-^ -" .  - ••<•••• 
-------
FOREST TASK GROUP ANNUAL STATISTICAL SUMMARY:
               A HYPOTHETICAL EXAMPLE
                            Authors

                          Kurt Riitters
                          Karl Hermann
                 NSI Technology Services Corporation
                      Environmental Sciences
                 Research Triangle Park, North Carolina

                        Rick Van Remortel
               Lockheed Engineering and Science Company
                        Las Vegas, Nevada
                           June 1990
              OFFICE OF RESEARCH AND DEVELOPMENT
              U.S. ENVIRONMENTAL PROTECTION AGENCY
             RESEARCH TRIANGLE PARK, NORTH CAROLINA

-------
                                TABLE OF CONTENTS


FOREWORD                                                                           i

ACKNOWLEDGEMENTS                                                                 iii

1.0 Introduction                                                                        1
1.1 Background for Forest Monitoring                                                      1
1.2 EMAP Goal and Objectives                                                           2
1.3 Description of Annual  Statistical Summaries                                             2
1.4 Organization of This Report                                                           5

2.0 Example Annual Statistical Summary: 2000                                             4
2.1 Forest Extent and Distribution                                                         4
2.2 Status and Trends of Response Indicators                                               7
2.3 Exposure and Habitat Condition                                                      10

3.0 Conclusions                                                                       17

Appendix A. Response Indicators                                                        18
A. 1 Forest Extent and Distribution                                                        19
A.2 Visual Symptoms                                                                   21
A.3 Growth  Efficiency                                                                  26

Appendix B. Exposure/Habitat Indicators                                                 29
B.1 Foliar Nutrients and Contaminants                                                    30
B.2 Soil Productivity                                                                    32

Appendix C. Ancillary Data                                                             34
C.1 Biotic and Abiotic Stress Indices                                                     35
C.2 Injury to Pollution "Indicator" Plants                                                   37
C.3 Landscape Indices                                                                 38
C.4 Tree/Shrub/Herb Species Number and Variety                                          39
C.5 Other Ancillary Data                                                                40
Appendix D. Uncertainty Estimation                                                      41

Appendix E. Classification and Aggregation Schemes for Reporting                            42

Appendix F. Literature Cited                                                             47

-------
                                         NOTICE

This document is a contribution to the Environmental Monitoring and Assessment Program. The contents
of this document do not necessarily reflect the views and policies of the Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endorsement or recommendation for
use. Statistics contained in the following text and figures are hypothetical. This document is in-
tended to illustrate a possible format of an EMAP-Forests Annual Statistical Summary. The
hypothetical data and  statistics do not reflect the actual status or trends in forest conditions.

The research described in this document has been funded by the United States Environmental Protection
Agency through  Contract 68-02-4444 to NSI Environmental Sciences and Contract 68-03-3249 to
 Lockheed Engineering  and Science Company.

-------
                                       FOREWORD

This document is a product of the Environmental Monitoring and Assessment Program (EMAP). EMAP-
Forests is part of an interagency forest health monitoring program which is being designed to collect, inter-
pret, and report forest status and trends data. The purpose of this document is to illustrate the format and
content of one type of report, the Annual Statistical Summary. This report is an example which is based
on hypothetical data:  it is not intended to portray actual forest conditions.

A sample report based on hypothetical data can be useful to:

1.   Inform potential users what can be produced by a fully implemented forest monitoring program, by
    illustrating possible questions and reporting formats.

2.   Help to ensure that the ultimate design for monitoring will satisfy user needs, by identifying specific
    design objectives to be built into the monitoring program.

3.   Help to identify goals for research to improve forest monitoring, by summarizing what is currently
    possible to accomplish.

-------
                              ACKNOWLEDGEMENTS

Kent Thornton of FTN Associates provided project guidance and comments on several drafts of this
manuscript. Introductory information was obtained from the October 1989 draft of the EMAP Design
Report and the December 1989 draft of the EMAP-Forests Indicator Report. The map of United States
forest types was digitized by S.H. Azevedo, Forest Ozone Team, United States Environmental Protection
Agency, Corvallis.  Members of the EMAP-Forests Task Group, with the technical direction of Craig Palmer
and coordination of Beverly Law, provided information for the Appendices and Brenda Huntley provided
assistance with the final production of this report.

-------
                                     1.0 Introduction

The need to establish baseline ecological conditions against which future changes can be documented
with confidence grows more acute with the increasing complexity, scale, and social importance of environ-
mental issues such as global climate change, acidic deposition, and deforestation. Monitoring programs
must be in place to provide quantitative,  scientific assessments of the complex effects of pollutants on
ecosystems. The goal of EPA's Environmental Monitoring and Assessment Program  (EMAP) is to monitor
the condition of the nation's ecological resources, to evaluate the success of current policies and
programs, and to identify emerging problems before they become more widespread or irreversible.

To help achieve this goal, the EMAP-Forests Task Group participates in a forest monitoring program with
the U.S. Department of Agriculture (USDA) Forest Service and other Agencies.  Cooperators share respon-
sibilities to monitor the health of forested ecosystems, but with different perspectives because of different
missions.  A cooperative approach can address EPA's concern for environmental regulation and can also
address issues related to forest resource management.  The purpose of this report is to illustrate the
format and possible content of one type  of report that would address EPA's concerns.

1.1 Background for Forest Monitoring

Forest lands provide many amenities important to the social, economic, and cultural aspects of life. The
most recent national assessment of forests (USDA Forest Service 1982) gives some indication of the
extent and value of forests and why there is public and congressional support for improved monitoring
programs.

    •   Forests cover approximately 737 million acres, or about one third of the total land area of the
       United States.

    •   Two thirds of the nation's water supply comes from forested watersheds.

    •   Forests have great social and cultural values.

    •   Forest ecosystems are important habitat for fish, birds, and wildlife.

    •   Timber-based economic activities are valued at more than $48 billion (4.1 % of the gross national
       product).  Employment for more than 3 million workers originates from forests, and in many rural
       locations, the primary employment sector is forest-related.

Public and Congressional support for forest monitoring has a very long history.  Almost 100 years ago, the
Organic Act of 1891 established the national forests and included provisions for the inventory of these
lands.  Later, the Forestry Research (McSweeney-McNary)  Act of 1928 required a current and comprehen-
sive inventory and analysis of all of the nation's forest resources.  This early legislation focused on timber
inventory,  and it was not until 1974 and the passage of the Forest and Rangeland Renewable Resources
Planning Act (RPA) that much consideration was given to monitoring nontimber resources. Since passage
of the RPA, Congress has passed the National Forest Management Act (1976), the Federal Land  Policy
and Management Act (1976), the Soil and Water Conservation Act (1977), and the Forest Ecosystems and
Atmospheric Pollution Research Act (1988). These legislative acts all contain the following assignments
(Lund 1986).
       prepare and maintain continuous natural resource inventories
       coordinate and cooperate among resource agencies and organizations to avoid duplication of
       inventory and planning efforts
       determine current and potential changes in renewable natural resources
       determine resource interactions and management alternatives
       submit periodic assessment reports of the natural resources to the Nation

-------
Forest monitoring is also relevant to assessments made pursuant to other legislation, including the Federal
Insecticide, Fungicide and Rodenticide Act, the National Environmental Policy Act, the Resource
Conservation and Recovery Act, and the Endangered Species Act. This legislation has given increasing
emphasis to nontimber resources, for example wildlife,  rangeland, water, and recreation amenities, and
has required quantifiable environmental information for establishing environmental policy at the national
level.

