projects) and the
implementation phase (regional monitoring
and data interpretation). For FHM, the DQO
process begins with the identification of
environmental questions brought forward by
EMAP directors, interest groups, and the
public. These questions will focus overall
monitoring objectives and help formalize and
quantify appropriate DQOs in an interactive
multi-staged process. Figure 8.5 highlights the
interactions between "top-down" and "bottom-
up" approaches in the development of DQOs.
Stage I defines major questions or
issues of concern by focusing on the data
user's information needs or requirements,
resource and time constraints, and
consequences of Type I and II errors. The
Type I error (false positive) indicates the
presence of adverse ecological effects when
such effects actually do not exist. On the
other hand, the Type II error (false negative)
indicates the absence of adverse ecological
effects when, in fact, real effects actually
exist. Establishing the data user's information
needs assumes that a consensus can be
reached on these data needs from different
data users. In reality, data quality will be
assessed and documented during the early
stages of the program. Reassessment and
development of DQOs will be a continuous
process.
STAGE I: PROBLEMS OF CONCERN
• Whet Is the purpose ol the environmental dale?
• What are the resources and time constraints?
• What are the consequences ol Type I and Type It errors?
STAGE III: SCIENTIFIC APPROACH
• What approaches to dais collection are available?
• Which approaches provide data quality commensurate with
Stage II requirements?
• Wh*l R&D activities are needed to meet Stage II requirements?
STAGE II: INFORMATION NEEDS
• What Is the population ol Interest?
• What level ol confidence must attend results?
• Does pertinent, usable data currently exist?
• What new data are needed?
Figure 8.5. The DQO process for continuous communication and feedback
among decision makers and scientists.
8-9
-------
Stage U defines those data that are
needed to answer the questions or make the
decisions identified in Stage I. This exercise
includes development of pertinent scientific
questions, definition of the populations of
interest, identification of specific design
constraints, examination of existing data, and
confirmation of requirements for collecting
new data.
Stage III determines the scientific
approach. Different approaches to data
collection, critical levels of data quality
required to meet Stage II constraints, and
research needed to address Stage I and II
concerns are identified.
A continuous chain of communication
persists during the DQO-setting process and
includes policy makers, program directors,
resource scientists involved in data analysis,
and scientists involved in the actual data
collection activities (see Figure 8.6).
8.5.3 Current Status
"Hie need for environmental information
has been stated in very qualitative terms at
Stage I. However, the EMAP steering
committee has not yet indicated precisely how
this information will be used by the various
client groups that have been identified.
Specific assessment criteria are being
developed. These will define ecological
condition, quantify its extent, and allow data
users to articulate their data quality
requirements in more quantitative terms than
are presently available. For Stage II, a group
of indicators must be developed that allow
scientists to establish techniques for making
overall assessments of resource condition.
Rationale statements must be developed
which describe the data to be collected for
each indicator and the way in which those
data will be used to provide information on
ecosystem condition. Such statements might
include critical values for condition and must
COLLECT DATA
PROBLEM SOLVING
CHANGE
THE PROCESS
STUDY
THE PROCESS
ADOPT
NEW PROCESS
EVALUATE PERFORMANCE
IOENTIFY
CRITICAL PROCESS
NEW PROCESS SUPERIOR
TO OLD PROCESS
Figure 8.6. The DQO continuous improvement process.
8-10
-------
be scientifically and statistically defensible.
Rationale statements should also relate each
indicator to assessment endpoints of societal
interest so that the ramifications of changes
in system condition can be understood and
appreciated by the wide variety of data users.
In addition, a thorough investigation of
data uncertainty sources must be conducted
for each candidate indicator. If possible,
actual estimates should be provided for each
source of uncertainty. In this way, factors
that contribute significantly to the overall
variability of the indicator are identified and
the effectiveness of various options in
resource allocation can be evaluated.
Until appropriate aggregation tech-
niques are developed, individual parameter or
indicator assessments will be used to make
discrete evaluations of condition. The
techniques for making aggregated ecosystem-
level assessments will be developed over time.
The DQO process should provide the
developmental framework by ensuring that
assessment techniques at all levels provide
information of sufficient quality to satisfy the
FHM objectives.
At all levels, it must be the
responsibility of the EMAP directors to initiate
and encourage the development of DQOs. The
EMAP QAC and a consulting group from QAMS
have indicated an interest in facilitating this
process. It is anticipated that the EMAP
Steering Committee will be responsible for
approving the IQOs developed for each
resource group. Use of the DQO process will
assure that the multi-level DQOs are
consistent with the overall EMAP goals and
objectives.
8.6 QA DOCUMENTATION AND REPORT-
ING
8.6.1 Quality Assurance Project Plans
(QAPjPs)
The EPA QA policy requires every
monitoring and measurement project to have
a written and approved QAPjP (Stanley and
Verner 1983). This requirement applies to all
environmental monitoring and measurement
efforts authorized or supported by the EPA
through regulations, grants, contracts, or other
formal means.
8.6.1.1 Purpose of the QAPjP
The QAPjP for FHM will specify the
policies, organization, objectives, and
functional activities for the specific project.
Each plan will also describe the QA activities
and assessment criteria that will be
implemented to ensure that the data bases
will meet or exceed all DQOs established for
FHM. The QAPjPs will be revised as necessary
to reflect changes in procedures that result
from continuous improvement. All project
personnel, especially indicator leads, should
be familiar with the policies and objectives
outlined in pertinent sections of the QAPjP to
ensure proper interactions among the various
data acquisition and management
components.
Due to the evolving nature of the
research plan, topics covered in the QAPjP are
generic in nature. As modifications to the
research plan are made, these descriptions
will be revised to address any modification.
8-11
-------
8.6.1.2 Content of a QAPjP
The QAPjP must identify all
environmental measurements within the scope
of the project goals and objectives and identify
specific processes within each measurement
that could introduce possible sources of error
or uncertainty in the resulting data. Methods,
materials, and schedules for assessing the
error contributed by each process must also
be addressed. The QAPjP must also define
the criteria and procedures for assessing
statistical control for each measurement
parameter.
Data collection activities must institute
sufficient control procedures, materials, and
techniques to minimize measurement errors.
Each process that could affect the quality of
the data (e.g., sample collection, preservation,
transportation, storage, preparation, analysis,
and data reporting) must be evaluated and
documented.
There are several QA requirements in
the QAPjP to monitor the step-wise process
for environmental measurements (Figure 8.7).
By using appropriate measurement quality
samples, it is possible to isolate the error
contribution and set control criteria based
upon the overall MQOs. This approach is
essential for providing diagnostic information
so that real time corrective action can be
taken to ensure control in satisfying these
DQOs.
Although the information in Figure 8.7
represents traditional data collection activities,
these guidelines might not be feasible for
identifying every error-contributing process
given technical and resource constraints. It is
essential that the QAPjP define the rationale
behind each QA application, describe the
specific measurements, and highlight the
application of QA in each case.
The QAPjP must also identify FHM
services that will be used to support QA and
assess the effectiveness of the QA program.
These include the QA responsibilities of
individuals in the project, preventative
maintenance of instrumentation, and
scheduling and scope of the audit program.
8.6.1.3 Responsibility
The TD for each EMAP resource group
delegates responsibility for generating the
QAPjP to the resource group QAO. Through
the QAPjP, program directors will establish
policies, criteria, and procedures related to QA.
Specific sign-off authority by program
directors has been established to assure
concurrence with the scope and content of the
QAPjP elements.
8.6.2 Standard Operating Procedures
Environmental monitoring SOPs are
devised for sampling, preparation, and
analysis, data management, QA, reporting
activities, accounting, project finance and
contracts, and analysis integration activities.
Written SOPs provide guidelines for
planning, implementation, and analysis
activities over time and among personnel for
routine activities within an organizational unit
(e.g., resource group), but not among units.
To ensure consistency in data among resource
groups, SOPs must be cooperatively
developed. In field and laboratory circles.
SOPs are often referred to as "methods" or
"protocols." The incorporation of QA into
methods development is very important. The
QA personnel should have an active
cooperating role in methods development with
each indicator lead. Subsequently, an active
sample exchange program will be initiated
when multiple laboratories are contracted for
analyses. Comparability of data from various
8-12
-------
SUBSTRATE
FIELD SAMPLE-
DUPLICATES
PROCESSED SAMPLE
Sample Collection
Sample Preparation
Transport & Storage
SAMPLE STABILITY
STUDIES
LAB REAGENT BLANKS
INTERNAL STANDARDS
LAB SAMPLE DUPLICATES
| Analysis
RESPONSE/OBSERVATION
NOTEBOOK AUDITS
DATA AUDITS
DATA BASE Qc
Data Reduction/Processing
DATA
CALIBRATION
CHECK STANDARDS
Figure 8.7. Assessment arid control of process errors within a measurement.
field and laboratory sources is a QA issue of
considerable importance.
The TD for FHM is responsible for
determining which activities require SOPs and
oversees the development, review, and
implementation of these SOPs. During
periodic audits, the NQAO for FHM should
document the status of all new SOPs.
8.6.3 Documentation
Current versions of the following
documents and information must be dis-
seminated among all appropriate FHM
participants and cooperating organizations:
• QAPjP - A document addressing all items
delineated by Stanley and Verner (1983).
These include clearly defined field and
laboratory protocols, including QA staff
responsibilities, and use of QA protocols,
and project DQOs.
• Laboratory methods manual - A
document containing detailed SOPs related
to laboratory and instrument operations.
• Field methods manuals - Documents
containing detailed instructions, including
field forms, for all field operations.
8-13
-------
8.6.4 Reports to Management
• Monthly Reports - The NQAO will submit
monthly reports to the TD and the QAC.
The report will discuss QA issues,corrective
actions, and accomplishments. The QAC
will submit a monthly report to the EMAP
Terrestrial Systems Associate Director.
This report will summarize the overall FHM
QA activities. Delivery dates have been
specified as the 5th of the following month
for the QAO report and the 10th for the
QAC report.
• Audit Reports - Management systems
reviews (MSRs) and technical systems
audits (TSAs) are stand-alone reports to be
delivered through specified channels on an
on-going basis. Audits of data quality
(ADQs) and performance evaluation audits
(PEAs) are incorporated into the monthly
reports. The QAO is responsible for
reporting the TSA, ADQ, and PEA reports.
• QA Annual Report and Workplan - The FHM
QA Annual Report and Workplan (QAARW)
discusses project activities during the prior
fiscal year. Recommendations for changes
in QA policy (with appropriate justifi-
cations) are also addressed. The workplan
component of the QAARW describes all
major QA activities for the coming year,
including DQO developments and
refinements, deliverables, audit schedules,
changes in the QA program, resource
requirements, active projects, tasks
involved in data generation, and approved
changes in the QAPjPs and SOPs. The
annual report of the QAC will be appended
to the QAARWs for each resource group.
8.7 QA OPERATIONS
The following subsections describe the
methods used to control and evaluate the
quality of data produced during the data col-
lection process (see Figure 8.8).
8.7.1 The Audit Program
The FHM QA staff will develop and
conduct laboratory and field audits and
reviews at the program and project levels.
These audits, which will be conducted with the
cooperation of the FHM indicator leads, will
aid in determining whether the QAPjPs are
being fully implemented, and if they are
adequate for the objectives of the project.
Audits will be conducted for all data collection
measurements.
8.7.1.1 Categories of Audits
QAMS has classified audits into four
categories (EPA, 1987) as follows:
• Management Systems Reviews (MSRs) -
assess the effectiveness of the
implementation of the EMAP QAPP.
> Audits of Data Quality (ADQs) - used to
check data accuracy, to determine whether
or not sufficient information exists within
the data set to support assessment of
data quality, and to verify that DQOs have
been satisfied.
• Technical Systems Audits (TSAs) - an on-
site visit used to verify conformance to the
QAPjPs and to confirm that good
laboratory and field practices have been
used in the generation of the environmental
data.
• Performance Evaluation Audits (PEAs) -
assess laboratory and field analyses
based on results achieved in the analysis
of blind samples; also serve as a check on
the comparability of data between
resource groups. Indicator leads will be
active participants with QA personnel in
developing and conducting these PEAs.
The possibility of field reference plots for
auditing the performance of field crews is
being considered.
8-14
-------
ESTABLISH STATISTICAL CONTROL
MAINTAIN QUALITY CONTROL CHARTS
NO
YES
ACCEPTABLE
EVALUATE PERFORMANCE
IDENTIFY
CRITICAL PROCESSES
DEVELOP A FLOW CHART
FOR THE ACTIVITIES
DEVELOP QUALITY CONTROL
MATERIALS/TECHNIQUES
RE-ANALYZE SAMPLE
CORRECT PROBLEM
IDENTIFY PROBLEM
REPORT DATA
Figure 8.8. Data collection process.
8.7.1.2 Corrective Action
The TD has the responsibility for verifying
corrective action for discrepancies
documented in the audit report. The NQAO is
responsible for tracking the corrective actions.
A Memorandum of Intent (MOI) must have a
concurrence page for sign-off. The MOI identi-
fies the problems and/or discrepancies, the
corrective action(s), and the remedial effects.
8.7.2 Data Verification
Verification is the act of determining
and controlling the quality of data. Verification
can be accomplished manually, electronically,
or through remeasurements.
8.7.2.1 Electronic Data Verification
All FHM data will eventually be placed
in an electronic format. Much of the data
collected in 1990 was entered into portable
data recorders in the field. Computer pro-
grams can be designed to perform logic
checks of most entries, automatically de-
termining if the code entered is valid and
logicalty correct.
Data can also be verified by relating it
to other correlated measurements (e.g., height
to diameter, and pH in different extracts). If
data verification is accomplished during or im-
mediately following data collection, outliers
8-15
-------
may be identified and spurious data may be
corrected.
Many basic data verification programs
have been developed for the field portable
data recorders. Extensive verification pro-
grams similar to those used in the DDRP sur-
veys (Papp et al. 1989) were developed for soil
samples collected during the 20/20 pilot study.
8.7.2.2 Remeasurements
Remeasurements of field parameters
are another method of data verification and
are a technique for verifying relatively sub-
jective field observations such as percent
defoliation. Remeasurements can be ac-
complished by the following methods:
• Field crews remeasuring a statistically
valid percentage of their own plots to
estimate within-field crew precision.
• Field crews measuring reference plots.
The data allow a measure of accuracy and
between-crew precision.
• Remeasurements by a check crew visiting
plots previously measured by other crews.
This is another method of estimating crew
precision.
In 1990, each field crew remeasured at
least one of their own plots and a check crew
remeasured a small percentage of plots. The
QA staff will attempt to establish reference
plots in the coming years. Issues such as
confidentiality and multiple visits will be
considered.
8-16
-------
9 LOGISTICS APPROACH
9.1 INTRODUCTION
Implementing a national Forest Health
Monitoring (FHM) Program will require detailed,
comprehensive logistics planning. A logistics
plan will be developed prior to implementing
any operational phases. The plan will assist
in the five operational phases:
1) Field sampling,
2) Sample and data handling/shipping,
3) Sample preparation,
4) Sample analysis, and
5) Sample archive.
The FHM Logistics Plan serves two
purposes: 1) to provide information on the
concept of FHM and detail the responsibilities
for logistics, and 2) to serve as a guide to the
development of regional logistics plans. The
active tense used in this section indicates the
author's opinion about how logistics could be
accomplished in FHM. The concepts within
this section must be reviewed and approved
through the FHM multi-agency process.
9.1.1 Current Status of Logistics
In October 1989, the USDA Forest
Service (FS) and the EMAP-Forests team
started discussing plans for a multi-agency
monitoring program and established national
technical committees (see Section 9.3). The
Logistics National Technical committee met at
Research Triangle Park, North Carolina, in
January 1990 to develop the Joint National
Monitoring Plan. The operational phases were
identified at this meeting and in the following
months, and the logistics plan was developed
and distributed for internal review. The
document is still being developed.
At the time of the January meeting, it
was known that FHM would be implemented
by the FS in the six New England states. The
logistics aspects of this field activity would be
the responsibility of the FS. Much of this
activity was turned over to state forestry
personnel. During the summer of 1990, a pilot
study was conducted to test indicators on 20
sites in New England and 20 sites in Virginia.
The EMAP-Forests logistics team assumed
primary responsibility for this activity.
In 1991, the FS plans to implement
FHM in the New England States, New Jersey,
Delaware, Maryland, Virginia, Georgia, and
Alabama. The FS will be responsible for
logistics. The EMAP-Forests team plans on
additional field work for indicators that are not
fully implementable in the states mentioned
above and will be responsible for the logistics
of this activity.
9.2 LOGISTICS ISSUES
9.2.1 National Planning
The Logistics National Technical
Committee, which is developing a draft of the
National Logistics Plan, has not been
appropriately represented both spatially (FS
regions) or organizationally (Forest Pest
Management, National Forest System) to
capture all the logistics issues. The National
Logistics Plan should be a guidance document
for implementation within the four proposed
mega-regions (see Figure 11.3). The document
should represent the minimum requirements
needed to satisfy the national program.
Therefore, efforts will be made to develop this
document in FY91 and gain acceptance of the
document over all mega-regions.