Many factors affect the health or condition of forests and thus are of potential concern for environmental
regulation. These include global warming, atmospheric deposition, weather, soil erosion, insects,
diseases, and human activities. Current public concerns about the possible deterioration of forest
condition center on global climate change, pollution (particularly air pollution), and land use. Rapid
changes in global climate may occur as a result of increases in the concentrations of trace gases in the
atmosphere (NAS 1983), and this could have severe ecological consequences (Abrahamson 1989). There
is also concern about air pollution impacts on forests (Mclaughlin 1985; American Forestry Association
1987; NAPAP 1988).  Finally, land use zoning and resource management guidelines reflect concerns for
good land stewardship and regulation of activities such as urban and agricultural development, mineral
extraction, and silviculture. Forest monitoring provides baseline data to address these and other issues of
public concern.

1.2 EMAP Goal and  Objectives

The goal of EPA's Environmental Monitoring and Assessment Program (EMAP) is to monitor the condition
of the nation's ecological resources to evaluate the success of current policies and programs and to
identify emerging problems before they become widespread or irreversible. EMAP provides a strategic
approach to meet the growing need to identify and bound the extent, magnitude, and location of
degradation or improvement in environmental condition.  EMAP focuses specifically on regional-scale
conditions over periods ranging from years to decades. This focus, along with EMAP's integrated design,
distinguishes the program from many current environmental monitoring efforts. When fully implemented,
EMAP will answer the following critical questions:

    •  What is the current ecological status, extent, and geographic distribution of each ecosystem class
       of interest?
    •  What proportion of each ecosystem class is currently in acceptable condition?
    •  What proportions of each ecosystem class are degrading  or improving, in what regions, and at
       what rate?
    •  What are the likely causes of degradation or improvement of ecosystem conditions?

To  provide the information needed to answer these questions, EMAP has three major objectives:

1.  estimate with known confidence the current ecological status, extent, changes, and trends in indicators
    of the condition of the Nation's ecological resources on a regional basis;

2.  monitor indicators of pollutant exposure and habitat condition and seek associations between human-
    induced stresses and ecological condition that identify possible causes of adverse effects; and

3.  provide to the public periodic statistical summaries and interpretive reports on ecological condition.

1.3 Description of the Annual Statistical Summary

The Annual Statistical Summary is one type of report that can be produced from forest monitoring data.
Its purpose is to provide timely descriptions of regional  status and trends in forest condition by addressing
a subset of EMAP questions:
1.  What is the current regional ecological status, extent, and geographic distribution of forest resources?

2.  What proportions of forests are degrading or improving, where, and at what rate?

-------
In general, EMAP statistical summaries 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 mitiga-
tion programs, or to report data from non-EMAP sources.  These questions are topics for other types of
EMAP reports such as periodic interpretive reports and integrated assessments (EPA 1990). The process
of compiling and analyzing data for an annual statistical summary sometimes can suggest specific topics
for these other reports.

The annual statistical summary will highlight status and trends in terms of a few key indicators because it is
not feasible to report all the measurements of forest structure, function, and composition in a concise
 format (Hunsaker and Carpenter 1990). These key "response indicators" are measures or indices
(aggregations, compositions, or functions of measurements) which quantify forest condition and which are
the basis for classifying forest condition as relatively "good" or "poor". Three response indicators have
been identified by EMAP-Forests for annual reporting (Appendix A).

Relatively few components of forest condition can be quantified and classified by a regional forest
monitoring program at this time. The initial focus of the response indicators is on trees, but efforts are
underway to develop or refine techniques which will broaden the scope of response indicators to include
birds and other wildlife, landscape diversity, soils, other vegetation, and water quality.

Based on currently available  response indicators, the annual statistical summary can address these
general questions:

1.  What are the extent and trends of forest atea of different types in different regions?

2.  According to the status and trends of visual indicators of tree condition, what proportions of forests of
    different types are degrading, where, and at what rate  in different regions?

3.  According to the status and trends of indicators of forest growth, what proportions of forests of different
    types are degrading, where, and at what rate in different regions?

The annual statistical summary will also report information about "exposure/habitat indicators" which are
measures of environmental stresses (Appendix B) and will  document the existence of selected ancillary
data (Appendix C). This information will be useful for confirming or interpreting trends of forest condition
suggested by the response indicators. The annual reporting mechanism is a vehicle for compiling and
documenting the basic data that will be needed for interpretive reports.

When a change in a given  indicator is only small,  in relation to the uncertainty about  components of the
indicator, conclusions might  be erroneous. Thus, the Appendices of the annual statistical summary will
provide estimates or indications of measurement and population uncertainties for each of the variables
measured (Appendix D).

1.4 Organization of This Report

The organization of the remainder of this report generally follows the format of a hypothetical Annual
Statistical Summary for the year 2000. The Summary will contain a highly condensed, heavily illustrated
main body; the condensed summary is illustrated for one EMAP Reporting Region of the United States
(Section 2.0). The Summary will also contain appendices which provide more complete summaries of
more variables (Appendix A through Appendix D of this report). Appendix E describes the forest classifica-
tion scheme (major forest types) and data aggregation categories (EMAP Reporting Regions) which will be
used in the annual statistical summaries.

-------
                2.0 Example Annual Statistical  Summary:  2000

IMPORTANT NOTE
Statistics contained in the following text and figures are hypothetical. This section and the Appen-
dices are intended only to illustrate a possible format of an EMAP-Forests annual statistical sum-
mary.  The hypothetical data and statistics do not reflect the actual status or trends in forest
conditions.

The Environmental Monitoring and Assessment Program (EMAP) was initiated by the U.S. Environmental
Protection Agency (EPA) in 1990 in response to the need to monitor ecological status and trends in a
national-scale, integrated, pollutant exposure and ecological condition monitoring network.  EMAP is a
comprehensive, interagency, multidisciplinary program which documents and periodically assesses the
condition of the nation's ecological resources. This report on the nation's forest conditions is the tenth in a
series of annual statistical summaries designed to provide observations of conditions during the past year.

The goal of forest monitoring is to detect regional trends in forest condition which are of potential concern
to land and environmental managers, the public, and scientists. This information is used to help develop
national environmental and land management policies. The EPA, the USDA Forest Service, and State
Agencies have measured indicators of forest condition at more than 5,000 locations since EMAP was
initiated in 1990. They have also measured indicators  of environmental stresses to help relate forest
conditions to natural  phenomena and human activities. This report highlights some of the regional
indications of forest condition from the past measurement season (1999), and considers trends in forest
condition observed over the past decade.

Many forest types are distributed across approximately one third of the United States (Figure 2-1).  Some
regions may have very little forest cover, yet in other regions forest is the predominant land cover.  In
either case, forests vary widely over any single region. Heterogeneous natural patterns and processes
and a variety of spatially patterned natural and human-caused environmental stresses complicate the task
of forest monitoring.  Statistical and ecological principles are used to select a small sample of forest land
for measurement, with confidence that the sample truly represents the regional status and trends of forest
condition. Information collected from a variety of forest and site conditions has been combined into a  few
categories for this report.

2.1 Forest Extent and Distribution

Major changes in forest area  have resulted from land-clearing for urbanization and agriculture and of
reestablishment of forests on abandoned agricultural land. In the eastern United States, approximately 18
million hectares of agricultural land were abandoned and recolonized by forests between 1920 and 1960,
and there has been a net loss of forest area in some regions due to urban development since 1960 (USDA
Forest Service 1982). These  changes affect aesthetic and economic values of forests, as well as their
quality as habitat for birds and wildlife.

EMAP-Forests monitors the areal extent of 22 major types of forests by periodically classifying and
measuring a statistical sample of about six percent of the total land area in the United States. Changes in
the amount, proportions, and distributions of major forest types in different regions can be estimated with
confidence from these data.  These changes provide an overall picture of the regional trends in forest area
which might be of concern, but the statistical sample does not allow identification of specific locations of
all forests.

There is more to be known about forest distribution than trends in the areal extent of major forest types
which are reported here. For example, stresses (e.g., Dutch elm  disease, American elm; chestnut blight,
American chestnut; ozone damage, ponderosa pine) have resulted in species changes but not in
reductions of forested areas.  Additional analyses of the EMAP-Forests data bases can help answer some
of these more specific questions about forest extent and distribution.

-------
                                                                                                 Eastern  Forests
                                                              SCALE 1:20,000,000
                                                                                                       White - Red - Jack Pine
                                                                                                       Spruce - Fir
                                                                                                       Longleaf - Slash Pine
                                                                                                       Loblolly - Shortleaf Pine
                                                                                                       Oak - Pine
                                                                                                       Oak - Hickory
                                                                                                       Oak - Gum - Cypress
                                                                                                       Elm - Ash - Cottonwood
                                                                                                       Maple - Beech - Birch
                                                                                                       Aspen - Birch
                                                                                                 Western Forests,
                                                                                                 Alaska,  and  Hawaii
                                                                                                       Douglas - Fir
                                                                                                       Hemlock - Sitka Spruce
                                                                                                       Redwood
                                                                                                       Ponderosa Pine
                                                                                                       White Pine
                                                                                                       Lodgepole  Pine
                                                                                                       Larch
                                                                                                       Fir Spruce
                                                                                                       Western Hardwoods
                                                                                                       Ohia
                                                                                                       Pinyon - Juniper
                                                                                                       Spruce - Hardwoods
                     Alasko
Hawaii
Digitized version produced in 1989,
by S.H. Azevedo, Forest  Ozone Team,
USEPA Environmental Research Lab, Corvallis, Oregon
FIGURE  2-1.    REGIONAL  DISTRIBUTION   OF   MAJOR  FOREST  TYPES
                     IN   THE  UNITED  STATES       (EYRE   1980)

-------
    Major Forest Types
                                     PIE  AREA IS PROPORTIONAL TO
                                     TOTAL FOREST AREA IN  REGION.
        White - Red - Jack Pine
        Spruce - Fir
        Loblolly - Shortleaf Pine
        Oak - Pine
        Oak - Hickory
        Oak - Gum - Cypress
        Elm - Ash - Cottonwood
        Maple - Beech - Birch
        Aspen - Birch
SCALE 1:6,000.000
                                                             14.6%
                                                              6.3X
                                                          REGION
                                                      2.2  MIRION  ACRfcS
                                           MILLION ACRES
                                     REGION 3
                                 42.5 MILLION  ACRES
                                           11.3%
                                                                         10.6X
FIGURE 2-2.   HYPOTHETICAL  EXAMPLE  OF  FOREST AREA  STATISTICS:
AREA  BY  FOREST  TYPE  AND EPA  RECION  IN  THE  NORTHEASTERN  UNITED STATES

-------
Northeastern United States

The hypothetical current areal extent of forest land area by major forest type for three EPA regions in the
Northeastern EMAP Reporting Region is shown in Figure 2-2. The pie charts in this Figure illustrate
hypothetical differences in the total amount of forest land and in the relative abundance of different types
of forests. The inset  map (taken from Figure 2-1) illustrates how some of these differences are related to
the general geographic distribution patterns of forest types as determined by climate and physiography;
the inset does not portray the land use or forest type at any given  location.

Hypothetical changes occurring over the 10 years since the 1990  measurement of forest extent in 1990 are
illustrated in Figures 2-3 and 2-4. Bar charts in Figure 2-3 illustrate 10-year changes by forest type in each
EPA Region, in both absolute terms and percentages.  Figure 2-4  shows the locations where forest area
was measured and the percentage change in forest area at each location.  The bar charts and geographic
displays identify forest types and locations for which changes in forest area are most pronounced.

2.2 Status and Trends of Response  Indicators

Stresses on forests are ubiquitous yet variable, and population-level summaries of forest monitoring data
are expected to reflect a wide range of forest condition in any one region or forest type.  Even when
conditions are "normal" everywhere, a certain proportion of forests will be in poor condition because of
"normal" stresses such as weather, pathogens, and stand maturation. Yet while the condition of individual
forests is dynamic, the condition of a population of forests will be  consistent if the population is subject to
the same stresses over time. For this reason, a statistical summary does not focus on the apparent
condition of any particular tract of forest.

EMAP-Forests classifies forest condition and reports the regional status and trends of forest condition in
terms of visual symptoms scores and growth efficiency values. Together, these indicators describe forest
condition as it is related to productivity, sustainability, and aesthetic appeal. Regional summaries of these
response indicators are snapshots of apparent forest conditions at a given time.  Full interpretation of
these conditions requires baseline information about the normal or expected patterns of conditions. It
may sometimes be difficult to define a "normal" value for any indicator, or to ascribe significance to the fact
that a certain proportion of forests appears, on the basis of indicators, to be in "abnormal" condition.  But
in any case, trends which signal increasing proportions of the population shifting towards "abnormal"
conditions are of concern.

Forests are complex ecosystems which should be described by more than just two indicators of condition.
Visual symptoms scores and growth efficiency values summarize  only a small part of forest ecosystem
condition; it is possible that conditions could change in important  dimensions which are not detectable by
these two indicators. A better characterization of conditions would consider forest structure, composition,
and function in terms of trees, lesser vegetation, animals, and soils. It would also include both spatial and
temporal measures of pattern and change and would provide ways to combine information from different
types of indicators.  It is critically important to improve the nation's abilities to monitor and interpret
additional measures of forest condition. The three current indicators and existing technology for making
the necessary measurements are but a starting point; future EMAP-Forests annual statistical summaries
will be made more comprehensive as our monitoring capabilities are improved.

Northeastern United States

The current status of forests as measured by the visual symptoms indicator and by the growth efficiency
indicator is illustrated in Figures 2-5 through 2-8. The pie charts in Figure 2-5 show the percentages of
samples which indicate "good", "marginal", and "poor" conditions based on one interpretation of the two
response indicators for each EPA Region.  These pie charts provide the most general picture of regional
condition, but they may in some cases obscure differences among forest types.  Figure 2-6 brings out
these differences among forest types by combining the information from different EPA Regions.

-------
                                KEY
             RED-WHFTE-JACK PINE
             SPRUCE-FIR
             LOBLOLLY-SHORTLEAF PINE
             OAK-PINE
CD
OAK-HICKORY
ELM-ASH-COTTONWOOD
MAPLE-BEECH-BIRCH
ASPEN-BIRCH
       L.

1 2 •

0.8-
0.4-
o;
0.4-
0.8-
1.2-



* -
0,8-
0.4:
°]
-0.4^
0.8 ^
•
-1.2-

MILLION
ACRES
R-nwr5)^
•-—§•-"



MILLION
ACRES
n n^\/7r7^\/^
inivre
- W— TT-


LJ
REGION 1
40 - PERCENT

'TIjillM'li^^^L
.
-20-
-40-

REGION II
40 J PERCENT
J
j_^
Tln][MTfl(^/S\L
op >1B
-20-
"
-40 H

1


,.-

_ ,J





•^n*


.
                               REGION II!

                                    40 - PERCENT
FIGURE 2-3. HYPOTHETICAL EXAMPLE OF NET CHANGES IN FOREST EXTENT
BY EPA REGION (l.ll.lll) AND FOREST TYPE.  CHANGES FROM 1990 TO
2000 ARE SHOWN  IN ACRES (LEFT) AND PERCENT (RIGHT).