9.2.2 Agency Responsibility
The success of FHM is dependent on
interagency cooperation with the FS (all
branches) and with other agencies that will
participate in the program such as the Soil
Conservation Service (SCS), National Park
Service (NPS), Bureau of Land Management
9-1
-------
(BLM), U.S Fish and Wildlife Service (FWS),
etc. The roles of the various agencies must be
established and responsibilities must be
defined.
Currently, the FS and the EPA have
been implementing all phases of logistics, the
FS during implementation and the EPA during
pilot testing and demonstrations. However,
each agency brings unique expertise to the
program. In 1991 the National Technical
Committee will review the logistics
implementation phases and decide which
agency could best implement specif ic phases.
9.2.3 Design and Indicators
A number of design and indicator
issues must be resolved before a definitive
logistics plan can be developed and
implemented.
As stated in the field scenario (Section
9.4.1), the FS and the EPA have agreed to
sample all forested (-5000) Tier 1 sites.
However, it has not been resolved whether all
sites or a subset of these sites will be visited
each year. In addition, a decision must be
made on whether to measure indicators
together (full-suite) on a plot or in different
years.
These issues have a significant impact
on the logistics of the program, particularly
staffing, reconnaissance, and procurement
activities. The design and indicator technical
committees will address these issues in 1991.
9.2.4 Resource Integration
Logistics efforts among EMAP
resource groups should be coordinated and
integrated as much as possible to defray
program costs. Shared, regional logistics
centers with permanent warehouse facilities
will aid in this integration. The Boise
Interagency Fire Center (BIFC) may serve as
a model for the future regional logistics
centers. A national logistical support center,
the BIFC is an interagency program with
agreements between the BLM, U.S. Bureau of
Indian Affairs, FS, NPS, National Weather
Service, and FWS. The BLM manages the land
and facilities and is host to the other five
agencies. Similar arrangements need to be
considered for interagency EMAP logistics
centers.
9.3 ORGANIZATIONAL STRUCTURE
Figure 9.1 is an example of the
organizational structure of logistics. This
organizational structure is dependent on FS
and EPA agreement at the Washington,
station, and work-unit level.
9.3.1 National Steering Committee
The National Steering Committee is
comprised of FS personnel from Forest
Inventory and Assessment (FIA), Forest Pest
Management (FPM), FS management, FS state
personnel, and EPA management. The
National Steering Committee, with assistance
from the National Technical Committee,
develops policy 1or the national program.
9.3.2 National Logistics Technical
Committees
The National Logistics Technical
committee consists of state, FS, and EPA
personnel. The technical committee develops
logistics protocols for the national program in
a manner that will provide consistent and
comparable information across the sampling
regions of the United States. The technical
committee is responsible for planning and
scheduling all phases of FHM data collection.
The EPA has identified two individuals
to assist on the National Logistics Technical
Committee. These two individuals will oversee
EPA regional logistics leads. Due to different
9-2
-------
National Steering Committer
Hational LogiBtj.es
Technical Committee
Region 1
Logistics Lead
Region 2
Logistics Lead
Region 3
Logistics Lead
Field
Sampling
Sample/Data]
handling
r
— 1
Sample |
Preparation|
i i
Sample
Analysis
Region 4
Logistics Lead
Sample
Archive
_J L.
Figure 9.1. Partial example of a flow chart for describing logistics staffing.
organizational structures, the FS may need to
represent national interests with more
personnel. Workshops will be held fre-quently
until the National Logistics Plan is completed.
After the plan is completed, semi-annual
meetings could be held to discuss issues
pertaining to implementation of sub-sequent
surveys and to wrap up current year
activities.
The technicalcommittee, with guidance
from the National Steering Committee, will
define the logistical responsibilities of each
agency. At the national level, the roles will be
distributed within the two primary agencies
(EPA/FS). Each agency will determine the
means to accomplish its responsibility (i.e.,
state assistance or subcontracting), then
forward their decision to the executive
committee for final approval.
9.3.3 Regional Logistics Leads
At a regional level, an organized
logistics team will be given responsibility for
certain elements of logistics. The regional
logistics leads will manage all logistics phases
of the program within the guidelines of the
National Logistics Plan.
The difficulty of coordinating the
logistics activities of 30 to 40 field crews
across the nation necessitates the creation of
regional logistics centers. These centers,
which could be established at regional
technical centers, could house regional
program leaders.
9.4 LOGISTICS IMPLEMENTATION
COMPONENTS
Table 9.1 identifies a number of
logistics issues within the five operational
phases (see Section 9.1). Responsibilities for
each of these components will be determined
by project managers and logistics leads.
These elements will be addressed fully in each
of the regional logistics plans prior to
implementation of field activities. This can
only be accomplished through long-range
planning and coordination. Each element is
not necessarily the responsibility of logistics;
however, the logistics plan identifies who is
responsible for completing the activity and
9-3
-------
Table 9.1 EMAP logistics elements for implementation of forests monitoring programs.
1. Planning
2. Review of Logistical Activities
3. staffing
4. procurement fc inventory
5. Reconnaissance
6. Training
7. comnunicationa
who provides a status summary. A timeline or
Gantt chart (see Figure 9.2) will delineate all
milestones or critical path activities. Timing
for most logistics activities is dependent upon
completion of prior logistics operations.
Therefore, the chart will be updated continually
as the schedule changes or as activities are
completed. Revisions will be dated. Sections
9.4.1 - 9.4.14 discuss each of the logistics
elements that are listed in Table 9.1.
9.4.1 Field Operations Scenario for FHM
The following field operations scenario
is presented to demonstrate that the proposed
field activities are logistically feasible within
the allotted timeframe. This scenario
represents only one of many that could be
developed at this time. All FHM personnel
have not agreed on all the assumptions of this
scenario. The scenario is based on field
implementation and pilot tests that were
implemented in 1990 and eight assumptions:
1. Implementation started in 1990 with 263
sites in six New England states. A pilot
study was conducted in New England and
Virginia. Additional regions and pilot
studies will be phased into the program in
subsequent years. The program will be
fully implemented across the nation by
1995.
2. All Tier 1 sites (-5,000) will be sampled for
Tier 2 sampling. It is estimated that one
quarter (1,250) of these sites will be
sampled each year; a complete sampling
cycle will take four years.
8. Contracting
9. safety
10. scheduling
11. QA/QC
12. information Managesent
13. Review/Recoamandations
14. Inventory/storage
3. The distance between sites at the Tier 1
density will be approximately 30 miles.
Travel between plots is estimated at 2 to 3
hours.
4. Site selection does not consider access.
5. Five indicators of forest health will be
sampled at each sampling site (see
Section 3). As new indicators are added or
the additional list is modified, the elements
within the logistics plan must change.
6. Data collection activities at a site are
limited to one day. Including field crew
travel between plots, it is estimated that a
crew could complete three plots in a week.
7. A field crew of five people will be required
to sample a site in a day.
8. Sampling will take place in an index period
that may vary for different sampling
regions. Generally, the index period will
last from June through August.
Based on these assumptions, 30 to 40
field crews will be necessary to complete the
national sampling program. Allowance for
downtime due to weather and other factors
will have to be considered in determining the
actual number of field crews. To organize and
coordinate the activities of the field crews,
four regional logistics centers (1 per mega-
region) will be established.
Under these assumptions, it is the
responsibility of logistics to provide the most
cost and time efficient method to collect data
9-4
-------
while maintaining data quality. Changes in
any of the assumptions could change the
logistics plan.
9.4.2 Planning
Planning is essential to any program
as large and complex as the FHM program. At
a national level, it will be critical that the
phases of logistics that are occurring in each
region be tracked and that the phases are
implemented within critical timeframes. The
national logistics team must also coordinate
with other technical groups to assure that
decisions affecting logistics occur within
appropriate timeframes. The following Gantt
chart (Figure 9.2) is an example of a timeline
the national logistics team would maintain.
Because the program is in its inception, the
national logistics team and the regional
logistics leaders should be identified; they
should meet on a regular basis (possibly every
3 months) to develop the National Logistics
Plan and the regional logistics plans. After the
plans are developed, planning meetings could
occur twice a year.
9.4.3 Staffing
Staffing and personnel requirements
encompass logistics personnel, field crews,
preparation laboratory personnel, office
personnel, training crews, and QA crews. In
this section, the qualifications of each position
will be addressed in terms of hiring. Timelines
will be developed, identifying when resumes
will be reviewed, when interviews will be
conducted, and when positions will be filled.
The organization through which each position
will be hired (FS, EPA, cooperators, or
contractors) will be determined. Determining
how key personnel will be kept on staff during
slack periods of the program will also be
discussed.
9.4.3.1 Logistics Personnel
Logistics personnel will provide
support in the following areas:
• Equipment and consumable procurement,
storage, maintenance, and repair.
• Vehicle procurement, maintenance, and
repair.
• Sample storage, tracking, packing, and
transfer.
• Lodging, timekeeping, etc.
The primary objective of field logistics
is to keep the samplers sampling. Field
logistics will attempt to accomplish any tasks
that would deter a crew from sampling.
Logistics personnel must be familiar
with the objectives of the program and the
requirements of each indicator. Each element
listed in Table 9.1 must be addressed by the
logistics personnel. The position requires an
eye for detail and experience in acquiring
goods and services.
After data and sample collection and
transfer, logistics personnel will be
responsible for the following activities:
• Staffing of preparation or analytical
laboratories.
• Training of preparation or analytical
laboratory staff.
• Scheduling.
• Contracting.
• Data Handling.
• Procurement and inventory.
• Sample storage, tracking, packing, and
transfer.
9.4.3.2 Field Personnel
The following issues need to be
addressed to adequately staff field crews
within a region:
9-5
-------
1-Imi Loqiatic* JLBtivltlei
ror*«t ii«Hh XonItozln?
180 Q.
ITB Q,
2»D Q,
3 NO 0
FUnOINC (THAA)
SELECTION Of IMDICATORS/'KKTHOOS
«A/«C PLANb
SfiLHUTiQN OP SITKS
RrCfi.vN/ACCESf/fUK HT-UP
is^S!
DATA BUNDLING
DPtHATiOBB SBSOBTi
rVUDlH (YBAB f
9BOJ1C? DlBtGH
ISUICTXOW Or mOICATOBt/JOTBQOl
aiLtcrroM or airsa
ntconM/Acciii/Ptor «tr-u»
PR0PAAAT1QIJ/AflALf 819
DATA SADDLING
O9SRATIOH0 RBPOItf
93
pPBeoNNB^/rrAyriMO - ru«oivs (1«a«
PRO/fCT oatjcu
DfucTiOB or ivotCATOM/iarBooa
QA/QC 9LAB0
saurrxo* or sitbs
¦aeon / ACCSM / PLOT
PKB?|JIAT ion/AW JO.T8 ! •
DATA BAHDLXMC
orBBATioiia snorrs
Figure 9.2. FHM/EMAP 3-Year Gantt chart of activities
-------
• Identification of the organization
responsible for staffing the various
activities.
• Determination of the total number of plots.
• Locations of the plots.
• Identification of the measurements to be
made on each plot.
• The levels of education, e^ierience, and
training required for field personnel.
• The standard operating procedures for
each measure and the time required to
carry out each procedure.
• The retention of key personnel.
• Work schedules (overtime, etc.)
• Contingency plans for replacing staff
members, either temporarily or perma-
nently.
• Acquisition of expertise outside EPA/FS
(SCS/universities/consultants)
Figure 9.3 depicts the organizational
structure of the field staff.
Project Manager
The project manager is the national
lead for the field monitoring phase of the
program. This title could be held by a single
individual or by a committee. The project
manager ensures that field monitoring is
accomplished in all regions and information is
properly disseminated.
Regional Project Leader
The regional project leader oversees
field monitoring activities in a particular region.
This individual disseminates information
(progress/problems) upward to the project
manager and downward to the field crew.
This person maintains adequate staff to
assist in different elements of this task (i.e.,
data management, logistics, reconnaissance).
Field Crew Leader
The field crew will be supervised by a
designated crew leader who is a member of
the crew. The crew leader will supervise all
field operations and resolve discrepancies or
issues as needed at the site. The field crew
leader is responsible for:
• Maintaining and revising sampling
schedules and itineraries.
• Assigning duties according to sampling
priorities.
• Ensuring that all sampling protocols are
followed.
• Ensuring proper use and maintenance of
field equipment.
• Maintaining the integrity of the site and
samples collected.
• Reporting problems or difficulties to proper
management staff.
• Returning all field equipment and supplies.
• Maintaining communications activities.
Field Crew
The information obtained from the pilot
study conducted in 1990 (Figure 9.4) indicates
that a five-person crew is needed to sample a
site for the five indicator measurements
presently proposed for sampling. The five-
person field crew will be comprised of:
• 2 Foresters (visual symptoms, growth)
with work-related experience in mensur-
ational-type measurements.
• 1 Soil scientist (soil sampling) who is
familiar with National Soil Survey
characterization and sampling methods.
• 2 Forest technicians (foliar sampling,
vertical vegetation, growth) who are
experienced in tree climbing and foliar
sampling techniques.
9-7
-------
Project Manager
1
1
1
i
Region 1
Project Lead
Region 2
Project Lead
Region 3 | Region 4
Project Lead j Project Lead
I '
i
Field
Sampler
Field
Sampler
Field Crew
Leader
Field
Sair.pl er
Field
Sampler
Figure 9.3. Partial example of a flow chart for describing field staffing.
Indicator | It People
Hours |Tot• brs
Soil Sampling
1
8
8
Foliar Sampling
1
3
3
Vertical Vegetation
"
2
2.5
5
<3ro*rth
2
2
4
Visual Symptona
2
5
10
3D
Figure 9.4. Estimated time requirements for indicator implementation.
As new indicators are developed and
others are replaced, staffing requirements will
be reevaluated.
To accommodate field personnel from
different organizations, a work schedule for
the crew will be developed. Ten-hour work
days are needed. Within the FS and EPA,
there are a number of work schedules that can
be adapted (e.g., four ten-hour days, eight
days on four days off, etc.). The most
efficient schedule will take data quality, crew
efficiency, and program expenditures into
consideration.
9.4.3.3 Preparation Laboratory Personnel
Logistics personnel will recruit
personnel and acquire facilities for the
preparation of soil and foliar samples. The
following issues need to be addressed:
• Identification of the organization
responsible for staffing the various
activities.
• Determination of the total number of
samples each year.
• The levels of education, experience, and
training required for laboratory personnel.
9-8
-------
• The standard operating procedures for
each measurement and the time required
to carry out each procedure.
• The retention of key personnel.
• Work schedules (overtime, etc.).
• Contingency plans for replacing staff
members either temporarily or permanently.
• Acquisition of expertise outside EPA/FS
(SCS/universities/consultants).
The work schedules of the preparation
laboratory staff will conform to the field
sampling schedule; therefore, the preparation
facility could be operational six days a week.
Preparation Laboratory Manager
The preparation laboratory manager is
responsible for maintaining the integrity of all
samples upon their arrival at the laboratory
facility. The laboratory manager must be
knowledgeable in laboratory methods and
procedures and have demonstrated ability to
track large numbers of samples and supervise
laboratory personnel.
Ultimately, the laboratory manager is
responsible for assigning duties according to
the specific project needs. The following
division of responsibilities is tentative and may
be adjusted.
• Coordinates laboratory operations and
time management.
• Communicates with QA manager and QA
representative.
• Communicates with sampling task leaders
and indicator leads.
• Oversees sample receipt and storage.
• Oversees all computer data entry and
evaluation procedures.
• Oversees sample preparation and analysis
activities.
• Organizes analytical samples into batches.
• Tracks all samples during processing.
• Assists other analysts after other duties
are complete.
Sample Preparation Staff
After the program is fully implemented,
adequate staffing will be provided to ensure a
fast and efficient turnaround of samples from
the field to the analytical laboratories. All
personnel must be thoroughly trained in the
protocols and safety procedures by the
laboratory manager before sample processing
begins.
The preparation laboratory staff will
complete the following activities:
• Sample receipt/tracking.
• Sample storage.
• Sample drying.
• Sample analysis.
• Sample disaggregation/sieving.
• Sample homogenization and subsampling.
• Sample batching.
• Sample archiving.
• Data entry, verification, and reporting.
9.4.3.4 Office Personnel
The regional and national programs
will need office personnel for word processing,
travel assistance, timecards, etc. Personnel
requirements, responsibilities, and funding
estimates will be identified.
9.4.3.5 Training Personnel
Before each field season, field crews
and preparation laboratory personnel will be
trained in the sampling, data collection, and
analysis methods. Experienced instructors will
be recruited. Logistics personnel may be
responsible for obtaining these individuals for
training. Training will include:
• Sampling/preparation methods.
• Quality assurance.
• Safety.
• Information management (data entry/
verification).
9-9
-------
• Sample shipping/handling.
• Logistics (equipment procurement,
maintenance, time keeping, etc.).
9.4.3.6 Quality Assurance Crew
QA crews will be visiting a prescribed
number of sites within each region. The
regional logistics leads will brief the QA crew
on site location, crew sampling schedules and
site completions, lodging accommodations,
and equipment/vehicle procurement.
9.4.4 Procurement and Inventory Control
The success of any survey depends on
appropriate equipment, supplies, and services
being supplied on time and at adequate levels.