-------
         Ten-Year Change in Total Forest  Area
             o

             •
> 20X Loss
10 - 20Z Loss
< 10% Change
10 - 20X Gain
> 20* Gain
                                                                     Region 1 \
                                                                  -1.38 Million
                                                       Region 2
                                                    -1.82 Million Acres
                                                        Each symbol represents one EMAP primary sampling
                                                        unit ("hexagon") of approximately 40 square kilometers.
                                                 Region 3
                                               + 1.40 Million Acres
FIGURE\-4.   HYPOTHETICAL EXAMPLE  OF  FOREST AREA  STATISTICS:
GEOGRAPHIC  DISTRIBUTION OF  NET TEN-YEAR  PERCENTAGE  CHANGES IN
TOTAL FOREST  AREA  OBSERVED ON EMAP  PRIMARY SAMPLING UNITS.

-------
In the hypothetical scenario, the three EPA Regions have generally similar percentages of forests in dif-
ferent apparent condition, and the two response indicators tell about the same regional story (Figure 2-5).
But differences in apparent condition among forest types, and as suggested by different indicators,
become evident in the more detailed summary (Figure 2-6).

The geographic distributions of apparent forest conditions are shown in Figures 2-7 and 2-8. In these
Figures, the values obtained for each response indicator are mapped according to the sample locations.
Inspection of these figures can suggest geographic subregions where the extremes of forest condition are
concentrated. The EMAP sample design measures only one quarter of the sample locations in any one
year; the values mapped in the figures come from the most recent measurements available (1997-2000).
The specific forest types sampled at each of the illustrated locations are likely to differ from the types
suggested by the regional map in Figure 2-1 because the  regional map loses information when the general
forest-type boundaries are drawn.

In the hypothetical scenario, both response indicators seem to suggest a concentration of relatively poor
forest condition in two or three forest types near the predominantly urban areas. These types of patterns
would be topics for additional analyses in other EMAP reports.

Time trends of forest condition are illustrated in Figures 2-9 and 2-10. In  Figure 2-9, the percentages of
samples which have been classified as indicating "good", "marginal", and "poor" conditions are charted
over time for each response indicator in each EPA Region. As an overall regional summary, this figure is
analogous to the display of current conditions seen earlier in Figure 2-5.  The breakdown of trends by
forest type (analogous to Figure 2-6) are shown in Figure 2-10. These charts suggest whether regional
conditions are improving or degrading, at what rate, and whether trends  in condition differ among forest
types.

In the hypothetical scenario, the three EPA regions have different time trends in overall forest condition
(Figure 2-9), even though it was suggested earlier that they are currently  in about the same general
condition. The more detailed  description of trends by forest type suggests that the conditions of different
types of forests are improving or  degrading at different rates.

2.3  Exposure and Habitat Condition

Forests everywhere are  under continual and variable stress from a variety of natural and  human-related
sources. An important objective  of the annual statistical summary is to provide regional characterizations
of the stresses which are measured at the sampled locations.  Other types of EMAP-Forests reports
consider these and other stresses in developing hypotheses about the possible causes of changing forest
condition.

Regional summaries of the status and trends of indicators of exposure and habitat condition summarize
environmental stresses which can affect forest condition.  But careful interpretation of the relationships
between various types of indicators is needed before associations may be drawn; experience has shown
that more than one environmental stress can usually be associated with detrimental changes in condition,
and partial "explanations" of changes can lead to ineffective environmental regulations.

Ideally, stresses would be characterized by a number of exposure and habitat condition indicators. When
monitoring was initiated  in 1990, however, it was not possible to accept simple indicators of stresses which
were appropriate for regional summaries. Scientists were able to identify key measures of foliage and soil
chemistry, air pollution exposure, landscape pattern, and biotic and abiotic stresses which, it was believed,
would ultimately be expressed as indicators of environmental stresses (see Appendices B and C).

Identifying appropriate indicators is difficult; we still cannot use many simple summary statistics for
regional reports. Nevertheless, we must Identify additional indicators of ecological condition and
additional measures of exposure and habitat condition,  so that future EMAP-Forests statistical summaries
and interpretive reports will be useful for their designed purposes.
                                                10

-------
       KEY !
                  CONDITION
                  INDICATOR
GOOD


 > 75
MARGINAL   POOR
     VISUAL SYMPTOMS
        (% foliated)

     GROWTH EFFICIENCY   >100
        (g/m2/yr)
[_	_			__	

  VISUAL SYh

        5%
                                         40-75
                                         40-100
           < 40
           <40
 REGION
 REGION !!
 REGION !!
                                           'H EFFICIENCY
                                             6%
                                                      64%
                                                       64%
                                                       56%
FIGURE 2-5. HYPOTHETICAL EXAMPLE OF FOREST CONDITION STATISTICS:
PERCENTAGES OF ALL SAMPLED LOCATIONS BY REGION  IN DIFFERENT
CONDITION AS MEASURED BY VISUAL  SYMPTOMS  (LEFT)  AND BY
GROWTH EFFICIENCY (RIGHT).  SEE APPENDIX A FOR DESCRIPTIONS.

-------
   KEY       CONDmON
             INDICATOR
           VISUAL SYMPTOMS
              (% foliated)
                            GOOD   MARGINAL   POOR

                             > 75    40-75     < 40

                                               <40
           GROWTH EFFICIENCY   >100    40-100
              (g/m2/yr)
        i		_			,__„		,	j
        AREA OF PIE IS PROPORTIONAL TO EXTENT OF FOREST TYPE.
WHFTE-RED-JACK PINE

       6%
   29%
   43
           65%
           40%
  OAK-PINE    ASPEN-BIRCH

      ©
38%(^52^

                13%
"'(I64*   43*<%4*
                                                LOBLOLLY-SHORTLEAF
                                                   24?
                                                 16
                                                 29%
                                                     4%
                                                        60%
                                                        67%
    24%
      SPRUCE-FIR
         5%
            ,71%
                       MAPLE-BEECH-BIRCH
    36%
ELM--ASH-COTTONWOOD
      '


       10%
      ^^
                         OAK-HICKORY
                             2%
                                 90%
                     FIGURE 2-6.  HYPOTHETICAL EXAMPLE OF FOREST
                     CONDITION STATISTICS: PERCENTAGES OF
                     SAMPLED LOCATIONS BY FOREST TYPE IN
                     DIFFERENT CONDITION AS  MEASURED BY VISUAL
                     SYMPTOMS (TOP) AND  BY GROWTH EFFICIENCY
                     (BOTTOM) IN THE NORTHEASTERN UNITED STATES.
                     SEE APPENDIX A FOR DESCRIPTIONS.

-------
           Visual Symptoms
            (% foliated)


           Poor     ( < 40)
           Marginal  (40 - 75)
           Good     ( > 75)
                        ; ov :• iV^--_
                                                   Values are adjusted to the year 2000.
                                                   Each symbol represents one sampled forest location.
FIGURE>Z-7.   HYPOTHETICAL EXAMPLE  OF FOREST  CONDITION  STATISTICS:
GEOGRAPHIC DISTRIBUTION  OF  APPARENT FOREST CONDITION AS MEASURED BY
THE  VISUAL SYMPTOMS  INDICATOR  IN  THE  NORTHEASTERN  UNITED  STATES.

-------
           Growth Efficiency
             (g/m2/yr)


           Poor      ( < 40)
           Marginal  (40 - 100)
           Good      ( > 100)
                                                    Values are adjusted to the year 2000.
                                                    Each symbol represents one sampled forest location.
FIGURE>2-8.   HYPOTHETICAL  EXAMPLE OF  FOREST  CONDITION  STATISTICS:
GEOGRAPHIC DISTRIBUTION  OF APPARENT  FOREST  CONDITION  AS MEASURED BY
THE  GROWTH  EFFICIENCY INDICATOR IN THE NORTHEASTERN  UNITED  STATES.