The appropriate methods for enumerating
supplies and functional equipment on hand,
assessing future needs, and ordering and
restocking replacement supplies and
equipment on a timely basis will be addressed.
The specific equipment and support needed to
satisfy each of the categories listed in Table
9.2 and the process by which equipment will
be procured will be determined. Close
coordination with design and indicator teams
to identify the equipment, supply, and service
requirements is essential. Procurement
schedules will be tracked very closely and
included in Gantt charts.
9.4.5 Reconnaissance
Reconnaissance of base sites and
sampling sites can reduce the time and effort
of field sampling.
9.4.5.1 Base Site Reconnaissance
Base sites may be necessary to assist
sampling personnel by supplying consumable
supplies, acquiring equipment, shipping
samples, and assisting field crews whenever
necessary.
Selection of base sites will be founded
upon the proximity of sampling sites to towns
or cities; the capabilities of towns or cities to
Table 8.2. List of supply needs for logistics.
1. SCIENTIFIC INSTRUMENTATION
a. measurement devices
b. recording devicss/data forms/log books
c. power sources
d. calibration gear
e. maintenance/repair gear
2. SAMPLING EQUIPMENT
a. containers
b. labels and markers
c. data forms/log books
d. collection devices
e. preservatives
f. shipping containers and accessories
3. SAFETY EQUIPMENT
a. clothing
b. communication
c. flotation
d. first aid
4. TRANSPORTATION
a. vehicles
b. canoes
c. maintenance gear
5. COMMUNICATION
a. radio
b. telephone
c. computer
d. facsimile
6. ADMINISTRATION
a. photocopier
b. forms (e.g., time cards)
9-10
-------
support base sites will also be a deciding
factor (see Table 9.3). These base sites may
be fixBd or mobile, depending on logistical
requirements. Geographical information
systems (GIS) can provide digital line graph
(DLG) maps overlaid with locations of support
services. The list of potential base sites can
be narrowed by using reconnaissance
activities.
Base site operations require specific
utilities. Telephone lines will probably be
essential for satisfying the communications
plan. If telephones are unavailable at remote
locations, alternate methods such as radio
communication will be considered. There must
be access to the appropriate fuels for access
vehicles.
ITiere must be adequate space for
calibrating field instruments and preparing
samples. Separate storage space is
necessary for equipment, reagents, samples,
and wastes. This may require climate
controlled environments such as refrigerators
or freezers. Security of storage areas will be
addressed in the base site plan.
If chemical or biological wastes are
generated, the base site must be located
either within the range of a qualified shipper or
have local facilities available to handle these
wastes.
Shipping facilities are necessary for
movement of samples, supplies, and mail.
Pickup and delivery service will match the daily
and weekly schedule dictated by the sampling
methodology. If the maximum holding time for
samples is minimal, then overnight service will
be available.
Local service and supply stores such
as hardware stores, sporting goods stores,
and auto service centers will be considered in
the selection of base sites. There should also
be personnel support services in the vicinity.
These include lodging, food, banking, and mail
services. If the same base site is used for
more than several weeks, hotel accommoda-
tions may become expensive and confining.
Leased homes can provide a good alternative.
9.4.5.2 Sampling Site Reconnaissance
The site access plan will address how
reconnaissance information about a site will
be collected and how written access
permission will be obtained. It will describe
how and when appropriate government
agencies will be contacted to obtain permits
Table 9.3. Base sita technical support requirements.
I. PROXIMITY TO SAMPLING SITES
II. UTILITIES
A. Phone
B. Fuel
C. Electric
D. Water
III. SPACE
A. Sample preparation/analysis
B. Storage
1. equipment/reagents
2. samples
3. wastes
IV. WASTE DISPOSAL FACILITIES
V. SHIPPING FACILITIES
A. Samples
1. pickup and delivery
2. overnight shipment
(if methods require)
3. high volume carrier
9-11
-------
and site information. Important site data
needed from these agencies include land
ownership information and physical access
information. Other information in the access
plan includes locations of the nearest
emergency services and the types of physical
or biological hazards near the sampling site.
A person will be identified to gather this
information and to disseminate it to the
personnel who will visit the site. Figure 9.5 is
a flowchart of an access plan.
All sites should be visited before field
sampling. If possible, access into the plots
will be marked for field crews. If this is
impossible, detailed sketch maps will be
drawn and obvious starting points will be
identified. Global positioning systems will be
tested to develop this capability. Sampling
sites identified as having potentially difficult
physical or legal access will be identified. If
the access problem is physical, additional
resources (e.g., addition of a crew member or
alternative access vehicle) required to obtain
samples from the site will be identified. Field
crews will be notified of such cases and
allocated additional time and resources for
sampling these sites. If access is legally
denied, the sites will be reported to the design
team. The design team will determine if the
site will be dropped or if an alternative site will
be selected. Field crews will be given a copy
of each site dossier pertaining to their
sampling sites in order to contact landowners
prior to visits and to access the sites
appropriately.
9.4.6 Training
Training is essential to the success of
data collection activities. Training enables
personnel to complete each aspect of
operations according to design and
management objectives and in a standardized
manner. Training will include practice in
standard operating procedures (SOPs) as
documented in a field training and operations
manual. This manual will not be included in
the logistics plan, but will be developed as a
separate implementation document. The
manual will include protocols for measure-
ments, sample collection, sample handling and
processing, sample shipment, data recording,
associated QA and quality control (QC) issues,
safety issues, communications, and preven-
tative maintenance. At the end of training
sessions and at debriefings at the end of
surveys, the manual will be reviewed and a
formal questionnaire for each section will used
to document changes that are needed in the
SOP for future surveys.
A training program plan will be
developed prior to the start of any field or
laboratory operation. Training will include
practice with each of the SOPs. Specialized
training such as safety training, training for
leadership personnel, or instruction on
instrument operation or maintenance will be
addressed. If outside organizations are
needed for any aspect of training (for example,
the American Red Cross for first aid/cardio-
pulmonary resuscitation (CPR) instruction),
their services will be arranged. Participants
will be evaluated for competency following
training.
As the program expands to new
regions, many trainers will be needed to train
all the field crews. Pre-training sessions,
which will be designed to maintain consistency
across and within regions will be needed to
"train trainers". Logistics personnel will be
responsible for locating training sites and
facilitating training sessions.
9.4.7 Communications
The logistics plan will establish
efficient communications methods to ensure
smooth operation of field sampling, laboratory
analyses, and data and sample tracking activ-
9-12
-------
FS, EPA, etate and county agencies
telephone or
letter request
ownership information
permit information
physical access information
Access Coordinator
telephone
request with
written
follow-up
I Landowners and/or appropriate agencies
' 1
Dossier of sits access information
dossier used
update dossier
for site
upon reconnsiiianct
reconnaissance
I Reconnaissance personnel/Field crews |
I I
written access permission,
perziite,
physical access information
written
access
permission(
permits,
physical
access
information
Figure 9.5. Example of a sampling site access plan.
ities. All communications lines required for
science, logistics, and safety (Figure 9.6) will
be determined. If working with other
agencies, additional communications lines to
disseminate information to these groups will
be needed.
The communications plan will describe
methods for tracking sample shipments to lab-
oratories and for sending and tracking data.
It will also provide a mechanism for field
crews to acquire information from labora-
tories, data management staff, and QA staff
about problems with data collection or sample
handling so that the problems will not recur.
9.4.7.1 Line of Communication
The basic line of communication is
illustrated as follows:
j Project manager
I '
| Regional project leader
Field crew leader
Field crew personnel
Project Managers
Project managers are responsible for
the dissemination of information vital to the
project (i.e., protocol changes, sampling
schedule changes, etc.) and will require
progress reports on all aspects of the project.
9-13
-------
PROJECT MANAGEMENT
News
releasee
MEDIA
d
Directions
Updates
1
•
Emergency
calls
COMMUNICATIONS CENTER (perhaps 2 4 hr)
Access
requests
1) Supply requests
2) Sa»ple/data
tracking info.
3) Daily activity,
and plans
4) Administrative
requests
5) Emergency needs
1) Supply status
2) Problems found
with samples/data
3) Activities of
other base sites
4) Administrative
requests
BASE SITES
Visit
notification
|Emergency calls
POLICE, AMBULANCE, PIRE
LAND
OWNERS
Figure 9.6. Example of a communications plan.
Regional Project Leads
The regional project lead is responsible
for relaying information to the project man-
agers and other technical support leads and
from field crew leaders (Figure 9.7). The
regional lead disseminates information back
to these groups and may need to contact land
owners or emergency services. Regional
project leads will be available for phone or
emergency communication during sampling
hours. However, a 24-hour line of
communication with the regional project lead
must be established. This could be
accomplished by using phone recorders,
electronic mail, or G1S capabilities.
Field Crew Leader
The field crew leader must inform
regional project leaders about sampling
progress and problems or emergencies that
occur in the field. Field crew leaders are
responsible for the direct communication of
emergencies to the appropriate authorities. If
injured, their duty will be transferred to one of
the field crew members. The field crew leader
will submit weekly travel itineraries to the
regional project leaders and is responsible for
disseminating information to field crew
personnel (i.e., status of sample shipments,
data discrepancies, supply disposition, etc.).
Field Crew Personnel
Field crew personnel are responsible
for their sampling assignments. They must
report to the field crew leader to establish an
efficient relay of information on progress,
problems, or emergencies occurring in the
field.
Conference Calls
As illustrated in Figure 9.7, the regional
project leaders link the field crew leader with
project managers, other technical leads, and
various groups. As problems occur in the field
or as protocols change, it is important that
decisions are consistent for all field crews and
regions. Therefore, a weekly conference call
will enable technical leads, regional leads, and
9-14
-------
| PROJECT MANAGERS |
I '
MEDIA
LOGISTICS
H LANDOWNERS |
H INDICATORS {
I I
EMERGENCIES
REGIONAL PROJECT LEAD
DESIGN
PREP LAB
CONTRACTORS
OA
CIS
PIELD CREW LEADERS
Figure 9.7 Flow of information to and from regional project leads.
project managers to discuss progress on all
operational phases, problems, and protocol
changes.
9.4.7.2 Geographical Information Systems
(GIS)
"The GIS group can assist the project
by locating facilities and services that may be
necessary for field crews, including hardware
stores, hotels, express mail, automotive repair
shops, hospitals, and fire stations.
The GIS group has used a software
package called Business LINE® which provides
business listings and summary reports for 7.6
million business establishments throughout
the United States. Latitudes, longitudes,
geocodes, addresses, and telephone numbers
are included in this data base. Maps of each
hexagon or state can be produced that include
location of important establishments.
9.4.8 Contracting
Many operations within the FHM
program will require some form of contracting.
Contracting may be required for phases of
field sampling, sample preparation, and
sample analysis.
9.4.8.1 Contracting Mechanisms
The following types of contracting
mechanisms can be used to acquire services
for EMAP-Forests:
• Federal government acquisition.
• Fixed price contract (IFB).
• Competition negotiation (RFP).
• Special analytical services.
• Interagency agreements.
• Cooperative agreements.
• Subcontracts.
9.4.9 Safety
The safety plan will consider preventive
safety measures and emergency action
procedures.
Clothing and other equipment will be
discussed in the safety plan and criteria used
for selection of safety gear for field personnel
will be explained. Other sections of the
logistics plan (training, waste disposal) will
outline additional preventive safety measures.
9-15
-------
Information on field personnel and
their travel itineraries (Table 9.4) will be
compiled. A person responsible for
maintaining this information will be identified.
Emergency action plans will be
developed. The American Red Cross principles
of first aid/CPR will be used as a guideline for
the initial treatment and evaluation of
personnel in emergencies. Criteria and
methods will be developed for initiating search
and rescue operations. The communications
plan will indicate who will be contacted during
emergencies.
Table 9.4. Safety information to be
logged by field personnel.
Travel Itinerary
roads to be travelled, order
flight plan
coordinates expected to be visited, order
time left base site
estimated time of return
Medical information
known allergies
existing conditions (eg., heart disease, diabetes)
Personal contacts (i.e., immediate family)
addresses
telephone numbers
Personal descriptions
color and types of clothing worn
height and weight
hair, eye, and skin color
age
vehicle used and its description
9.4.10 Scheduling
The sampling design task group will
provide various input parameters: the
geographical area to be studied, the sampling
site locations, and the sampling index period.
The methods task group will provide other
required information such as the sampling and
transportation equipment to be used and the
required sample holding times. Based on the
design and methods requirements, an efficient
sampling schedule will be developed.
Geography will be considered when
preparing the sampling schedule. The
locations of sampling sites in relation to each
other and to other points of interest will
determine how much time and fuel will be
required to travel to and from sampling sites.
The distribution of sampling sites relative to
refueling stations, base sites, and courier
services will be determined.
"Down time" will be accounted for in
the schedule by considering typical climatolog-
ical conditions for the area during the
sampling window. Precipitation, cloud cover,
temperature, and winds can affect the quality
of samples and sampling site accessibility.
The difficulty of site access will also
be of concern. Physical constraints include
mountains, brush, soft substrate (mud or
marsh), and the lack of paved roadways.
Legal constraints include lack of access
permission and any conditions imposed by
landowners. Parklands, wilderness areas, or
publicly owned water supplies may forbid
motorized access. Military reservations may
be restricted or require escorts.
When the above factors are examined,
a list of potential schedules and potential
base sites can be created (Figure 9.8).
Schedules will also be developed for
other data collection activities such as sample
preparation and laboratory analysis. Volume
of samples and turnaround time for reports
will affect schedules and must be determined
prior to field activities to secure proper
facilities to handle the activities.
9.4.11 Quality Assurance/Quality Control
In accordance with the QA plan, there
will be regular site audits to assure that field
9-16
-------
personnel are operating according to protocols
established by the Field Training and
Operations Manual. The logistics plan will
indicate who will audit and when audits will
occur. If QA objectives require Held crews to
be unaware of impending audits, then the
logistics plan will provide windows of
appropriate times for audits. The logistics
r - - - i
EXPERIMENTAL I
DESIGN ,
the entry programs and how they will be set
up on hardware.
Data transfer mechanisms from
temporary field data bases to the central data
base will be developed (see Section 10.3.1.4).
The plan will indicate when data will be sent,
how often it will be sent, and what security
rv
GEOGRAPHY
LIST OF POTENTIAL
BASE SITES
VARIOUS SAMPLING
SCHEDULE SCENARIOS
CONSTRAINTS TO
ACCESS
(LEGAL, PHYSICAL)
SAMPLING SCHEDULE
+
CONTINGENCIES
RECONNAISSANCE
Figure 9.8. Development of a sampling schedule.
plan will also describe how audit comments
will be addressed and provide a mechanism to
correct protocols.
9.4.12 Information Management
Four aspects of information manage-
ment will be discussed within the logistics
plan:
1) Data recording.
2) Data transfer.
3) Data security.
4) QA.
The data management plan will
describe a standard method for recording data
(see Section 10.3.1.3.2). The logistics plan will
address who will record the data. It will also
address who will develop the forms and how
the forms will be printed. If software Is to be
used, the plan will address who will develop
measures will be taken to ensure that data
will not be lost (i.e., properly backing-up data).
The verification procedures will be
described. The data review process will be
established. This process includes how and
when the data will be reviewed and how the
reviewer(s) will signify that the review has
been completed.
A computerized barcode sample
tracking system will be developed (see Section
10.3.1.5). This will make sample tracking and
chain-of-custody more efficient and reliable. A
similar system will be used to inventory
equipment and consumables.
9.4.13 Review/Recommendations
The logistics activities will be sum-
marized each year of the study. Discussions
9-17
-------
will include debriefing sessions, resolution of
problems, and planning activities for the
following year.
9.4.14 Inventory/Storage
After an operational phase is com-
pleted, the equipment and supplies used
during the phase will be inventoried and
examined for damage. Bar code readers, used
for sample tracking, will be used to inventory
and track equipment. Equipment and
consumables will be ordered for the next
year's activities and stored at regional
locations.
9-18
-------
10 STRATEGY FOR THE INFORMATION
MANAGEMENT SYSTEM
This chapter describes the information
management (IM) system for a National
Forest Health Monitoring (FHM) program that
is achievable five years from now. The current
IM system is embedded in the description of
the future IM system. The steps necessary to
move from the present system to the future
are presented in this section. The level of
detail reflects the level of uncertainty
concerning the future direction of FHM and the
future of technology. Flexibility is a key
concept in the IM system. An IM system that
can not or will not adapt to change will be
obsolete before it is implemented.
10.1 INTRODUCTION TO INFORMATION
MANAGEMENT
Information management supports and
facilitates many aspects of environmental
monitoring. The IM personnel work with the
technical directors, project managers, logistics
staff, quality assurance/quality control
(QA/QC) personnel and scientists throughout
the FHM project. This starts with planning and
coordination to assure an IM system that is
responsive to overall project needs. During
implementation and the operational phases of
data collection and transfer, software systems
will be in place to support the timely
acquisition of data into the IM system. After
data collection, IM supports the scientists
working on integration and analysis of data
and presentation and reporting of results. IM
will also support the dissemination of data
and information to users outside of the FHM
program.
The FHM IM system will be distributed
with nodes on both EPA and FS computer
networks. A user will be able to connect to
the FHM data base system and access data
without knowledge of the location of those
data. Eventually this networked IM system
will include computer networks from other
agencies participating in FHM such as the
National Park Service (NPS), Bureau of Land
Management (BLM), states, and universities.
A key element in the FHM IM system is
the Forest Information Center (FIC). The FIC,
staffed by personnel from all agencies
participating in FHM, is the nexus for software
development, data collection (both FHM-
generated and historical data), data
cataloging, data processing, and data
dissemination. The FIC staff will work with
appropriate personnel in the FS and the EPA
to assure that the automated data processing
(ADP) requirements of FHM are met.
10.2 GOALS AND OBJECTIVES
The design and development of the IM
program is guided by the following goals:
• Assure that the data in the system are of
the highest possible quality.
• Assure that FHM scientists have access to
the data as quickly as possible.
• Make the data available to users both
within the project and outside the FHM
group.
To achieve the above goals, an IM
program will be developed to meet the
following objectives:
• Design an IM program to be responsive to
user requirements from within and outside
the FHM program.
• Commit to achieving complete data
collection and transfer electronically.
• Ensure access to FHM-generated data,
auxiliary data, and historical data.
• Provide an IM program that effectively
collects, processes, documents, stores,
catalogs, and distributes the FHM data
within accepted time frames.
• Develop a flexible IM program that can
adapt to the program's future needs.
10-1
-------
• Develop an integrated 1M program that
provides access to GIS systems, other
EMAP and FS monitoring components, and
other programs.
• Develop a system that is responsive to the
needs of the national FHM program, but is
flexible enough to accommodate regional
differences.
• Provide training and support to the field
crews and users of the FHM IM system.
10.3 DESIGN OF THE FHM IM SYSTEM
The IM system for the FHM program
will have two major components: (1) a field
and laboratory data collection system, and (2)
a data management system. The field and
laboratory system handles data coming into
the F1C. The data management system
handles data in the FIC and distributes data
to the users.
10.3.1 Field and Laboratory Systems
The field and laboratory systems
provide input to the FHM FIC. These systems
have close ties to the cross-cutting activities
of QA/QC and logistics. The objectives of the
field and laboratory systems are twofold: (1)
to develop a system to ensure that
Measurement Quality Objectives (MQOs) are
satisfied, and (2) to ensure that data are sent
to the FIC in a timely manner. This mandates
electronic data collection in the field and the
laboratory and electronic data transfer to and
from the FIC. These systems must be flexible
enough to accommodate changes in data
requirements, indicators, and technology.
Verification checks are placed as close
to the point of data entry as possible in the
field and the laboratory. Close cooperation
with the QA staff will be essential in the
development of the computerized verification
checks.
Electronic sample, shipment, and crew
tracking will be used to give project managers
daily updates of field and laboratory activities.
These tracking systems will be developed in
conjunction with the logistics staff.
10.3.1.1 Field Crew Hardware and Software
Field crew equipment will include
portable data recorders (PDRs), laptop
computers, portable printers, global
positioning system (GPS) hardware, and bar
code readers. The first three items in the list
were used in the 1990 field season. Except for
the laptop and the printer, which remain in the
motel room, all the equipment listed is used in
the field.
10.3.1.2 Field Logistics Data Base (not
implemented in 1990)
A logistics data base with a user-
friendly interface will be installed on the laptop
computers to simplify field logistics. Informa-
tion describing sample site locations and log-
istics information will be entered into a
geographic information system (GIS) data
base. The GIS system will produce maps
showing the locations of sample sites and
support services. With these data, the crew
will easily be able to locate sample sites,
express mail facilities, motels, airports,
hospitals, repair centers, etc. Sampling site
information will include location of the site,
location of the starting point, field
measurements to be taken, and samples to be
collected.
10.3.1.3 PDR Programs
The EMAP-Forests indicators are
dependent on field measurements such as
forest mensuration data, pedon descriptions,
and visual damage data. To ensure that field
10-2
-------
data are of the highest quality possible, EMAP-
Forests will be committed to electronic data
collection.
To facilitate electronic data collection,
each field crew will have one or more PDRs.
The PDR is a rugged field computer. The PDR
currently used by EMAP-Forests is MS-DOS
compatible, which allows for flexibility in
programming. Custom software, written in C
and BASIC, was developed for the PDR for
use in the 1990 field season. The current
software will be refined and new programs will
be developed for the PDR as the FHM program
continues. The PDR programs will include the
GPSdata collection programs, sample tracking
information, and communications. A user-
friendly menu will allow the crew to choose the
appropriate program. Sections 10.3.1.3.1-
10.3.1.3.3 provide details about the programs
envisioned for use on the PDR.
10.3.1.3.1 Global Positioning System (not
implemented in 1990)
The GPS will interface with software
on the PDR, allowing the crew to determine
their field position within tens of meters.
Software will be written to utilize the GPS data
to guide the crew to the plot. This system will
help crews find new as well as established
plots.
10.3.1.3.2 Field Data Collection Programs
(implemented in 1990)
Data entry will be performed directly on
the PDR in the field. Paper forms will be used
only for back-up in case the PDR fails in the
field. A spare set of PDRs will be available;
these can be shipped via express mail to a
crew within 24 hours.
The PDR will have various data
collection programs. Menu choices, based on
the data requirements of the current indi-
cators, will include soil pedon descriptions,
forest mensuration (including visual damage
data), vertical vegetation profile, and
ceptometer data transfer.
If, for example, the user chooses the
forest mensuration data collection program,
an electronic tally sheet will be displayed on
the PDR screen.
Using electronic data entry allows for
QA checks at the point of data entry. These
QA checks, which include range checks,
validity checks, and logic checks, will be
designed in close cooperation with the QA
staff and the indicator leads. Data from
previous years will be loaded on the PDR for
further QA checks. For example, the user will
be notified if the diameter of a tree is
significantly smaller this year than it was in a
previous survey. Additionally, data from
previous surveys can help the field crews
locate specific trees. The distance and
direction to a tree will ensure that the same
tree is sampled in all surveys, a requirement
for some indicators. This feature was not
implemented in 1990.
10.3.1.3.3 Sample Tracking on PDR (not
implemented in 1990)
Many types of samples will be
collected in the field. Currently, these include
soil, root, foliar, and increment cores. The field
crews must be sure that all necessary
samples are collected and that samples are
correctly identified and tracked. A bar-coding
system will link data on the PDR to samples
collected in the field. This will permit a
relational join between the sample ID and data
in the PDR. The system will check that all
samples have been collected before the crew
leaves the field. Sample tracking is described
in more detail in Section 10.3.1.5.
10-3
-------
10.3.1.4 Field Communications System (partial-
ly implemented in 1990)
When fully implemented, the
communications systems will allow for two-
way communications between field crews and
the FHM FIC. Data and tracking information
will be uploaded from the crews to the FIC.
Messages, data, and program updates will be
sent from the FIC to the crews. In 1990
communications were unidirectional, from field
crews to the FIC.
10.3.1.5 Computerized Shipment Tracking (not
implemented in 1990)
"The field crews will collect a plethora
of samples, many of which are perishable,
requiring proper handling and quick shipment
to the laboratory. A computerized sample and
shipment tracking system that utilizes bar
codes is necessary to ensure that samples get
to the proper laboratory in a timely manner.
The field crews will have pre-printed sample
labels with bar codes. When a sample is
collected, data about the sample will be
entered in the PDR. The sample will be
labeled, the bar code scanned, and the sample
number recorded on the PDR. Before leaving
the field, a program on the PDR will check that
all samples have been collected.
When the data from the PDR are
uploaded to the laptop, the sample tracking
data base on the laptop will automatically be
updated. The crew will use the bar code
reader to scan the samples as they are
packing the shipment cases. The system will:
• Ensure that the correct samples are
packed together.
• Ensure that samples are shipped to the
correct laboratory.
• Check that all samples have been shipped.
• Provide information about special handling
required.
After all samples are ready for
shipment, the crew will enter data about the
shipment on the laptop. These data include:
shipment number, carrier name, air bill number,
destination laboratory, and estimated time of
arrival at the laboratory. These data are
entered into the tracking data base which is
sent to the FIC and then to the receiving
facility.
This communication system also
allows crew tracking. A crew tracking data
base, including location of crews, locations
sampled, data collected, samples collected,
and shipments sent, will be updated daily.
10.3.2 Laboratory Systems (partially imple-
mented in 1990)
The FHM program will employ a variety
of laboratories for processing different sample
types. Computerized laboratory sample
tracking, verification, and communications
systems will be used by the laboratories
employed by the FHM program. The FHM
program will have two types of laboratories,
preparatory and analytical. This section
describes the components common to both
laboratory types.
Each laboratory will have an IBM-
compatible computer with a modem and bar
code reader. The FHM laboratory system
software will be installed on the computer.
The tracking portion of the system will
interface with the tracking system described
above to create a complete sample trail from
field to laboratory. The verification portion of
the program ensures that results from the
laboratory meet the quality standards of the
FHM program.
The communications are similar to the
field system. The laboratory will send the
following information to the FIC via modem:
10-4
-------
• Results, including QA/QC data, since last
upload.
• Samples received at the laboratory.
• Samples shipped from the laboratory (for
preparatory laboratories only).
• Messages from the laboratory to the
central system.
• Tracking data.
The following information will be sent
from the FIC to the laboratory:
• The tracking data base.
• Software updates, when required.
• Messages from the FIC to the laboratory.
Each laboratory will have a bar code
reader. As shipments arrive at the laboratory,
the bar code label on each sample will be
scanned. Those data will be compared
against the tracking data base that was
downloaded from the FIC.
10.3.2.1 Preparatory Laboratory Systems
(partially implemented in 1990)
Preparatory laboratories receive field
samples, process the samples, then ship the
samples to analytical laboratories. A data
base will be maintained based on the
information entered in the preparatory
laboratory. The data base will relate batch
numbers to sample numbers and will record
data that describes archived samples.
10.3.3 Data Management System
This data management section
envisions a five-year scenario. This scenario
includes the assumptions that a high speed
connection between the EPA and FS computer
networks and a relational data base
management system (RDBMS) that is
compatible between the two agencies are in
place.
The core of the distributed FHM data
management system is the FHM FIC. As
referenced in this document, the FIC is a
logical concept; the physical structure of the
FIC will be determined after detailed design
work is completed. The FIC will be staffed by
both EPA and FS personnel and will support
the exchange of data with other agencies and
organizations. Information management
personnel are responsible for maintaining a
comprehensive data inventory, data set index,
code libraries, and data dictionary. They will
also maintain and disseminate FHM data and
ensure that appropriate data are incorporated
into the FIC.
10.3.3.1 Data Types
The FHM IM system will contain data
generated by the FHM program and data from
outside sources. The following types of data
will be maintained by the FIC:
• Project management and logistics data.
• Raw data files.
• Summarized data.
• QA/QC data.
• Laboratory data and associated QA/QC
data.
• Spatial data in GIS format.
• Historical data.
• Pointers to auxiliary data (e.g., climate
data).
10.3.3.2 Users
Users of FHM data will include the
following four groups.
Group I Users: FHM Core Group -
Responsible for day-to-day field operations
and data verification and validation activities.
The group wilt include field crews, logistics
staff, QA/QC staff, IM staff, indicator leads,
and the technical directors of the FHM
program. Both FS and EPA staff are in this
group.
10-5
-------
Requirements - This group will need to
have access to a comprehensive data set,
including project management information,
sample and shipment tracking, raw data files,
QA/QC reports, logistics, summary reports,
and verified and validated data sets.
Timing of Access - This group will
require access to the data on a real time
basis. The data need not be quality assured
prior to access. All raw data used by this
group must be used with the understanding
that the data have not been verified or
validated. This group needs access to all data
described in the other categories.
Group II Users: FHM Team -
Individuals and groups who will participate in
the FHM effort but will not be active in the
day-to-day operations of the field programs or
the data verification and validation processes.
These participants will include FHM staff
involved in reporting, the FHM Integration and
Analysis Team, GIS support personnel, FHM
design and statistical staff, and program
reviewers.
Requirements - This group will require
access to summary information regarding
logistics, project management, and QA/QC.
They will also require access to some
validated and verified raw data files but will
not require real time access to the data.
Timing of Access - Group II users will
require data one month from the time of
collection.
Group III Users: Inter-Agency
Research Group - Includes all researchers who
will be active in the design, implementation,
and analysis of the national EMAP program,
the other FS-FHM groups, and scientists from
other participating agencies. These individuals
will include members of other EMAP resource
groups, EMAP cross-cutting groups, the FS
evaluation monitoring team, and the FS
research monitoring team.
Requirements - This group will require
final summaries regarding logistics, project
management, and QA/QC. They will require
access to some validated and verified raw
data files. Document summaries with
interpretation and graphic outputs will be most
useful.
Timing of Access - Group III users will
require data approximately six months from
the time of data collection.
Group IV Users: Other Users -
Includes all potential users outside of those
listed above. This group will include state and
federal agencies, universities, research
organizations, citizen's groups, administrators,
and legislators.
Requirements - This group will require
access to validated and verified data including
QA/QC data that is integrated to the plot level.
They will need summarized characterization
data for each plot sampled and access to an
index of available data. They will also require
access to some validated and verified raw
data files. Document summaries with
interpretation and graphic outputs will be most
useful.
Timing of Access - Group IV users will
require data one year from data collection.
10.3.3.3 Data Base Access
Users on either the EPA or FS
computer networks will be able to access the
FHM IM system directly through the networks.
Users who are off the network can access the
system through a dial-up line into the system.
Users in Groups I and II (see Section 10.3.3.2)
will access the data through the FHM IM
system. Users in Groups 111 and IV may have
10-6
-------
the option of accessing the FHM IM system,
but it is more likely that they will use the
EMAP-wide EIC.
A user-friendly interface will guide the
user to required data. A data catalog and a
data dictionary will detail the data available
through the data base system. For wide
distribution of the data, use of a commercial
service such as Compuserve will be explored.
At periodic intervals, the data in the data
bases will be published on CD-ROM.
10.3.3.4 Data Base Security
The four user groups will have different
access privileges to the data bases. Until the
data have been verified and validated, very
strict security measures will be employed.
Only members of Group I, the core group, will
have access to raw data from the field and
the laboratories and project management
data. During the QA/QC process, only the IM
staff will be allowed to change the data
bases. If discrepancies are found during the
QA checks, those data will be communicated
to the IM staff. The IM staff will update the
data bases and record the change, the person
requesting the change, and the reason for the
change in a data base. This is to ensure that
there is only one official version of the data
base that is maintained by the IM staff.
After the data bases have passed
QA/QC, the security will be changed so that
members of Group II (the FHM analysts) will
have access to the data. At this point,
members of Group III can have access to the
data with permission of the TD. After the
yearly statistical summaries have been
published, the data will be made available to
other users. At this point, the FHM data will
be made available to the EMAP-wide EIC.
10.3.3.5 Data Confidentiality
Certain types of data, both FHM
collected and from external sources, may have
to remain confidential. Locational data are the
most likely candidates for confidentiality.
These data include FHM plot location, location
of plots in other data bases used by the FHM
program (e.g. FIA plot locations), and
locations of rare and endangered species.
The GIS representations of point data
will be "fuzzed" to hide the exact locations of
plots, or the data will be represented on a
regional basis to hide the exact plot locations.
The locational data in the public data base will
be reported at the Tier 1 hexagon center level.
Analysts outside of Group III who need exact
locational data will need written permission
from the senior administrators of the FHM
program and will be required to sign a non-
disclosure document.
10.3.3.6 Data Base Management System
The FHM data base management
system will include a data set index (DS1) also
known as a data catalog, a data dictionary,
code look-up tables, and a user-friendly
interface. An RDBMS will be the engine of the
system.
The DSI will provide users with important
information about the contents of each data
set. It will also describe how to access a
particular data set. The DSI will also store a
catalog of FHM-generated, historical, and
auxiliary data.
The on-line data dictionary will provide
users with information about parameters
stored in the data bases.
10-7
-------
10.3.3.7 Yearly Statistical Summaries
Standardized, yearly, data statistical
summaries will be one product of the FHM
program. Standard software will be developed
to automatically produce the tables, graphs,
and maps that go into the yearly statistical
summaries.
10.3.3.8 GIS Interface
A major requirement of the FHM FIC
will be to create maps and perform
geographically-based analyses. Therefore, the
data generated for FHM will be referenced to
a spatial entity such as a latitude and
longitude. Spatial analyses will be
accomplished using ARC/INFO, a GIS that is
used throughout the EPA and the FS.
ARC/INFO is not user-friendly. Therefore,
user-friendly interfaces for routine data
analysis and display will be developed by the
FIC.
10.3.3.9 EMAP Information Center
The EIC will be the entry point to EMAP
data bases. The EIC will allow users to
access data from the EMAP resource groups
and cross-cutting activities.
For the overall EMAP goals to be met,
scientists must have access to all data
collected in connection with EMAP data,
including FHM data. The design of the FHM
IM system must be compatible with the EIC
design to allow other EIC users access to the
data.
10.4 STRATEGY TO MOVE TOWARD THE
FHM IM SYSTEM
This section outlines what is needed to
make the preceding vision of the FHM IM
system a reality. Most importantly, the plan
for the IM program must undergo continual
review and update.