-------
         KEY
PERCENT
100
 80
 60
 40
 20-
 0

100
 80
 60
 40
 20
 0

100
 80
 60
 40
 20
 0
                    CONDITION
                    INDICATOR
                                          CONDITION
                 VISUAL SYMPTOMS
                    (% foliated)
                 GROWTH EFFICIENCY
                    (g/m2/yr)
                            POOR    MARGINAL   GOOD
                            < 40     40-75      > 75
                     < 40     40-100
                                              >100
                   LINES SEPARATE PERCENTAGES BY CONDITION
                   (SEE I
                                REGION I
GOOD
MARGINAL
POOR
                                REGION
                                                               100
• 80
• 60
• 40
• 20
 0
                                                               100
                                                               80
                                                              I- 60
                                                               40
                                                               20
                                                               0
                                                               100
                                                                80
                                                                60
                                                               - 40
                                                               • 20
                                                                0
          1990
                       98  2000  1990  92
                                  94    96
                                   YEAR
                                                         98 2000
  FIGURE 2-9. HYPOTHETICAL EXAMPLE OF FOREST CONDITION STATISTICS:
  TIME TRENDS OF THE PROPORTIONS OF FORESTS IN DIFFERENT
  CONDITION AS MEASURED BY THE VISUAL SYMPTOMS (LEFT) AND
  GROWTH EFFICIENCY (RIGHT) INDICATORS IN THE NORTHEAST.
PERCENT

-------
LL1
O
cc
111
Q.
    100 •
     80-
     60-
     40-
     20-
     60.
     40.
     20.
     0
                   CONDITION
                   INDICATOR
                 VISUAL SYMPTOMS
                   (% foliated)

                 GROWTH EFFICIENCY
                   (g/m2/yr)
                                   POOR   MARGINAL   GOOD

                                   < 40    40-75      > 75
                           <40     40-100
                                                     > 100
                  LINES SEPARATE PERCENTAGES BY CONDITION
         WHITE-RED-
         JACK PINE
                         SPRUCE-FIR
         GOOD
         MARGINAL
         POOR
                                 LOBLOLLY-
                                 SHORTLEAF
  OAK-PINE
LU
O
CC
LU
OAK-HICKORY
                         ELM-ASH-
                         COTTONWOOD
                                         MAPLE-
                                         BEECH-BIRCH
ASPEN-BIRCH
1UU
80'
60"
40'
20"
0 •
80-
60-
40-
on .
C.\J
0 .





'^-~-JL^~-



•^ 	 	 	 ~^~
       1990   95  2000   1990
                              i
                             95
                          2000  1990   95  2000   1990   95  2000
                            YEAR
 FIGURE 2-10. HYPOTHETICAL EXAMPLE OF FOREST CONDITION STATISTICS:
 TIME TRENDS OF THE PROPORTIONS OF FORESTS IN DIFFERENT
 CONDITIONS AS MEASURED BY THE VISUAL SYMPTOMS (TOP) AND GROWTH
 EFFICIENCY (BOTTOM) INDICATORS, BY FOREST TYPE, IN THE NORTHEAST.

-------
                                     3.0 Conclusions

The "conclusions" section of the annual statistical summary will emphasize the most significant patterns of
indicators observed during the past year, and relate them to any important trends observed over the past
several years.  The patterns of primary interest are those related to long-term trends which occur over
large regional areas. Obvious short-term changes which have occurred over large regional areas of the
United States are also of interest.  The conclusions section will compare the results obtained for the
different EMAP reporting regions (only one such Region was illustrated in Section 2.0).

It will be difficult to conclude that the forests in a given region of the United States are in "better" or "worse"
condition in comparison with the immediate past year; most short-term changes are likely to be in
response to immediate environmental conditions such as weather that would vary from year to year.
Apparent short-term changes can also be due to sampling error. At most, the annual statistical summary
can refer to the continuation, or interruption, of long-term regional trends in forest conditions as evidence
that forests are improving or degrading. Similar precautions are in order when attempts are made to
compare results for different regions of the country.

The conclusion of the annual statistical summary should emphasize the limits of interpretability, as well as
the necessarily arbitrary definition of cut-off indicator values for "good", "marginal",  and "poor" conditions.
Interested readers should be invited to read the appendices, which contain enough information to relate
the sensitivity of the major conclusions to the choice of these cut-off values.  Scientists may request data
sets in order to pursue any particular hypothesis of interest to them.

The EMAP-Forests annual statistical summary is a regional summary of data collected from a national
monitoring network.  Regional status and apparent trends  of key indicators of forest condition and
associated indicators of environmental stresses can be used to identify specific stresses in specific forest
regions which must be examined in more detail in subsequent analyses.
              i
The need for timely collating and processing of monitoring data to produce an annual summary of forest
condition precludes detailed reporting of all of the measurements, or investigation of the causes of any
forest changes which might be indicated.  However, the emphasis on timely reporting ensures that the
necessary data will be available without delay, in a concise and documented format, for more detailed
follow-up analyses.
                                                17

-------
                         Appendix A. Response Indicators

This appendix describes statistics related to forest extent and to the "response" indicators proposed for the
annual statistical summary. As defined by Hunsaker and Carpenter (1990), response indicators are used
to describe and to classify the status and trends of forest condition. A brief rationale for each indicator is
followed by a general description of the methods of sampling and measurement.

This appendix will include two types of "core" (standard) tables which will be prepared for each EMAP
Reporting Region: (1) descriptive statistics of the measured variables, derived indicators, or summary
indices, and; (2) data quality statistics for the measured variables. In addition, cumulative distribution
functions will be prepared to display statistics about the derived indicators or summary indices for selected
populations of interest.

An important consideration for response indicators is the quantitative definition of "good", "marginal", and
"poor" categories. It is desirable to classify the apparent condition of each sampled site in terms of the
calculated values of one or more response indicators.  It is a reasonable expectation, given the objectives
of EMAP and the expense of collecting data, to be able to summarize what can be learned from the data in
just a few tables and charts that can be grasped easily. At the same time, it is important to state that  other
interpretations are possible and to provide statistics in the appendices so that alternate sets of assump-
tions can be explored.

The assumption has been that the expected values and normal variation of the different response
indicators are known sufficiently to enable rigorous definition of the different categories. This assumption
has proven tenuous for many response indicators which have been suggested; in fact, this is the primary
reason why there are so few of them reported in the annual statistical summary.

Although it is usually possible to ascribe value to observing a relatively higher or lower value of some
indicator, it is not easy to identify the specific cut-off values which define distinct categories of  condition.
One reason is that the functional relationships between measures of "goodness" and indicators are usually
smooth and continuous, and they usually do not exhibit obvious breakpoints  which would simplify bound-
ary definitions. Another reason is that definitions may differ for the variety of forest types and sites which
are sampled.  Finally, there is concern that univariate definitions are not  appropriate.

For current and new response indicators, it may be necessary to develop cut-off indicator values which are
not constants, or to develop multivariate interpretive methods. This would improve the confidence in the
simple summaries that were illustrated in Section 2.0, but it need  not alter the basic reporting format.
These improvements are logical refinements that need  to be made as a  result of research and  experience.
The preliminary definitions made in this Appendix are considered to be starting points for further review
and development.
                                                18

-------
                             A.1 Forest Extent and Distribution

Rationale

All forests are potentially at risk, so the total extent of forests sets the upper limit of the amount of land
resource which is of potential concern. Thus, the status and trends of forest extent are useful information.
Trends in forest extent are also of concern, for example, as indications of the trends of habitat or of the
amount of resource available for recreational uses.

The size, species composition, age, density, and diversity of individual  tracts of forest can vary markedly.
Because EMAP's regional perspective utilizes a relatively coarse sampling grid, it is necessary to think of
forests as comprising major classes of forest types existing across their natural ranges (see Appendix E).
Finer resolution of forest types is essential for other types of analyses conducted at local scales.