10.4.1 Recognition of the FHM as a New and
Different Program
All participants must recognize the
FHM as a new program. It may have its roots
in other programs, such as FIA, FPM, Forest
Response Program (FRP), and the
Direct/Delayed Response Project (DDRP), but
FHM is fundamentally a new program with a
new set of goals and objectives. The program
should borrow the good points from its
ancestors, but it should look beyond its
antecedents to determine its own future. This
is especially true of IM. The IM goals of the
FHM program may be best served by
establishing a new data processing program.
10.4.2 Commitment to an Interagency IM
Program
All participating agencies, EPA, FS,
NPS, and others, must make a commitment to
IM. All parties must realize that without good
cooperative IM at all levels, from the field to
data bases, the FHM program will likely fail.
This commitment includes adequate funding,
adequate staffing, and the recognition that
difficult decisions must be made. The
program must recognize that there are no EPA
data and there are no FS data; there are FHM
data that will be shared with all participants in
the program.
The FIC will be the focal point for data
collection and dissemination. The FIC will
receive data from all field locations and make
data available to users of all agencies
concerned after QA/QC is complete. The FIC
will manage data on a system that ties the
EPA and FS computer networks together.
Aphased approach to the development
of the IM program will be needed. The current
level of resources, technology, and interagency
cooperation dictates that for the next several
years each agency will maintain separate data
10-8
-------
base systems. Commitment to and planning
for the FIC should begin immediately.
10.4.3 Public Access to the Data
The FHM program must make a strong
commitment to making the data available to
the public. After the data have been through
the verification and validation processes and
the yearly statistical summaries have been
produced, the data should be available to all
interested users.
10.4.4 Interactions with the EMAP
Information Center (EIC)
EMAP-Forests must adhere to the
requirements of EMAP. One such requirement
is having all EMAP data accessible through the
EIC. Ml participating agencies must agree
with the policy that data collected in
connection with EMAP-Forests will be made
available to the EIC. Release of the data to
the EIC will be subject to data security and
confidentiality constraints (see Section 10.3).
10.4.5 Standards
Standards are necessary for FHM to
be a truly national program. The FHM program
and its IM system must be flexible enough to
accommodate regional differences, but at the
same time be comparable at some level
throughout the country. Standards that are
used throughout the program are necessary to
meet that objective. An interagency work
group should be formed to resolve standards
issues. For example:
« Codes - Standards for codes that are used
across the country, such as species, must
be adopted. The FIA has a standard set of
some codes. It is recommended that
those codes be adopted.
• Computational Algorithms - A standard set
of FHM computational algorithms that
correspond to ecological, not political,
boundaries must be established. Post-
stratification along political boundaries will
always be possible, if required.
• PDRs - PDRs must be standardized to the
extent that all those used by FHM will run
the same programs without modifications.
• PDR Software - The same software should
be used on all the PDRs used by FHM. The
software should be flexible to allow for
regional differences.
• Measurement Units - The FHM should use
the same measurement units, preferably
SI, in all regions of the country.
• Word processing software - A standard
word processing program should be
adopted for producing reports and
documents. If institutional constraints
prohibit this, a standard interchange
format should be adopted.
10.4.6 Data Sharing and Access
All agencies concerned must come to
an agreement on data access. One proposal
for data access is given in section 10.3.3.2 of
this document.
If this model of data sharing is not
acceptable to all participants, an interagency
committee should be formed to draft an
alternative policy. A clearly stated policy on
data access should be adopted for the entire
FHM program.
10.4.7 Staffing
The FHM IM program needs a full time,
quality staff of adequate size. The FIC
minimum staff and their functions include:
• Information Manager - responsible for
system design, management of the
information staff, liaison with other
ecosystems and agencies.
• Systems Programmer(s) - responsible for
development of the software for the field
10-9
-------
systems, communications, and data
analysis.
• Data Base Manager/Program mer(s) •
responsible for data base programming,
assists the Information Manager in data
base design and works with the other staff
in satisfying users data requests.
• Data Clerk(s) - responsible for
documenting FHM data sets including
historical data obtained from other
agencies, responding to data requests
from members of the FHM team and
routinely processing FHM-generated data.
An interagency programming team,
using state-of-the-art programming tools, such
as object-oriented programming, should be
formed to develop flexible programs for FHM.
The FHM program can use existing software
for several more years, but FHM can not afford
to institutionalize software that does not have
the ability to evolve with the program.
10.4.8 Interagency Computer Links
For FHM to function efficiently, there
must be a link between the computer
networks of all participating agencies. These
agencies include the EPA, FS, NPS, BLM, and
possibly others. The link should start with an
EPA/FS connection, then progress to other
agencies. The interagency links will provide
services such as E-Mail capability, file transfer,
and data base access to all participants
across the FHM program. Additionally, links to
other networks such as Bitnet, Internet, and
LTERnet should be explored. Those additional
links will allow easy access to university
cooperators.
10.4.9 User Needs Analysis
The FHM Information System must be
responsive to users' needs. A study must be
undertaken to identify categories of potential
users of FHM data. After users have been
identified, a user needs analysis will be
performed. Users will be queried to determine
the types of data needed, the modes of
access, and the interfaces desired. Once the
users and their needs have been identified, a
revaluation of the user categories presented
in section 10.3.3.2 will be undertaken. The
data compiled in the user needs analysis will
be one input in the design of the FHM 1M
system.
10-10
-------
11 STRATEGY FOR REPORTING
In contrast to the term "assessment",
which refers to the intellectual synthesis and
interpretation of forest environmental data
(Section 7), "reporting" refers to the
mechanical aspects of document scheduling,
production, review, and clearance. This
chapter considers all documents produced by
the interagency Forest Health Monitoring
(FHM) program and focuses on the reports
and roles of EMAP-Forests.
Reporting activities operate simul-
taneously within several programs and must
serve many needs. The reporting strategy of
EMAP-Forests is coordinated by EMAP and is
implemented in cooperation with the Forest
Service (FS) and other agencies that also
produce monitoring reports. Teams of
analysts are comprised of individuals from
several organizations. Success within these
multiple contexts requires cooperation among
agencies and individual participants, and
division of labor is an essential ingredient of
the strategy.
11.1 CURRENT STATUS OF REPORTING
EMAP-Forests operates as a national
reporting unit, with specific laboratory roles
assigned according to the overall EMAP
scheme (i.e., design, statistics, and indicators
at ERL-C; integration, assessment, weather,
air quality, and pollution deposition at AREAL-
RTP; logistics, quality assurance, and
information management at EMSt-LV).
Reports are coordinated with national program
counterparts in the FS. EMAP-Forests does
not have regional reporting capabilities, except
that national staff participate in some regional
reports prepared by FS regions.
In contrast, the FS operates as
regional reporting units, making use of EMAP-
Forests national staff to produce regional
reports. Currently, there is not an identified
national reporting responsibility within the FS.
There is not a formal interagency
agreement concerning the preparation,
publication, and distribution of reports;
however, each agency has identified the types
of reports that are expected.
11.2 REPORTING DESCRIPTIONS
This section describes the purpose,
scope, scheduling, and review and clearance
procedures for various reports. These
descriptions are based on earlier EMAP plans
and on the summaries of the "Analysis and
Assessment" and the "Reporting" work groups
at the Joint EPAAJSDA-FS Meeting on Forest
Health Monitoring Coordination (January 23-24,
1990, Research Triangle Park, NC). The
characteristics and purposes of the reports
are summarized in Tables 11.1 and 11.2. A
logical sequence of the reports considered in
this chapter is shown in Figure 11.1.
11.2.1 Types of Reports
Reports delineated in this section
include plans, operations reports, data base
summaries, data quality reports, statistical
summaries, interpretive reports, and technical
proceedings. Plans describe the rationale and
intentions of the monitoring program and are
a focal point for peer and management
reviews. Operations reports summarize the
actual operations of the monitoring program.
Data base summaries and data quality reports
document the existence and quality of the data
that are collected. Statistical summaries
provide timely descriptions of regional status
and trends in forest condition in terms of a
few key indicators. Interpretive reports
address specific environmental issues, in-
corporate additional data and detail in the
analysis of status and trends, and con-
sider linkages between forests and other
11-1
-------
Table 11.1 Summary of reports to be produced by the interagency
forest monitoring program.
Report
Freqency/
Authors/
type
Scheduling
Publishers
Interagency
once/
multi-agency/
Monitoring Plan
spring '92
unknown
Interagency QAPP
once/
multi-agency/
spring *92
EPA
Research Plan
varies/
varies/
as needed
varies
Regional
annual/
multi-agency/
Work Plan
June 30*
FS regions
Regional QAPjP
annual/
multi-agency/
June 30*
EPA
Regional
annual/
multi-agency/
Operations Report
January 30b
FS regions
National
annual/
multi-agency/
Operations Report
March 30b
EPA
National
annual/
EPA/
Data Base Summary
June 3Qb
EPA
National Data
annual/
EPA/
Quality Report
June 30"
EPA
Regional
annual/
multi-agency/
Statistical Summary
September 3Qb
FS regions
National
annual/
multi-agency/
Statistical Summary
September 30"
EPA
Regional
3-5 yr./
multi-agency/
Interpretive Report
FS region
National
3 yr.
multi-agency/
Interpretive Report
EPA
(Continued)
11-2
-------
Table 11.1 (Continued)
Report Freqency/ Authors/
type Scheduling Publishers
MAP Integrated 3-yr. EPA/
Assessments EPA
Proceedings multi-agency/
FS national
'In the year prior to the year of data collection.
bIn the year following the year of data collection.
Table 11.2 Summary of purposes of reports to be produced by the
interagency forest monitoring program.
Report Type
Brief statement of purpose
Interagency
Monitoring Plan
Interagency QAPP
Research Plan
Regional
Work Plan
Regional QAPjP
Regional
Operations Report
National
Operations Report
Data base
Summary
Data Quality
Report
Basis of Interagency Agreement (IAG) between USDA-FS and EPA
for monitoring rationale, approach, and implementation of inter-
agency monitoring program
Basis for interagency Quality Assurance (QA) program
Initiate improvement of monitoring capabilities by exploring,
for example, alternate designs, indicators, and assessment models
Describe program management, logistics, QA, information
management, and reporting procedures for monitoring
Articulation of QA activities for each region
Summarize operational experiences and accomplishments and recommend
changes to implementation
Same as regional operations report
Signal agreement on, document contents of, and describe access to the
forest data base
Document the quality of data contained in the
forest data base
11-3
-------
(Continued)
Table 11.2 (Continued)
Report Type
Brief statement of purpose
Regional
Statistical
Summary
Provide timely summary of status and trends in forest condition; provide
regional authorship opportunity
National
Statistical
Summary
Provide timely national summary of status and trends in forest condition
Regional
Interpretive
Report
Investigate status arid trends in relation to regional environmental issues
and policies
National
Interpretive
Report
Investigate status and trends in relation to national environmental
issues and policies
Report
title
Brief statement of purpose
EMAP Integrated
Assessments
Investigate status and trends of forest condition in relation to
status and trends of other ecosystems
Proceedings
Provide publication opportunity and document monitoring procedures
YEAfl 1
YtAH2
YEARS
YEAR >3
IMRjeMEtfTATTOM
pum
I OATA
QUALITY
REPORT
STATISTICAL
SUMMARY
RESEARCH
PLANS
DATA
COLLECTED
VfWRED
DATABASE
OWHATIOHS !
DATABASE
suuuAflrr ! -
SUMMAKV
IHTWHt HVE
REPORT
iMTKMATCD
Figure 11.1 Sequence of reporting in EMAP-Forests.
11-4
-------
ecosystems. Technical proceedings provide
an opportunity for analysts to summarize and
publish findings and research and to
document monitoring techniques and
procedures.
11.2.2 Review and Clearance Procedures
In general, both the EPA and the FS
have similar review and clearance procedures.
They include informal peer review, formal peer
review, laboratory/station clearance, and
Washington Office clearance if needed.
Cooperative monitoring could require
review and clearance of multi-agency reports
by several agencies. EMAP-Forests
recommends the establishment of an
interagency (including but not necessarily
limited to EPA-EMAP and FS-FHM) monitoring
review committee to facilitate the review and
clearance of interagency reports.
If a document has single agency
authorship, publication and distribution is
handled by that agency. It must also pass
through the interagency review committee prior
to agency clearance by the authoring agency.
If a document has interagency authorship, a
pre-designated agency handles publication
and distribution. For clearance it must pass
through the interagency review committee and
clearance procedures for both agencies. Pre-
designated authorship of various reports
should be agreed upon early in the program.
11.3 FOREST MONITORING PLANS
The rationale for the interagency
monitoring program is described in regional
and national research plans. These plans
include monitoring design, data analysis,
indicator development, and assessments. The
implementation intentions of the multi-agency
program are described in regional work plans
that include program management, field and
laboratory logistics, QA, information
management, and reporting. Research and
work plans consider essential linkages and
coordination with other groups. For example,
research plans consider intra- and inter-
coordination among EMAP and FHM
monitoring tiers, and the work plans consider
coordination with other agencies for data
collection and reporting.
Research plans are difficult to
anticipate; flexibility is essential. The
participating agencies will prepare separate or
multi-agency, national or regional, research
plans with frequencies and contents that
depend on their separate and mutual needs.
A separate agreement for cooperative
research will usually accompany each inter-
agency research plan.
Annual regional work plans will have
multi-agency authorship and will be published
and distributed by the FS. The EPA will be
primarily responsible for QA, information
management, air quality and deposition,
meteorology, and laboratory sample collection
and analysis sections. The FS will be primarily
responsible for field data collection sections.
Program management and reporting sections
will be co-authored.
The target publication date for regional
work plans is June 30. The publication of
annual work plans should precede data
collection by at least one year. For example,
work plans produced during FY95 would
specify data collection for FY96. This
scheduling allows FY95 planning to make use
of Agency budget projections from FY94, and
the lead time allows full organization of data
collection. The multi-agency program is
currently operating on a compressed planning
schedule. One way to maintain current
activities while moving to the desired schedule
is to prepare a single two-year work plan in
FY92 (for FY 92 and FY 93 field work) to be
followed by the annual plans starting in FY93
(for FY 94 field work).
11-5
-------
The QAPP and QAPjP are described in
Section 8.
In contrast to this strategy for plans,
the first national plan (scheduled for 1992) will
be a multi-agency research and work plan that
will guide initial implementation and set the
stage for additional research. Publication and
distribution of this plan have not been
determined.
11.4 OPERATIONS REPORTS
An annual operations report describes
activities completed during the preceding data
collection cycle, evaluates the performance of
the monitoring program, and includes
recommendations for future work plans.
These reports have multi-agency authorship
and are published and distributed by the EPA
(national) and by the FS (regional). The target
publication dates are Jan. 30 (regional) and
March 30 (national) of the year following data
collection.
11.5 DATA BASE SUMMARIES
Data base summaries are needed for
three reasons. First, their publication signals
the existence of, and describes access to, a
validated data base that is available for data
analyses. A second purpose for the data base
summary is subtle but important. In an
interagency program, the report is a vehicle for
participants to agree what the data are as
opposed to what the data mean. The
distinction will enable data analyses to
proceed from an agreed-upon, common data
base, even if different conclusions are reached
by different analysts. This should eliminate
the question of whether different conclusions
result from different interpretations or different
data bases. Finally, the data base summary
provides an unambiguous reference for the
specific versions of the data base that are
used in any analyses.
The data base summary describes all
verified data (field and laboratory) collected by
the FHM program, as well as summaries (only)
of off-frame data that may be prepared
specifically for EMAP-Forests assessment
purposes. Data collected or summarized by
other EMAP Task Groups, for example EMAP-
Air and Deposition or EMAP-Landscape
Characterization, will be treated by separate
summaries prepared by those groups.
The data description should include a
data dictionary, formats, logical relations
among data base tables, and instructions for
access to data and QA information. It should
also describe the amount of data contained in
various data base tables for different years
and regions of the country.
The data base summary is produced
annually in addition to other information
management reports (see Section 10).
Because laboratory analyses, data verification,
and QA analyses require time beyond the "field
season" for completion, the data base
summary is scheduled for a June 30
publication date.
A single, national, data base summary
will be prepared by FHM and will be published
and distributed by the Forest Information
Center. The format will be comparable to
other EMAP resource group data base
summaries, thus providing all forest analysts
easier access to all ecosystem data that may
be managed by the overall EMAP Information
Center (EIC).
11.6 DATA QUALITY REPORT
EMAP-Forests adheres to EPA
requirements for QA and quality control (QC).
These requirements include certain types of
planning and evaluation reports (see Section
8). The data quality report, like the data base
summary, is designed specifically to assist
11-6
-------
analysts who use the FHM data base. The
purpose is to provide analysts with a concise
summary of the quality of the verified data
that are contained in the data base. Data
quality statistics and other descriptors are
essential for analysts so that the uncertainty
of assessments can be quantified.
All field and laboratory data collected
by the FHM program will be included, but data
derived from auxiliary sources will be excluded
from data quatity reports. The quality of
auxiliary data will be reported separately by
the collecting agencies.
Data quality descriptors may include
detectability, precision, accuracy, compara-
bility, and completeness. Measures of
uncertainty associated with system error may
be included.
The data quality report should be
prepared in a format that is very similar to the
data base summary. Like the data base
summary, this report is scheduled for an
annual June 30 release in the year following
data collection. A single, national data quality
report will be prepared by EMAP-Forests and
will be published and distributed by the EPA.
11.7 STATISTICAL SUMMARY
The statistical summary is designed to
provide timely descriptions of regional status
and trends of forest indicators monitored by
EMAP (see Section 3). The statistical
summary will rely mainly on pre-planned
("canned") analyses to produce compatible
reports among regions and over time. Such
analyses produce timely reports, but do so at
the expense of novel analyses and interpreta-
tions. Therefore, these reports are not
designed to diagnose specific causes of
specific changes in forest condition, to track
the recovery of particular forests in response
to particular control and mitigation programs,
or to report data from non-EMAP sources.
Pre-planned summaries may include EMAP
data collected off the forest sampling frame
(e.g., air quality, weather, and landscape
characterization data).
The statistical summary will highlight
the regional status and trends of forest condi-
tion and environmental stresses and exposure
in terms of a few key indicators because it is
not feasible to report all the measurements of
forest structure, function, and composition, as
well as measurements of environmental
stresses and exposure, in a concise format.
The statistical summary may also
show associations among indicators if the
associations are either pre-planned or known
to be non-controversial. This may be hard to
define. It is appropriate to use known
associations to identify potential "false
positives" that might otherwise be cause for
undue concern. For example, an association
between abnormal growth rates and abnormal
rainfall would be appropriate because there is
no real controversy about an underlying causal
relationship. The discussion of these
associations must emphasize that correlation
implies neither causality nor degree of effect.
One criterion of whether or not to include a
particular association will be interagency
clearance.
When a change in a given indicator is
small in relation to the uncertainty about
components of the indicator, conclusions
might be erroneous. Thus, the appendices of
the statistical summary will provide estimates
or other indications of measurement and
population uncertainties for each of the
reported indicators.
Prototypes of the statistical summary
(albeit with different indicators) include the
"core tables" common in FS FIA state reports.
These prototypes illustrate that raw data can
be summarized for routine and timely reporting
in a fashion that is consistent among regions
11-7
-------
and over time. Early reports from the FS-FHM
program utilize the "core table" concept. Table
11.3 suggests an outline for a typical
statistical summary.
Regional and national versions of the
statistical summary will be prepared annually.
The regional version is scheduled for a
September 30 release in the year following
data collection. The national version of this
report Is also scheduled for a September 30
release.
Statistical summaries will have
interagency authorship. The FS will assume
responsibility for publication and distribution
of regional reports, and the EPA will oversee
publication and distribution of national reports.
EMAP-Forests recommends that data
are released to the general public after
publication of the regional statistical
summaries (i.e., within one year after data are
collected). Prior to this release, verified data
may be accessed only by analysts authorized
by the interagency management structure.
11.8 INTERPRETIVE REPORTS
The interpretive reports are designed to
provide deeper analysis and assessment of
regional status and trends in forest condition.
In general, these reports are not designed to
diagnose specific causes of specific changes
in forest condition. An exception Is to track
the recovery of particular forests in response
to particular control and mitigation programs,
where supported by particular measurements
and procedures that augment the basic
monitoring (Tiers 1 and 2) design. The deeper
analyses can, however, suggest plausible
causes of changes and explore alternatives in
much more detail than can be accomplished in
the statistical summaries.
These reports may utilize data from
other EMAP resource groups and from non-
EMAP sources. They may consider the several
tiers of monitoring, explore the sensitivity of
apparent associations to varying assumptions,
and offer multiple interpretations of status and
trends.
The scope and content of these
reports is being discussed by EMAP and by
FS-FHM in the context of each total
monitoring-research program. At a minimum,
It is recommended that regional and national
interagency interpretive reports of forest
condition be prepared every three to five years.
These reports will be aimed at interpreting the
status and trends previously reported in the
statistical summaries. The reports will have
multi-agency authorship. The regional reports
will be published by the FS and the national
reports by the EPA Publication and distri-
bution dates of these multi-agency efforts has
not been determined.
In addition, EMAP-Forests will
participate every three years in EMAP-wide
integrated assessments that are coordinated
by the overall EMAP Integrated Assessment
Strategy. These reports will draw on forest
and other ecosystem data bases, will be
authored by the EPA and other agencies, and
will be published and distributed by the EPA
Among other uses, these reports are a link to
the EPA risk assessment reports.
11.9 TECHNICAL PROCEEDINGS
It is recommended that the
interagency program sponsor technical
conferences where cooperators would have
the opportunity to present their most recent
findings or research. It is further recom-
mended that the program publish a series of
monitoring research and applications reports
11-8
-------
Table 11.3 Typical outline of a statistical summary.
I. Executive summary
A. Statement of forest extent: What is the current extent of forest ecological resources, and
how are they distributed geographically?
B. Statement of forest condition: What proportions of the forest resources are currently in
good or acceptable condition?
C. Statement of regulatory concern: What proportions are degrading or improving, in what
regions, and at what rate?
D. Recommendations for monitoring and research.
II. Methods
A. Forest data collection.
B. Auxiliary data access.
C. Bio-geo-statistical methods
- Sampling, classification, and stratification
- Estimation of indicators and indices
- Statistical estimation of status and trends
- Special displays of information
D. Interpretive methods and caveats.
III. Results
A. Status and extent of apparently "healthy" forests.
B. Status and extent of apparently "unhealthy" forests.
C. Regional patterns and trends of forest health.
D. Regional patterns and trends of weather and pollution.
E. Atlases and spatial/temporal correlations.
IV. Discussion
A. Status and trends of forest condition in relation to trends of environmental stresses.
B. Possible/plausible associations (caveated).
V. Appendices
A. Data uncertainty estimation.
B. Data tables and documentation.
11-9
-------
as part of this conference. The purpose of
these publications is to document analysis
procedures and findings that would otherwise
not be appropriate for peer-reviewed journals
and would be lost with employee turnover. A
quarterly proceedings would be edited by
members of the interagency review committee
and/or advisory boards and published by the
FS.
11.10 ORGANIZATION OF REPORTING
EFFORT
The purpose of this section is to
assign, within EMAP-Forests, responsibilities
for contributions to the reports mentioned in
the previous sections.
It is premature to suggest specific
staffing requirements. However, the EMAP-
Forests staffing strategy should emphasize
extramural participation and cooperation to
avoid duplicating efforts and accumulating a
large in-house staff. A minimum in-house
staff will be responsible for technical areas in
which the EPA makes a unique contribution
such as air quality, weather, and deposition.
This staff will also coordinate EMAP-Forests
assessments and reporting and provide
analytical services needed by many
cooperators. Justification for "analytical
services" comes from familiarity with the
EMAP-wide data bases. In most cases,
EMAP-Forests reporting personnel will have
additional regional and national imple-
mentation or research responsibilities.
A reporting organization for EMAP-
Forests is depicted in Figure 11.2. A national
reporting unit and four regional units that are
coordinated by the national unit are required.
Regional units will be responsible for
cooperative reporting within their assigned
region (currently defined as FS mega-regions
[Figure 11.3]: North-Northeast, South-
Southeast, Rocky Mountain-Intermountain,
and Pacific Northwest-Pacific Southwest).
The national unit will be responsible for
cooperative national reporting and for
coordinating the reporting by regional units.
Reports will be written by both
"implementation" and "assessment" personnel.
Implementation personnel will oversee
contributions to work plans, operations
reports, data base summaries, and data
quality reports. Assessment personnel will
assume responsibility for contributions to
statistical summaries, interpretive reports, and
integrated assessments. Both groups
contribute to the Interagency Monitoring Plan,
research plans, and the proceedings series.
Table 11.4 summarizes the EPA
laboratory and national reporting assignments.
These assignments maintain regional
implementation and assessment reporting
capabilities and assign national leadership
roles that are consistent with EMAP's current
national assignments of laboratory roles.
Assignments of individual EPA Laboratories to
mega-regions are anticipated in FY91 or FV92.
The one exception to the regional-
national alignment within EPA is that both
national and regional units are assigned to
national interpretive reports. EMAP-Forests
scientists are organized according to
assessment topics with a national
perspective; therefore, participation by all
regional units in all regional and national
interpretive reports is possible. In some
cases, individual regional units may be
assigned, based on laboratory expertise, to
complete particular interpretive reports. The
national unit is identified specifically to
recognize a closer association with the overall
EMAP Integration and Assessment Project,
focusing on EMAP-wide integrated
assessments, and to coordinate regional and
national interpretive reports.
11-10
-------
FOREST
SERVICE
OTHER EPA
NATIONAL
UNIT
OTHER
FOREST
SERVICE
S/SE
UNIT
PNW/PSW
UNIT
NATIONAL
UNIT
RM/INT
UNIT
PNW/PSW
UNIT
S/SE
UNIT
NC/NE
UNIT
RM/INT
UNIT
NC/NE
UNIT
OTHER EMAP
TASK GROUPS
OTHER
AGENCIES
EMAP-WIDE
INTEGRATION
ft ASSESSMENT
Figure 11.2 EMAP-Forests reporting organization and liasons.
-------
NC/NE
PNW/PSW
Figure 11.3 Forest service mega-regions.
-------
Table 11.4 Laboratory and national reporting unit assignments to produce
documents described in this chapter for EMAP-Forests*
Staff type and
report type
National
unit
Regional
units
Implementation staff:
Regional Implementation
Plan and Operations
Report
National Implementation
Plan and Operations
Report
EMSL-LV
EMSL-LV
ERL-C
AHEAL-RTP
Data base Summary
Data Quality
Report EMSL-LV
EMSL-LV
Assessment staff:
Regional Statistical
Summary
EMSL-LV
ERL-C
AREAL-RTP
National Statistical
Summary
Regional Interpretive
Report
AREAL-RTP
EMSL-LV
ERL-C
AREAL-RTP
National Interpretive
Report and EMAP
Integrated Assessments
AREAL-RTP
EMSL-LV
ERL-C
AREAL-RTP
"Excludes Research Plans and Proceedings.
11-13
-------
11.11 ACTION PLAN
The plan for the evolution of reporting
capabilities assumes a phased, regional
implementation of monitoring and coincident
need for reporting. There is a need within
EMAP-Forests for regional capabilities in the
North-Northeast and South-Southeast regions.
In FY92, or FY93 at the latest, the other two
mega-regions will require reporting
capabilities. There is also a need to organize
the national reporting capability within EMAP-
Forests, especially with regard to EMAP-wide
assessments.
With respect to the interagency
reporting strategy, the following specific
recommendations are made;
• Utilize this strategy document as a basis
for reaching a formal interagency
agreement (IAG) in FY91 on:
1) The regional-national, implementation-
assessment framework.
2) The types of reports and the roles of each
agency.
3) The review, clearance, and scheduling of
reports.
• Allow for extramural (i.e., non-FS and non-
EPA) participation in reporting as part of
the IAG and circulate a joint strategy later
in FY91 to invite such participation.
• Identify an interagency review committee in
FY91.
• Include a multi-agency plan for reporting in
the Interagency Plan scheduled for FY92.
• EMAP-Forests should identify regional
reporting units for the two eastern mega-
regions in FY91 and for the two western
mega-regions no later than FY92.
• EMAP-Forests should identify the specific
roles and responsibilities for the national
reporting units in FY91.
• The FS should identify national reporting
units in FY91.
• The EPA and the FS should achieve the 3-
year planning, implementation, and
reporting schedule by FY93 at the latest.
This mainly affects the length of the
planning horizon in FY92.
• Include regional and national reports as
deliverables in the EMAP-Forests budgeting
process starting in FY92.
With respect to strictly EPA reporting
requirements, the following specific
recommendations are made:
• The national reporting unit should be
available to participate in the overall EMAP-
Integration and Assessment Project
(AREAL-RTP) strategies and plans in FY91-
FY92.
• As appropriate, the overall EMAP-
Integration and Assessment Project should
include EMAP-Forests personnel and
EMAP-Forests data bases in "example
integrated assessments" that may be
planned for FY91-92.
11-14
-------
12 REFERENCES
Albari, D.H. 1985. Seasonal changes in nutrient
concentration and content of aspen suckers in
Minnesota. For. Sci. 31:785-794.
Alexander, S.A. and J.A. Carlson. 1989. Visual
Damage Survey - Project Manual. Forest
Pathology Laboratory, Department of Plant
Pathology, Physiology and Weed Science.
Virginia Polytechnic Institute and State
University, Blacksburg, VA 24060-0330. 53 pp.
Anderson, R.L and R.P. Belanger. 1986. A
crown rating method for assessing tree vigor
of loblolly and shortleaf pines. Proceedings of
the Fourth Biennial Silvicultural Research
Conference. Atlanta, GA. pp. 538-542.
Anderson, R.L., C.M. Huber, R.P. Belanger, J.
Knighten, T. McMartney, and B. Brook. 1989.
Recommended survey procedures for
assessing ozone injury on bioindicator plants
in Region 8 Class I Wilderness Areas. USDA
Forest Service, Forest Pest Management,
Asheville Field Office Report No. 89-1-36 6pp.
Anonymous. 1987. Forest damage and air
pollution: Report of the 1986 forest damage
survey in Europe. Convention on Long-range
Transboundary Air Pollution. International Co-
operative Programme on Assessment and
Monitoring of Air Pollution Effects on Forests.
Prepared by the Programme Co-ordinating
Centres, with the assistance of the United
Nations Environment Programme (UNEP) and
the Secretariat of the United Nations Economic
Commission for Europe (ECE). 47 pp.
Anonymous. 1988. Forest ecosystem research
initiative. Draft. U.S. Environmental Protection
Agency Environmental Research Laboratory.
Corvallis, Oregon.
Anonymous. 1990. Report of the 1990 EMAP
tier 3/4 workshop. September 10-12, 1990.
Snowbird, Utah.
Atkeson, T.D. and A.S. Johnson 1979.
Succession of small mammals on pine
plantations in the Georgia Piedmont. Amer.
Midi. Nat. 101: 385-392.
August, P.V. 1983. TTie role of habitat
complexity and heterogeneity in structuring
tropical mammal communities. Ecology 64:
1495-1507.
Beaufils, E.R. 1973. Diagnosis and
Recommendation Integrated System (DRIS).
Soil Sci. Bull. No. 1. University of Natal,
Pietermaritzburg, South Africa. 132 pp.
Bickelhaupt, D.H., R. Lea, D.D. Tarbet, and A.L.
Leaf. 1979. Seasonal weather regimes
influence interpretation of pinus-resinosa foliar
analysis. Soil Sci. So. 43:417-420.
Bouma, J. 1989. Using soil survey data for
quantitative land evaluation. Advances in Soil
Science 9:177-213.
Bruce, R.R., A.W. White, Jr., A.W. Thomas,
W.M. Snyder, G.W. Langdale, and H.F. Perkins.
1988. Characterization of soil-crop yield
relations over a range of erosion on a
landscape. Geoderma 43:99-116.
Byers, G.E., R.D. Van Remortel, M.J. Miah, J.E.
Teberg, M.L. Papp. B.A. Schumacher, B.L.
Conkling, D.L. Cassell, and P.W. Shaffer. 1990.
Direct/Delayed Response Project: Quality
Assurance Report for Physical and Chemical
Analyses of Soils from the Mid-Appalachian
Region of the United States. EPA/600/4-90/001.
U.S. Environmental Protection Agency, Las
Vegas, Nevada.
12-1
-------
Childers, E.L., T.L. Sharik, and C.S. Adkissori.
1986. Effects of loblolly pine plantations on
songbird dynamics in the Virginia Piedmont. J.
Wildl. Manage. 50: 406-413.
Church, M.R., K.W. Thorton, P.W. Shaffer, D.L
Steven, B.P. Rochelle, G.R. Holdren, M.G.
Johnson, J.J. Lee, R.S. Turner, D.L Cassell,
D.A. Lammers, W.G. Campbell, C.I. Liff, C.C.
Brandt, L.H. Liegel, G.O. Bishop, D.C.
Mortenson, S.M. Pierson, and D.D. Schmoyer.
1989. Future effects of long-term sulfur
deposition on surface water chemistry in the
Northeast and Southern Blue Ridge Province.
EPA-600/3-89/061. Washington, DC: U.S.
Environmental Protection Agency.
Cochran, W.G. 1977. Sampling techniques.
Third Edition. John Wiley and Sons, New York,
NY.
Cochran, W.G., and G.M. Cox. 1957.
Experimental designs. J. Wiley & Sons, New
York. Pg. 256-263.
Commoner, B. 1971. The closing circle.
Knopf, New York, NY.
Cook, E.R. 1987. The decomposition of tree-
ring series for environmental studies. Tree-
Ring Bulletin 47:37-59.
Dueser, R.D. and H.H. Shugart. 1978.
Microhabitats in a forest-floor small mammal
fauna. Ecology 59: 89-98.
DuMouchel, W.J., and G.J. Duncan. 1983.
Using sample survey weights in multiple
regression analyses of stratified samples.
Journal of the American Statistical Association
volume 78, number 383, pp. 535-543.
Dwire, K., B. Huntley, and M. Miller-Weeks.