Existing Federal inventory systems provide estimates of forest extent according to the EMAP-Forests
classification scheme (e.g., USDA Forest Service 1982; see USDA Forest Service 1985 and Hazard and
Law 1989).  These estimates may in fact be  more precise or more accurate than EMAP estimates derived
from measurement of about six percent of the total United States. More precise information should not be
ignored, and experience will help to decide the best method for combining useful information from several
sources.

At the same time, EMAP requires statistics about the extent of forests which are consistent with the
estimated extent of other types of ecosystems such as agricultural lands, deserts, and grasslands.
Existing systems will not necessarily produce consistent estimates because of minor differences in the
definitions of forests, the sampling frequencies, and classification protocols.  The descriptions which follow
assume that only EMAP data are used to develop estimates of the extent of forests.

Definition and measurement

The definition, photointerpretation, and measurement of forest land area are described by Muchoney et al.
(1990).  The forest classification system utilized by EMAP recognizes 22 major forest types. At this time,
plans of the EMAP-Characterization Task Group (see EPA 1989a) rely mainly on delineation of forest cover
types on 1:40,000 leaf-off photography, supplemented where appropriate by other sources of information.
A minimum mapping unit of 1.0 hectare has been proposed.  Land use and cover type information
collected by the EMAP-Characterization Task Group will be available for analyses on separate "layers"
within the EMAP Geographic Information System (EPA 1989a).

Sample selection

A systematic sampling grid placed over the  United States identifies approximately 12,600 locations of
primary sample units (EPA 1989b). At each location, remotely sensed data will be used to classify and
delineate the land use and forest cover types within an area of approximately 40 km2 (within a "hexagon")
surrounding the sample point. In this fashion, about six percent of the total land area of the United States
will be characterized.  Remeasurements are currently planned at 10-year intervals.

Core tables

    pescriptive statistics
   The extent of forests will be referred to as "forest area". The following tables will be prepared on the
    basis of observations obtained per primary sample unit ("hexagon").

    1.   Minimum, maximum, median, quartiles, mean, and standard deviation of forest area, total and by forest
       type
   2.   Minimum, maximum, median, quartiles, mean, and standard deviation of number of forest types
                                               19

-------
3.   Minimum, maximum, median, quartiles, mean, and standard deviation of average delineated forest
    polygon size, total and by forest type
4.   Minimum, maximum, median, quartiles, mean, and standard deviation of annualized change in forest
    area since the previous measurement, total and by forest type

Data quality statistics
Data quality statistics will be specified by the EMAP-Characterization Task Group and are not available at
this time.
                                            20

-------
                                    A.2 Visual Symptoms

Rationale

Trees experience many stresses.  Very often a tree will lose foliage when combined stresses cause an
undesirable change in its function or condition. While foliage loss is not always a result of these stresses,
trees which have lost foliage have experienced a change in their status which is generally considered to be
detrimental to overall function or condition. Thus, foliage loss or defoliation is considered to be a useful
response indicator.

People care about foliage amount for many reasons. For example, there is the aesthetic quality of a walk
in the woods, which is thought to be dependent upon the appearance of the trees, especially the condition
and amount of foliage. In another view, tree foliage is the forest "factory1 upon which productivity
depends. Without leaves, trees do not produce wood or other amenities which have economic value. In a
third example, foliage is considered to be an important component of ecosystem processes, serving as
food for animals, insects, and microorganisms.

Definition and measurement

One way to measure the visual symptoms indicator is as the actual percentage of the "full" complement of
foliage for each sample tree. With this method, the foliation of each tree is observed from several perspec-
tives, and the indicator is estimated by comparing the amount of foliage to the amount expected for each
tree. The difference, expressed as a percentage of full or expected foliation, is the visual symptoms score
for a tree.

There are other ways to measure the amount of foliage.  A  common alternative is to measure foliage den-
sity, or equh/alently, the "transparency" of a tree crown. The advantages of this approach are that it does
not require knowledge of an expected amount for each tree, and transparency can be measured by a
slightly more objective technique. The disadvantage is that transparency is expected to vary more widely
under "normal" circumstances than defoliation. Crown transparency is being explored as an alternative
technique for measuring visual symptoms.

Sample selection

Visual symptoms are scored on several trees at each monitoring station (plot). At present, the specific
rules for selecting sample trees vary in different regions of the United States, because EMAP is utilizing the
plot designs which have been established differently in each region by the  USDA Forest Service. For
every design, however, trees are selected with known probability, and so comparable plot-level statistics
can be developed.

Visual symptoms score statistics will be reported by species and by forest type. Any species or forest type
must be observed at a minimum of 50  sample locations to  qualify for reporting in any given category.
Recognized forest types are listed in Appendix E. In the eastern United States, species recognized by the
USDA Forest Service, Forest Inventory and Analysis (Hansen et al. 1990), will be utilized by EMAP
Western species have not yet been decided.

Per-plot averages are to be reported. Per-plot species averages are estimated as the average score for all
sampled trees of a given species. Per-plot overall averages are estimated as the weighted (by estimated
number of trees per  unit area) average score for all sampled trees at a given location. The per-plot overall
averages are considered to represent the forest type at any given location.
                                               21

-------
The following cutoff values will be used to classify condition.

    Condition      Score
                   (% foliated)
    Good          >75
    Marginal        40-75
    Poor           < 40

These definitions are derived from the scoring system utilized by the United Nations Economic Commis-
sion for Europe (UN ECE 1987). Nilsson (1989) concluded, on the basis of available literature, that
defoliation up to 25% (i.e., a visual symptoms score of 75) was within the limits of normal stress responses
for conifers; for other species the limit is 10% (i.e., a score of 90).  In the UN ECE system, trees with scores
from 40 to 75 are classified as "moderately damaged" (here, "marginal") and trees with scores less than 40
are classified as "severely damaged" (here, "poor").

The UN ECE commonly reports visual symptoms scores in terms of individual trees.  EMAP-Forests is
extending the categories developed for individual trees to per-plot species, and per-plot overall, average
scores. It is possible that a given location could indicate one condition if only one species is considered,
but another condition if the overall average is considered  instead.

Core tables

    Descriptive statistics
    The following tables will be prepared on the basis of the per-plot average visual symptoms scores.

    1.  Minimum, maximum, median, quartiles, mean, and standard deviation of species average scores, by
       species
    2.  Minimum, maximum, median, quartiles, mean, and standard deviation of number of trees per species
       average score, by species
    3.  Minimum, maximum, median, quartiles, mean, and standard deviation of overall average scores, by
       forest type
    4.  Minimum, maximum, median, quartiles, mean, and standard deviation of number of species, and
       number of sample trees, per overall average score,  by forest type

    Data quality statistics
    Data quality statistics will be reported for the following measured variables:

       1.  species
       2.  percent of full foliation

    The data quality of mensurational variables used for developing expansion (or weighting) factors is
    reported in Appendix A.3.  See Appendix D for additional discussion of data quality statistics.