1990. Forest Service - Forest Health
Monitoring: Environmental Protection Agency -
Environmental Monitoring and Assessment
Program; Methods manual for field
measurements and sample collection, (in
preparation).
Environmental Protection Agency. 1987.
Guidelines and specifications for preparing
quality assurance program plans and quality
assurance annual report and workplans for
EPA national program offices and the office of
research and development. U.S. Environmental
Protection Agency, Washington, DC.
Felix, A.C. III, T.L. Sharik and B.S McGinnes.
1986. Effects of pine conversion on food
plants of northern bobwhite quail, eastern wild
turkey, and white-tailed deer in the Virginia
Piedmont. South. J. Appl. For. 10: 47-52.
Flelss, J.L. 1973. Statistical methods for rates
and proportions. John Wiley and Sons, New
York, NY.
Friedman, J. 1987. Exploratory projection
pursuit. J. Am. Stat. Assoc. 82:249-262.
Forman, R.T.T., and M. Godron. 1986.
Landscape ecology. John Wiley & Sons, New
York.
Graves, R.L 1990. Environmental monitoring
and assessment program quality assurance
program plan. Environmental Monitoring
Systems Laboratory, Office of Research and
Development, U.S. Environmental Protection
Agency, Cincinnati, Ohio.
Graybill, D.A. 1982. Chronology development
and analysis. In Hughes, M.K. Kelly, P.M.,
Pilcher, J.R., and LaMarche, V.C Jr., Editors.
Climate from Tree Rings, Cambridge University
Press, Cambridge UK. pp. 339-403.
Harper, J.L 1977. Population biology of plants.
London. Academic Press. 892p.
12-2
-------
Hazard, J.W., and B.E. Law. 1989. Forest
survey methods used in the USDA Forest
Service. EPA600/3-89/065. U.S. Environmental
Protection Agency, Office of Research and
Development, Corvallis, OR.
Hedges, L. and I. Olkin. 1985. Statistical
methods lor meta-analysis. Academic Press.
New York.
Horvitz, D.G. and D.J. Thompson. 1952. A
generalization of sampling without re-
placement from a finite universe. J. American
Statistical Association.
Huddle ston, J.H. 1984. Development and use
of soil productivity ratings in the United States.
Geoderma 32:297-317.
Hunsaker, C.T. and D.E. Carpenter, eds. 1990.
Ecological indicators for the Environmental
Monitoring and Assessment Program.
EPA/600/3-90-060. U.S. Environmental
Protection Agency, Office of Research and
Development, Research Triangle Park, NC.
Johnson, A.R. 1988. Diagnostic variables as
predictors of ecological risk. Environmental
Management 12:515-523.
Karr, J.R. 1981. Assessment of biotic integrity
using fish communities. Fisheries 6:21-27.
Karr, J.R., D.D. Faush, P.L Angermeier, P.R.
Yant, and I.J. Schlosser. 1986. Assessing
biological integrity in running waters: A
method and its rationale. Spec. Pub. No. 5.
Illinois Nat. Hist. Surv. Champaign, IL 28p.
Katz, G.M. 1984. Policy and Program
Requirements to Implement the Mandatory
Quality Assurance Program. Office of
Research and Development, Environmental
Protection Division. Washington, DC.
Kincaid, W.B., and Nash III, T.H. 1988.
Detection of a sulfur dioxide signal in a tree-
ring record: A case study from Trail, British
Columbia, Canada. GeoJournal 17:189-192.
Knapp, C.M., D.R. Marmorek, J.P. Baker, K.W.
Thornton, J.M. Klopatek, D.P. Charles. 1990.
The indicator development strategy for the
Environmental Monitoring and Assessment
Program. DRAFT DOCUMENT, U.S. EPA,
Environmental Research Laboratory- Corvallis.
Knight, F.B. and H.J. Heikkenen. 1980.
Principles of forest entomology. 5th Ed.
McGraw-Hill Book Co., New York. 461p.
Leaf, A.L., J.V. Berglund and R.E. Leonard.
1970. Annual variation in foliage of fertilized
and/or irrigated red pine plantations. Soil Sci.
So. 34:677.
Leopold, A. 1949. The land ethic. In A Sand
County Almanac and Sketches Here and There.
Oxford University Press, New York, NY. pp.
201-226.
Levins, R. 1966. The strategy of model
building in population biology. American
Scientist 54:421-431.
Li, K.C. 1989. Data visualization with SIR: a
transformation based projection pursuit
method. UCLA Statistical Series #24.
Linthurst, R.A. 1990. Keynote Address,
International Symposium on Ecological
Indicators, October 1990.
Linthurst, R.A., D.J. Landers, J.M. Eilers, D.F.
Brakke, W.S. Overton, E.P. Meier, and R.E.
Crowe. 1986. Characteristics of lakes in the
eastern United States. Volume I: Population
descriptions and physicochemical
relationships. EPA/600/300-89/037. U.S.
Environmental Protection Agency.
12-3
-------
Loftis, J.C., R.C. Ward, R.D. Phillips, and C.H.
Taylor. 1989. An valuation of trend detection
techniques for use in water quality monitoring
programs. EPA/600/3-89/037. U.S. EPA
Environmental Research Laboratory, Corvallis,
Oregon.
MacArthur, R.H. and J.W. MacArthur 1961. On
bird species diversity. Ecology 42: 594-598.
MacCracken, M.C., and Moses, H. 1982. The
first detection of C02 effects: Workshop
summary. Bulletin of the American
Meteorological Society 63:1164-1178.
Madow, W.G. 1949. On the theory of
systematic sampling II. Ann. of Matb. Stat.
20:333-354.
Magasi, LP. 1988. Acid rain national early
warning system. Manual on plot establishment
and monitoring. Information Report DPC-X-25.
Canadian Forestry Service, Government of
Canada, Ottawa, pp. 3/1-4/8.
Messer, J.J. 1990. EMAP Indicator Concepts.
In Hunsaker, C.T. and Carpenter, D.E., editors,
Environmental Monitoring and Assessment
Program Ecological Indicators. EPA 600/3-
90/060. U.S. Environmental Protection Agency,
Office of Research and Development, Research
Triangle Park, NC. pp. 2-1 - 2-26.
Miller, D.L, P.M. Leonard R.M. Hughes, J.R.
Karr, P.B. Moyle, LH. Schrader, B.A.
Thompson, R.A. Daniels, K.D. Fausch, G.A.
Fitzhugh, J.R. Gammon, D.B. Halliwell, P.L
Angermeier, and D.J. Orth. 1988. Regional
applications of an index of biotic integrity for
use in water resource management. Fisheries
13:12-20.
Miller-Weeks, M., and Gagnon, D.I.
(Compilers). 1990. Work plan for New Eng-
land forest health monitoring. USDA Forest
Service, Durham, NH.
Millers, I. and D. Lachance. 1989. North
American sugar maple decline project. Cooper-
ative field manual. U.S. NAPAP Terrestrial
Effects Task Group, Forest Response Program
- Eastern Hardwoods Research Cooperative;
Government of Canada, Forestry Canada and
USDA Forest Service. 16 pp. and appendices.
Mosteller, F., and Tukey, J.W. 1977. Data
analysis and regression: A second course in
statistics. Addison-Wesley, Reading, MA.
NRC (National Research Council). 1987.
Biological markers in environmental health
research. (Committee on Biological Markers cf
the National Research Council). Environ. Health
Perspect. 74:3-9.
NRC (National Research Council). 1989.
Biologic markers of air-pollution stress and
damage in forests. Committee on Biologic
Markers of Air-Pollution Damage in Trees,
Board on Environmental Studies and
Toxicology, Commission on Life Sciences,
National Research Council. National Academy
Press, Washington, DC. 363 pp.
NRC (National Research Council). 1990a.
Managing troubled waters: The role of marine
environmental monitoring. Committee on a
Systems Assessment of Marine Environmental
Monitoring, Marine Board, Commission on
Engineering and Technical Systems, National
Research Council. National Academy Press,
Washington, DC. 125 pp.
NRC (National Research Council). 1990b.
Forestry research: A mandate for change.
Committee on Forestry Research. Board on
Biology, Commission on Life Sciences, and
Board on Agriculture, National Research
Council. National Academy Press, Washing-
ton, DC. 84 pp.
12-4
-------
Ohmann, L.F., and Grigal, D.F. 1990. Spatial
and temporal patterns of sulfur and nitrogen in
wood of trees across the north central United
States. Canadian Journal of Forest Research
20:508-513.
Omernik, J.M. 1987. Ecoregions of the
coterminous United States. Ann. Assoc. Am.
Geog. 77:118-125.
O'Neill, R.V. 1988. Hierarchy theory and global
change. In Rosswall, T., Woodmansee, R.G.,
and Risser, P.G. (Editors), Scales and Global
Change: Spatial and Temporal Variability in
Biospheric and Geospheric Processes. John
Wiley and Sons, New York, NY. pp 29-45.
O'Neill, R.V., DeAngelis, D.L, Waide, J.B., and
Allen, T.F.H. 1986. A hierarchical concept of
ecosystems. Princeton University Press,
Princeton, NJ.
OTA. 1987. Technologies to maintain
biological diversity. OTA-F-331. Office of
Technology Assessment, Washington, DC. 47
PP-
Ott, W.R. 1978. Environmental indices: theory
and practice. Ann Arbor Science Publications.,
Ann Arbor, MI.
Overton, W.S. 1985. Working draft, analysis
plan for the Eastern Lake Survey. March 1985.
Technical Report 113, Dept. of Statistics,
Oregon State University.
Overton, W.S. 1987. Phase II analysis plan,
National Lake Survey - working draft. April
1987.
Overton, W.S. 1990. A strategy for use of
found samples in a rigorous monitoring
design. Technical Report 139, Department of
Statistics, Oregon State University, Corvallis,
OR.
Overton, W.S., and S.V. Stehman. 1987. An
empirical investigation of sampling and other
errors in the National Stream Survey: Analysis
of a replicated sample of streams. October
1987. Technical Report 199, Dept. of Statistics,
Oregon State University.
Overton, W.S., D. White, and D.L. Steven.
1990. Design report for EMAP, Environmental
Monitoring and Assessment Program, Part I.
Draft.
Palmer, C., J. Barnard, R. Brooks, and N. Cost.
1990. Forest health monitoring plot design and
logistics: A joint USFS/EPA study plan.
DRAFT DOCUMENT. U.S. EPA. Environmental
Research; Laboratory, Corvallis, OR.
Papp, M.L., R.D. Van Remortel, C.J. Palmer,
G.E. Byers, B.A. Schumacher, R.L Slagle, J.E.
Teberg, and M.J. Miah. 1989. Direct/Delayed
Response Project: Quality assurance plan for
preparation and analysis of soils from the Mid-
Appalachian region of the United States.
EPA/600/4-89/031. U.S. Environmental
Protection Agency, Las Vegas, Nevada.
Penrose, R. 1989. The emperor's new mind:
concerning computers, minds, and the laws of
physics. Oxford University Press, New York,
NY.
Plafkin, J.L., M.T. Barbour, K.D. Porter, S.K.
Gross, and R.M. Hughes. 1989. Rapid
bioassessment protocols for use in streams
and rivers: Benthic macroinvertebrates and
fish. EPA/444/4-89/001. U.S. Environmental
Protection Agency, Office of Water Regulations
and Standards, Washington, D.C.
Potter, V.R. 1988. Global bioethics: building
on the Leopold legacy. Michigan State
University Press, East Lansing, MI.
12-5
-------
Prigogine, I., and Stengers, I. 1984. Order out
of chaos: Man's new dialogue with nature.
Bantam Books, New York, NY.
Radford, P.J., and West, J. 1986. Models to
minimize monitoring. Water Res. 20:1059-1066.
Rapport, D.J. 1989. What constitutes
ecosystem health? Persp. Biol. Med. 33:120-
132.
Repenning, R.W. and R.F. Labisky. 1985.
Effects of even-age timber management on
bird communities of the iongleaf pine forest in
northern Florida. J. Wildl. Manage. 48: 895-
911.
Riitters, K.H., Barnard, J.E., Wester, A.R., Ford,
E.D., Saint, C.G., and VanDeusen, P.C. 1988.
Research to design long-term monitoring and
a prototype design of forest health monitoring.
Internal Forest Response Program report to
the Federal Management Group, Synthesis and
Integration Project and the National Vegetation
Survey, US Forest Service, Research Triangle
Park, NC.
Riitters, K.H., Law, B., Kucera, R., Gallant, A.,
DeVelice, R., and Palmer, C. 1990a. Indicator
strategy for forests. In Hunsaker, C.T. and
Carpenter, D.E., editors, Environmental
Monitoring and Assessment Program
Ecological Indicators. EPA/600/3-90/060. U.S.
Environmental Protection Agency, Office of
Research and Development, Research Triangle
Park, NC. pp. 6-1 - 6-13.
Riitters, K.H., K. Hermann, and R. VanRemortel.
1990b. Example statistical summary for
EMAP-Forests. Internal Report, U.S.
Environmental Protection Agency, Office of
Research and Development, Atmospheric
Research and Exposure Assessment
Laboratory, Research Triangle Park, NC.
Ripley, B.D. 1981. Spatial statistics. John Wiley
and Sons. New York.
Schmidt, R.A. 1978. Diseases in forest
ecosystems: the importance of functional
diversity. Pages 287-315 IN J.G. Horsfall and
E.B. Cowling (eds.). Plant diseases: an
advanced treatise; Vol. 2. How disease de-
velops in populations. Academic Press, New
York. 436p.
Schreuder, H.T., and J.P. McClure. 1991.
Modifying forest survey procedures to
establish cause-effect: Should it be done?
Manuscript submitted. US Forest Service,
Rocky Mountain Forest and Range Experiment
Station, Ft. Collins, CO.
Simmleit, N., and H.R. Schulten. 1989. Pattern
recognition of spruce trees: An integrated,
analytical approach to forest damage.
Environmental Science and Technology
23:1000-1006.
Smith, W.H. 1981. Air pollution and forests.
Springer-Verlag, New York, NY.
Smith, W.H. 1984. Ecosystem pathology: A
new perspective for phytopathology. Forest
Ecology and Management 9:193-219.
Stanley, T.W. and S.S. Verner. 1983. The U.S.
Environmental Protection Agency's Quality
Assurance Program. IN: Quality Assurance for
Environmental Measurements. ASTM STP 867.
pp. 12-19. American Society for Testing and
Materials, Philadelphia, Pennsylvania.
Stehman, S.V., and W.S. Overton. 1987a.
Estimating the variance of the Horvitz-
Thompson estimator in variable probability,
systematic samples. Proceedings of the
Section on Survey Research Methods of the
American Statistical Association.
Stehman, S.V., and W.S. Overton, 1987b. An
empirical investigation of the variance
estimation methodology prescribed for the
national Stream Survey: Simulated sampling
from stream data sets. October 1987.
12-6
-------
Technical report 118, Dept. of Statistics,
Oregon State University.
Strayer, D„ J.S. Glitzenstein, C.G. Jones, J.
Kolasa, G.E. Likens, M.J. McDonnel, G.G.
Parker, and S.T.A. Pickett. 1986. Long-term
ecological studies: An illustrated account of
their design, operation, and importance to
ecology. Occasional Publication 2, Institute of
Ecosystem Studies, New York Botanical
Garden, Mary Flagler Cary Arboretum,
Millbrook, NY.
Suter, G.W., II. 1990. Endpoints for regional
ecological risk assessments. Environ.
Manage. 14(1):9-23.
Treshow, M. 1984. Diagnosis of air pollution
effects and mimicking symptoms. In Treshow,
M., Editor, Air Pollution and Plant Life, John
Wiley and Sons, New York, NY. pp. 97-112.
USDA Forest Service. 1989. Interim resource
inventory glossary. U.S. Government Printing
Office. 96 pp.
US EPA. 1987. Unfinished business: a
comparative assessment of environmental
problems. Office of Policy Analysis and Office
of Policy, Planning and Evaluation.
Washington, D.C. 100 p.
US EPA. 1990. Environmental monitoring and
assessment program integrated assessment
strategy (DRAFT). U.S. Environmental
Protection Agency Office of Research and
Development, Washington, D.C.
US Forest Service. 1989. Plan for forest health
monitoring in the Northeast. DRAFT
DOCUMENT.
Wahba, G. 1990. Spline models for
observational data. Soc. Ind. Appi. Math,
volume 55 in the CBMS-NSF Regional
Conference Series in Applied Mathematics.
Wallace, H.R. 1978. The diagnosis of plant
diseases of complex etiology. Annual Review
of Phytopathology 16:379-402.
Walters, C.J., and Holling, C.S. 1990. Large-
scale management experiments and learning
by doing. Ecology 71:2060-2068.
Waring, R.H. 1990. Ecosystem stress and
disturbance. In Comparative Analysis of
Ecosytems: Patterns, Mechanisms, and
Theories. Springer-Verlag, New York, NY.
Waring, R.H., and Schlesinger, W.H. 1985.
Forest Ecosystems: Concepts and Manage-
ment. Academic Press, Orlando, FL
Willson, M.F. 1974. Avian community
organization and habitat structure. Ecology
55: 1017-1029.
Woodwell, G.M. 1974. Variation in nutrient
content of leaves of quercus-alba, quercus-
coccenea, and pinus-rigida in Brookhaven-
Forest from bud-break to abscission. Am. J.
Botany 61:749.
World Commission on Environment and
Development. 1987. Our common future.
Oxford University Press, New York, NY.
Zahner, R., Saucier, J.R., and Myers, R.K. 1989.
Tree-ring model interprets growth decline in
the southeastern United States. Canadian
Journal of Forest Research 19:612-621.
12-7
-------
13 GLOSSARY (after Hunsaker and Carpenter 1990)
Area frame - A sampling frame obtained by dividing a region into well-defined, identifiable
subregions that in aggregate comprise the total area of the region of interest. The subregions
are sampling units defined on maps or other cartographic materials.
Assessment endpolnt • A quantitative or quantifiable expression of the environmental value being
considered in the environmental analysis; examples include a 25% reduction in gamefish biomass
or local extinction of an avian species (Suter 1990).
Association rule - A rule that unambiguously links a single resource sampling unit with a grid
point if there are any resource units in the 40-hex centered at that grid point. Several such rules
have been identified in selecting a Tier 2 sample via the EMAP grid.
Augmented sample - A grid-based sample whose size has been increased by using a denser
grid.
Best management practices - Management practices targeted at minimizing specific watershed
disturbances, such as soil erosion, pollutant transport, stormwater runoff, or similar land-use-
related disturbances.
Bias - In a sampling context, the difference between the conceptual weighted average value of
an estimator over all possible samples and the true value of the quantity being estimated. An
estimator is said to be unbiased if that difference is zero.
Biodiversity - A conceptual term referring to the variety and variability among living organisms
and the ecological complexes in which they occur; diversity can be defined as the number of
different Items and their relative frequencies. For biological diversity, these items are organized at
many levels, ranging from complete ecosystems to the chemical structures that are the molecular
basis of heredity. Thus, the term encompasses different ecosystems, species, genes, and their
relative abundance (OTA 1987).
Blomarker • An indicator of cellular or physiological processes that signal events in biological
systems or samples. A biological marker of effect may be an indicator of an endogenous
component of the biological system, a measure of the functional capacity of the system, or an
altered state of the system that is recognized as impairment or disease. A biological marker of
exposure may be the identification of an exogenous substance within the system, the interactive
product between a xenobiotic compound and endogenous components, or other event in the
biological system related to the exposure (NRC 1987).
Bottom-up approach - Assessing ecological condition based on first principles, i.e., pollutant
effects are related causally to pollutant sources by transport and fate models.
13-1
-------
Candidate indicator - Indicator identified for each resource category by using a combination of
literature review, expert workshops, and interviews with scientists and environmental managers,
which was then judged against specific EMAP criteria to determine its feasibility as a research
indicator.
Characterization - Determination of the attributes of resource units, populations, or sample units.
A prominent use in EMAP is characterization of 40-hexes.
Classification - The process of assigning a resource unit to one of a set of classes defined by
values of specified attributes. Example: forest sites will be classified into the designated forest
types, depending on the species composition of the forest.
Core Indicator - EMAP indicator that is selected for long-term, routine monitoring based on its
performance as demonstrated in a regional demonstration project.
Cumulative frequency distribution - A distribution generated by a function (F(x)) such that at
any value for the variable x, F(x) represents the proportion of the resource sampling units in the
target population having a value for the variable that is less than or equal to x In EMAP, x is
usually a measurement of physical extent or an indicator measurement.
Deconvolutlon - Extraneous variation such as random errors in measurement has the effect of
inflating observed variation relative to true population variation. The cumulative distribution
function (cdf) that will be estimated when extraneous variation is present is the convolution of the
population (which is the cdf of interest) and the distribution of the extraneous variable. The
convolution cdf will be flatter (have longer tails) than the population cdf. Deconvolution is the
process of removing the influence of extraneous variation from an apparent cdf.
Developmental Indicator - An EMAP indicator that has passed evaluation for expected
performance (existing data analyses, simulation, and small-scale field tests) and, with the
concurrence of scientific peer reviewers, is deemed suitable for actual performance testing in a
regional demonstration project.
Diagnostic Indicator - Characteristics of the environment measured for the purpose of correlative
analysis to determine plausible explanations for subnominal conditions; a collective term for
EMAP exposure, habitat, and stressor indicators.
Digital line graph (DLG) • A standard U. S. Geological Survey computer format for representing
linear features of the earth, such as streams and roads, as they appear on maps.
Ecological Indicator - Response indicator.
Ecological resource category (resource category) - The aggregations of ecological resource
classes that are conveniently dealt with by ecologists with specific disciplinary expertise; six
categories currently are identified: near-coastal waters, surface waters, wetlands, forests, arid
lands, and agroecosystems. ecosystems.
13-2
-------
Ecological resource class (resource class) - A subdivision of an ecological resource category;
examples include small lakes, oak-hickory forests, emergent estuarine wetlands, field cropland,
mesohaline estuaries, and sagebrush dominated desert scrub.
Ecological risk assessment • The application of a formal framework to estimate the effects of
human action on a natural resource and to interpret the significance of those effects in light of
the uncertainties identified in each component of the assessment process. Steps in the
framework include initial hazard identification, exposure assessment, dose-response assessment,
and risk characterization.
Ecosystem - A local complex of interacting plants, animals, and their physical surroundings which
is generally isolated from adjacent systems by some boundary, across which energy and matter
move; examples include a watershed, an ecoregion, or a biome.
Ecosystem function - Attributes of the rate of change of structural components of an
ecosystem; examples include primary productivity, denitrification rates, and species fecundity
rates.
Ecosystem structure - Attributes of the instantaneous state of an ecosystem; examples include
species population density, species richness or evenness, and standing crop biomass.
Environmental indicators - A collective term for response, exposure and habitat, and stressor
indicators.
Explicit sampling frame - The representation of a target population (resource category, class, or
subclass), each unit of which has a unique identification code, used to implement a sampling
strategy; an example includes a list of all lakes greater than 4 ha in the Northeast.
Exposure Indicator - A characteristic of the environment measured to provide evidence of the
occurrence or magnitude of a response indicator's contact with a physical, chemical, or biological
stress.
Grid enhancement - Increasing the grid density; method for augmenting the sample. When the
Her 1 sample size is too small, as will occur for rare resources, the grid density may be increased
in order to obtain a sample size adequate for population description.
Grid, triangular (EMAP) - A lattice of points in exact equilateral triangular structure. The EMAP
grid points are 27.1 km apart.
Grid, baseline (EMAP) - The fixed position of the EMAP grid as established by the position of
the global hexagon covering the United States. This is distinguished from the random position of
the grid as used for sampling.
Grid randomization - The process of randomly positioning the grid so that each (compact and
small) unit of area of fixed size is equally likely to contain a grid point. This is the basis for the
probability-sample designation of the EMAP sample.
13-3
-------
Habitat Indicator - A physical attribute measured to characterize conditions necessary to support
an organism, population, or community in the absence of pollutants.
Hazard - A state that may result in an undesired event; the cause of risk. In EMAP, any human-
related event or activity that unintentionally or inadvertently can affect ecological condition;
examples are acidic deposition that may decrease the acid-neutralizing capacity of surface water,
or application of fertilizer to a forested watershed that may increase nutrient levels in adjacent
streams.
Hazard Indicator - Measures that reflect human activities that unintentionally affect ecological
resources (e.g., measures of pollutant release, number of permits issued for construction activity,
and rates of application of fertilizers to forests and crops that influence nutrient concentrations in
adjacent streams).
Hexagon - A regular six-sided polygon. A tessellation of hexagons is the dual of a triangular
point grid; each point in the grid is the center of a hexagon, and the hexagons tile the surface.
These hexagons on the EMAP grid have size of 634.5 km.
40-hex - The landscape description hexagon that is established on each of the grid points in the
EMAP grid. Actual size of these hexagons is 634.5 116 = 39.7 km2.
Hierarchical (grid) - Having nested levels and structure. Density of the EMAP grid is readily
increased or reduced in a regular manner into hierarchical levels of density. Adjacent levels may
differ in density by a variety of factors: 3, 4, 7 or many combinations of these base factors.
Typically, the grid of points at one level will be contained in the grid at a higher density.
Implicit sampling frame - A set of rules or criteria used to select resource sampling units that
cannot be listed a priority a unique identification code (upon which indicators will be measured);
the rules are developed as part of the landscape characterization activities performed on the
landscape sampling units.
Inclusion probability - The probability of including a specific sampling unit in the sample.
Index (Indices) - Mathematical aggregation(s) of indicators or metrics; one example is the Index
of Biotic Integrity (IBI), which combines several metrics describing fish community structure,
incidence of pathology, population sizes, and other characteristics.
Index period - Sampling period that yields the maximum amount of information during the year,
which may vary from one indicator or resource class to another.
Indicator - A characteristic of the environment that, when measured, quantifies the magnitude of
stress, habitat characteristics, degree of exposure to the stressor, or degree of ecological
response to the exposure.
13-4
-------
Interpenetrating design - The monitoring survey design used in EMAP, in which a new set of
resource sampling units (RSUs) is selected each year during four successive years. "The four-year
cycle is repeated by using the same set of RSUs as in the first cycle; therefore, the same set of
RSUs sampled in year 1 would be resampled in year 5.
Kriglng - A weighted, moving-average estimation technique based on geostatistics that uses the
spatial correlation of point measurements to estimate values at adjacent, unmeasured points.
Landscape - The fundamental traits of a specific geographic area, including its biological
composition, physical environment, and anthropogenic or social patterns.
Landscape characterization - The documentation of principal components and patterns of
landscape structure, including attributes of the physical environment, biological composition, and
cultural patterns. In EMAP, a term referring to the process of describing land use or land cover
within the landscape sampling units.
Landscape ecology - The study of the distribution patterns of communities and ecosystems, the
ecological processes that affect those patterns, and changes in pattern and process over time
(Forman and Godron 1986).
Landscape Indicator - A characteristic of the environment, calculated from remotely sensed
data, used to describe spatial distribution of physical, biological, and cultural features across a
geographic area.
Landscape sampling unit - The selected units (e.g., 40-km2 hexagons) upon which landscape
characterization will be performed.
Management Indicator - Measures that reflect human activities that intentionally alter an
ecological resource to meet some management objective; for example, the dredging or filling of a
wetland for the purpose of housing development.
Maximum/minimum operators approach - A mathematical aggregation scheme used to produce
an ecological condition index based on several response indicator values; the index assumes the
value of the most subnominal indicator.
Natural process Indicator - Measures that reflect cyclic or acyclic phenomena that affect
ecological condition, regardless of the presence of management actions or environmental
hazards; examples include natural climatic fluctuations, predator-prey cycles, and insect and
disease epidemics.
Nominal - The state of having desirable or acceptable ecological condition.
Population estimate - A statistical estimate of some characteristic (or distribution of
characteristics) that applies to an explicitly defined target population (category, class, or
subclass), e.g., the median acid-neutralizing capacity (or the cumulative frequency distribution of
acid-neutralizing capacity) for all small lakes in the Northeast.
13-5
-------
Probability sample/sampling - A sample chosen in such a manner that the probability of each
selected unit is known; for EMAP, each resource sampling unit (e.g., a lake, a forest, an estuary)
upon which indicator measurements are to be made will have a known probability of being
selected.
Randomization - The process of imposing an element of chance on the selection of a sample.
This may take many forms; this step of the design protocol is the basis for determining "design-
based" properties.
Region - Any extensive geographic area that generally corresponds in size to EPA administrative
Regions III through X (e.g., physiographic regions, ecoregions, major river basins).
Regional ecological resource class (regional resource class) • An ecological resource class
that is distributed over some natural spatial range, e.g., southeastern oak-hickory forests or small
lakes in the Northeast.
Regional reference site - One of a population of benchmark or control sites that, taken
collectively, represent an ecoregion or other broad biogeographic area; the sites, as a whole,
represent the best ecological conditions that can be reasonably attained, given the prevailing
topography, soil, geology, potential vegetation, and general land use of the region.
Research Indicator - A candidate indicator identified for an EMAP resource category which has
been prioritized on the basis of several criteria (e.g., regionally applicable, integrates effects,
monotonic, conducive to synoptic monitoring) and, following peer review, has been selected for
further evaluation for use in EMAP, as possible developmental indicators; evaluation of expected
performance includes analyzing existing data, performing simulation studies with realistic
scenarios and expected spatial and temporal variability, and conducting limited field tests.
Resource sampling unit - A particular ecological resource (e.g., a stream segment, a forest
stand, a wetland, an estuary) upon which indicator measurements will be made; more than one
resource sampling unit can occur in a landscape sampling unit.
Response Indicator - A characteristic of the environment measured to provide evidence of the
biological condition of a resource at the organism, population, community, or ecosystem process
level of organization.
Sample - A subset of the units from a frame. A sample may also be a subset of resource units
from a population or a set of sampling sites.
Sampling design - A sample consists of a set of sampling units or sites that will be
characterized. Sampling units are defined by the frame; they may correspond to resource units,
or they may be artificial units constructed for the sole purpose of the sampling design.
Stratum/strata - A stratum is a sampling structure that restricts sample randomization/selection
to a subset of the frame. Inclusion probabilities may or may not differ among strata.
13-6
-------
Stressor Indicator - A characteristic measured to quantify a natural process, an environmental
hazard, or a management action that effects changes in exposure and habitat.
Stressor - Measurements used to provide information on human activities or externalities that
can cause stress in ecological entities; three types of stressor indicators are considered in EMAP;
hazard indicators, management indicators, and natural process indicators. Examples are the
incidence of fertilizer application, which can increase nutrient concentrations in lakes; incidence of
dredging/filling, which can diminish availability of wetland habitat; and climatic fluctuations, which
can promote damage by pathogens.
Subnomlna) - The state of having undesirable or unacceptable ecological condition.
Systematic sample - A sampling design that utilizes regular spacing between the sample points,
in one sense or another. The EMAP design selects samples via the triangular grid; spatial
arrangement of the selected resource units is not always strictly systematic, but the systematic
grid is an important aspect of the design.
Target population - The set of ecological resources from which a sample is drawn.
Threshold - The value for a particular response indicator used to distinguish nominal from
subnominal ecological condition.
Tier 1 resource sample - Ail resource sampling units of each resource class within all landscape
sampling units.
Tier 2 resource sample - A subsample of the Tier 1 resource sample used for field sampling of
indicators.
Top-down approach - Assessing ecological condition based on correlative analyses; i.e.,
pollutant effects are associated temporally or spatially with pollutant sources by statistically
correlational analysis.
Validation - The process of determining the legitimacy of data, involving internal consistency
checks for outlier removal and definition of levels of confidence.
Verification The process of confirming the integrity of data, involving discrepancy, precision,
and accuracy checks.
Weights - In a probability sample, the sample weights are inverses of the inclusion probabilities;
these are always known for a probability sample. In other contexts, statistical weights are
indicated for other reasons.
13-7
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before camp' '
1 REPORT NO. 2.
EPA/600/4-91/012
PB92-146208
4. TITLE ANO SUBTITLE
MONITORING AND RESEARCH STRATEGY FOR FORESTS-
ENVIRONMENTAL MONITORING AND ASSESSMENT PROGRAM
(EMAP)
5. REPORT DATE
March 1992
6. PERFORMING ORGANIZATION CODE
7. AUTHORtS)
C.J. Palmer, K.H. Rittters, T. Strickland, D.L. Cassell,
G.E. Byers, M.L. Papp, and C.I. Liff
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME ANO ADDRESS
University of Nevada-Las Vegas
Environmental Research Center
Las Vegas, NV
10. PROGRAM ELEMENT NO.
11. dONTRAtT/6RANt noUNLV-ERC
/rCR81470, LESC. 068-CO-OO49,
Lantech Corvallis J'68-CO-OOo,
^anter_h-RTP Dfi
12. SPONSORING AGENCY NAME AND AODRESS
Environmental Monitoring Systems Laboratory-LV, NV
Office of Research and Development
U.S. Environmental Protection Agency
Las Vegas, NV 89193-3478
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/07
15. SUPPLEMENTARY NOTES
16. ABSTRACT
To protect, manage, and use forest resources effectively,
the condition of these resources must be known. Concern about
documented and potential effects of air pollutants in combination
with other multiple, interacting stresses has been a major
impetus behind the development of monitoring programs in forests.
During the past two years, the forest component of the
Environmental Protection Agency's Environmental Monitoring and
Assessment Program (EMAP-Forests) has been working closely with
the Forest Service's Forest Health Monitoring (FS-FHM) program
and other government agencies to develop a multi-agency program
to monitor the condition of the nation's forested ecosystems.
The purpose of this document is to present a strategy that
can be used as a starting point by all government agencies
interested in participating in a nationwide FHM program.
Monitoring issues such as design, indicator selection, and
assessment are presented along with approaches to resolving these
i ssues.
17. KEY WOROS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b. IDENTIFIERS/OPEN ENDEDTERMS
c. COSati Fielil/Croup
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS 'This Report)
UNCLASSIFIED
21. NO. OP PAGES
191
20 SECURITY CLASS iTIuspagr,
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (R«v. 4-7?) PREVIOUS ECITiOn is obsolete
14-1
------- |