Cumulative distribution functions

    The status of forest populations of interest as measured by the visual symptoms indicator will be
    displayed in a series of cumulative distribution functions (CDFs). Each CDF charts the proportion of plots
    with average (by species or overall) visual symptoms  scores less than  or equal to some average score.
    Upper confidence limits are also shown on these CDFs. Associated with each CDF, the estimated
    proportion of plots with average (by species or overall) visual symptoms scores 40, and the estimated
    proportion with scores 75, are also given.  The CDFs may be inspected to estimate population proportions
   with alternate cutoff values.
                                               22

-------
Per-plot overall average visual symptoms scores are to be reported by forest type in a format illustrated
in Figure A.2-1. Per-plot species average visual symptoms scores are to be reported by species for in a
format illustrated in Figure A.2-2.
                                             23

-------
                                     ET
               POPULATION: SPRUCE-FIR FOREST TYPE PLOTS.
               VARIABLE: PLOT AVERAGE % TREE FOLIATION
               POPULATION EXTENT: 12.1 MILLION ACRES
               SAMPLE SIZE: 53 PLOTS, WITH 212 TREES SAMPLED.
               1 APRIL 2001
                                                     4.6% IN "GOOD
                                                     CONDITION
         37.6% IN "POOR"
         CONDITION
                                                                  100
                    PLOT AVERAGE PERCENT TREE FOLIATION
FIGURE A.2-1. HYPOTHETICAL EXAMPLE OF FOREST CONDITION STATISTICS:
CUMULATIVE DISTRIBUTION FUNCTION, F(x), OF PER-PLOT AVERAGE VISUAL
VISUAL SYMPTOMS SCORES FOR THE SPRUCE-FIR FOREST TYPE IN EPA
REGIONS I, II, AND III, WITH UPPER CONFIDENCE LIMITS ON F(x).

-------
                POPULATION: BALSAM FIR TREES ON ALL PLOTS.
                VARIABLE: PLOT AVERAGE % TREE FOLIATION
                POPULATION SIZE: 375 MILLION TREES.
                SAMPLE SIZE: 78 PLOTS, WITH 423 TREES SAMPLED.
                1 APRIL 2001
                  T
                  20           40           60

                    PLOT AVERAGE PERCENT TREE FOLIATION
                                                                     - 0.0
                                                                     - 0.2
                                                                     - 0.4
; 4.6% IN "GOOD
! CONDITION
         37.6% IN "POOR"
         CONDITION
                                                                     - 0.6
                                                                     - 0.8
               100
FIGURE A.2-2. HYPOTHETICAL EXAMPLE OF FOREST CONDITION STATISTICS:
CUMULATIVE DISTRIBUTION FUNCTION, F(x), OF PER-PLOT AVERAGE VISUAL
VISUAL SYMPTOMS SCORES FOR BALSAM FIR IN EPA REGIONS I, II, AND III,
WITH UPPER CONFIDENCE LIMITS ON F(x).

-------
                                    A.3 Growth Efficiency

Rationale

Tree stem growth occurs when assimilated carbon is allocated to stemwood structure. This allocation of
carbon has relatively low priority relative to other carbon sinks, and so stemwood growth is a relatively
sensitive indicator of how well a given tree or stand of trees is functioning as a whole. In most cases, if
trees are not growing (i.e., are not accumulating stemwood), then they are in a relatively poor condition
insofar as the probability of their continued presence is concerned. Thus, tree growth is a measure of
forest vigor.

Most trees are under stresses which limit growth all the time. Growth efficiency (e.g., Waring and
Schlesinger 1985) is one of the most sensitive measures of the combined effects of those stresses; it
measures the allocation of assimilated carbon to stemwood growth per unit of leaf area. There are normal
cycles of growth efficiency for individual trees as stands regenerate, mature, and senesce (Matson and
Boone 1984). In general, trees and stands that have low growth efficiencies are more attractive targets for
insects and other pathogens (Waring 1983).  Growth efficiency also reflects vigor changes in response to
air pollution stresses (Oren et al. 1988).

People care about growth efficiency for two main reasons. First, there is economic value in wood
production, and wood is not produced efficiently when growth efficiency is low. Second, aesthetic and
cultural values of stands are partly based on their apparent vigor or health. Biologists care about growth
efficiency because it is a good integrative measure of a stand's ability to maintain itself in an ecosystem.

Definition and measurement

Growth efficiency has been defined differently as the capabilities to measure its components have  evolved.
The definition suggested by EMAP is the ratio of periodic gross total stemwood biomass growth per unit
area to leaf area index of the tree and tall shrub canopy at the time of full leaf-out.

The numerator of growth efficiency is estimated from measurements of tree dimensional growth. Usually
these measurements include diameter at breast height (dbh) and tree height (sometimes to a minimum top
diameter), but regional and local variations are common. Tree dimensional growth is then converted into
biomass (dry weight) change by using appropriate factors  (e.g.,  Hansen et al. 1990), and the results are
expressed on a per-unit-area basis.  The denominator of the growth efficiency indicators is estimated from
measurements of the average percentage of photosynthetically active radiation (PAR, 400-700 nm)
absorbed by the canopy (%APAR). %APAR measurements are converted to leaf area index according to
the Beer-Lambert Law, and an extinction coefficient of 0.5 is assumed. (Note: conversion methods sug-
gested here are preliminary approximations which are expected  to be refined with experience and
research.)

Sample selection

Growth efficiency is measured on a fixed-area plot at each  monitoring location. The specific plot design
varies in different regions of the United States, because EMAP is utilizing the different plot designs  which
have been established in each region by the USDA Forest Service. For every design, it is possible to
estimate comparable per-unit area biomass growth statistics. %APAR is a percentage which is obtained
by comparing measurements of PAR under the canopy defined by the fixed-area plot at each location with
simultaneous measurements made at an open (i.e., not under the canopy) and  nearby location.
                                               26

-------
The following cutoff values will be used to classify condition.

    Condition      Value fg/yr/m2)
    Good              >100
    Marginal            40-100
    Poor               <40(

These definitions are general estimates obtained from the literature. They were reported for individual
trees, but they will be applied to plot average growth efficiency estimates by EMAP-Forests.

Core tables

    Descriptive statistics
    The following tables will be prepared on the basis of the per-plot average growth efficiency values.

    1.   Minimum, maximum, median, quartiles, mean, and standard deviation of growth efficiency, by
        forest type
    2.   Minimum, maximum, median, quartiles, mean, and standard deviation of volume growth per acre,
        by forest type
    3.   Minimum, maximum, median, quartiles, mean, and standard deviation of %APAR, by forest type

    Data quality statistics
    Data quality statistics will be reported for the following measured variables.

    1.   Tree diameter at breast height (dbh) and  height or bole length
    2.   Ambient (i.e. not under the tree canopy) PAR
    3.   Under-canopy PAR

    See Appendix D for additional discussion of data quality statistics.

Cumulative distribution functions

    The status of sampled populations as measured by the growth efficiency indicator will be displayed in a
    series of CDFs.  Each CDF charts the proportion of plots with growth efficiency values less than or equal
    to some average score. Upper confidence limits are shown on these CDFs.  Associated with each CDF,
    the estimated proportion of plots with growth efficiency values 40, and the estimated proportion with
    values 100, are also  given. The CDFs may be inspected to estimate population proportions with alternate
    cutoff values.

    Per-plot growth efficiency values are to be reported by forest type in a format illustrated in Figure A.3-1.
                                               27

-------
                POPULATION: RED-WHITE-JACK PINE FOREST TYPE PLOTS.
                VARIABLE: PLOT GROWTH EFFICIENCY (g/m2/ha)
                POPULATION SIZE: 8.6 MILLION ACRES.
                SAMPLE SIZE: 50 PLOTS OF ONE-FIFTH ACRE EACH.
                1 APRIL 2001
                                                 41.1% IN "GOOD
                                                 CONDITION
        9.7% IN "POOR
        CONDITION
                                                                     t- 0.8
                                                                     - 1.0
                                           120
                           PLOT GROWTH EFFICIENCY
160
200
FIGURE A.3-1. HYPOTHETICAL EXAMPLE OF FOREST CONDITION STATISTICS:
CUMULATIVE DISTRIBUTION FUNCTION, F(x), OF PER-PLOT GROWTH
EFFICIENCY VALUES FOR RED-WHITE-JACK PINE FOREST TYPE PLOTS IN EPA
REGIONS I, II, AND III, WITH UPPER CONFIDENCE LIMITS ON F(x).

-------
                    Appendix B.  Exposure/Habitat Indicators

This appendix describes the "exposure/habitat" indicators which are measured at locations sampled by
EMAP-Forests, and which can be used to describe the status and trends of stresses which might be
associated with forest condition.  In most cases, the indicators actually comprise a set of measurements
rather than a single number. A brief rationale for each indicator, or set of measurements, is followed by a
general description of the methods of sampling and measurement.

This appendix will include two types of "core" (standard) tables which will be prepared for each EMAP
reporting region: (1) descriptive statistics of the measured variables, and (2) data quality statistics for the
measured variables. These data are not utilized in the main body of the annual statistical summary
because it has not proven possible to define useful summary indicators for that purpose.

Some of the indicators described in this section may, upon further development, be utilized later as
response indicators to classify forest condition.  In most cases, experience or a research basis for the
construction and interpretation of a suitable index is required before measurements can be used in this
fashion. This is an ongoing task within EMAP and other programs.
                                               29

-------
                          B.1 Foliar Nutrients and Contaminants

Rationale

Healthy functioning of forests depends on a sufficient supply and the proper balance of critical nutrients in
foliage, and on the absence of contaminants which interfere with normal functioning. Shortages ,
imbalances, or contaminations evident in foliar chemistry may be related to changes in forest condition
which are identified by response indicators. Thus, it is important to obtain measurements of key chemicals
in foliage. It is also useful to make these measurements because of the potential for developing an index
of foliar  nutrition as a response indicator.

Definition and measurement

Foliage  chemicals can be measured on current-year foliage (or perhaps leaf litter) during an index
sampling period. Standard laboratory procedures exist for the measurements. Units of expression are
shown in Table B-1.

Sample selection

Plot and tree sampling rules have not yet been specified, although it is expected they will generally follow
the protocols established for the Visual Symptoms indicator as described in Appendix A.2. It is also
expected that foliar samples from a  given plot location will be composited by species for laboratory
analyses. It should then be possible to develop per-plot species averages and per-plot overall averages,
although the latter may not be particularly useful unless expressed in some index form that gives ap-
propriate weights to different species.

Core tables

    Descriptive statistics
    The following tables will be prepared on the basis of the per-plot species average concentration data.

    Minimum, maximum, median, quartiles, mean, and standard deviation of dry weight concentrations
    of the elements listed in Table B-1, by species

    Data quality statistics
    Data quality statistics will be reported for the measured variables shown in Table B-1.

Cumulative distribution functions

    It will be possible to display cumulative distribution functions of the variables in Table B-1, by species and
    perhaps by forest type.  The format would be similar to Figure A.2-1, except that categories of "good",
    "marginal", and "poor" conditions cannot be identified.
                                               30

-------
         Table B-1. Foliar Analytical Parameters.

Parameter     Description of Parameter

TOT_N        Total nitrogen, (wt.%)
TOT_S        Total sulfur, (mg/kg)
KJL_N        Kjeldahl nitrogen, (wt.%)
TOT_K        Total potassium, (wt.%)
TOT_P        Total phosphorus, (wt.%)
TOT_CA       Total calcium, (mg/kg)
TOT_MG       Total magnesium, (mg/kg)
TOT_B        Total boron,  (mg/kg)
TOT_NA       Total sodium, (mg/kg)
TOT_CU       Total copper, (mg/kg)
TOT_ZN       Total zinc, (mg/kg)
TOT_FE       Total iron, (mg/kg)
TOT_AL       Total aluminum,  (mg/kg)
TOT_MN       Total manganese, (mg/kg)
TOT_CD       Total cadmium,  (mg/kg)
TOT_CR       Total chromium, (mg/kg)
TOT_PB       Total lead, (mg/kg)
TOT_V        Total vanadium,  (mg/kg)
TOT_HG       Total mercury, (mg/kg)
EXT CL       Extractable chloride,  (mg/kg)
                              31

-------
                                     B.2 Soil Productivity

Rationale

Healthy functioning of forests and forest soils depends on a sufficient supply and the proper balance of
critical nutrients in soils, on the absence of contaminants which interfere with normal functioning, and on
the existence of appropriate physical characteristics.  Shortages, imbalances, or contaminations evident in
soil chemistry may be related to changes in forest condition which are identified by response indicators.
Thus it is important to obtain measurements of key soil chemicals and physical structure. It is also useful
to make these measurements because of the potential for developing an index of soil productivity as a
response indicator.

Definition and measurement

Soil chemicals and physical variables can be measured on composited samples of soils from several soil
horizons. Standard field and laboratory procedures exist for the measurements. Units of expression are
shown in Table B-2.

Sample selection

Soil nutrients are measured on samples taken near each location which is identified for forest monitoring.
As described earlier, this is essentially a systematic sample of forested locations and thus of forest soils.
At any plot location, four soil sampling stations are located adjacent to the permanent plot which has been
established for tree remeasurements (see Appendix A.3).  The precise locations of these stations will vary
among regions because the permanent tree plot design varies among regions.  At each station, a 50-cm
diameter hole is excavated to a depth of 100 cm, or to hardpan or bedrock, whichever is shallower.

A composite sample of approximately 3 kg (approximately 750 g per sample station)  is made up from
each of the organic (excluding  Oi), A, E, B, and C horizons. In addition, soil  depth and horizon thickness
are recorded at each station, and a detailed description of profile characteristics is made at one station.
Excavated holes are filled with inert materials, and samples are cold-stored and shipped to a laboratory for
analysis.

Core tables

   Descriptive statistics
   The following tables will be prepared on the basis of the per-plot averages.

   Minimum, maximum,  median, quartiles, mean, and standard deviation of dry weight concentrations of
   the variables listed  in Table B-2, by forest type

   Data quality statistics                            _,.,,_
   Data quality statistics will be reported for the measured variables shown in Table B-2.

Cumulative distribution functions

   It will be possible to display cumulative distribution functions of the variables in Table B-2, by forest type.
   The format would be similar to Figure A.2-1, except that categories of "good", "marginal",  and "poor"
   conditions cannot be identified.
                                               32

-------
Parameter
Table B-2. Soil Preparation and Analytical Parameters.

            Description of Parameter
RF_FG        Fine gravel is the portion of the rock fragments in the soil with particle diameter between
              2 mm and 4.75 mm. (wt. %)
RF_MG        Medium gravel is the portion of the rock fragments in the soil with particle diameter between
              4.75 mm and 20 mm.  (wt. %)
BIO_S         Soil biomass (organic matter) is the sum weight of all organic constituents (fresh,
              decomposing, and decomposed) in a given area or volume of forest floor material, (wt %)
BD_COR       Bulk density is the density of the oven-dry, mm soil (minus rock fragments), (g/cm   )
MOIST_A      Air-dry soil moisture, (wt. %)
SAND         Sand is the portion of the sample with particle diameter between 0.05 mm and 2.0 mm.  (wt. %)
SILT          Silt is the portion of the sample with particle diameter between 0.002 mm and 0.05 mm.  (wt. %)
CLAY         Clay is the portion of the sample with particle diameter less than 0.002 mm. (wt. %)
PHJH20       Ph in deionized water.  (Ph units)
PH_01         pH is determined in a 0.01 M calcium chloride solution. (pH units)
CA_CL        Exchangeable calcium.  (meq/100g)
MG_CL        Exchangeable magnesium.  (meq/100g)
K_CL         Exchangeable potassium. (meq/100g)
NA_CL        Exchangeable sodium.  (meq/100g)
FE_CL        Exchangeable iron. (meq/100g)
AL_CL        Exchangeable aluminum. (meq/100g)
CEC_CL       Cation exchange capacity.  (meq/100g)Table B-2, continued
AC_BACL      Total exchangeable acidity. (meq/100g)
N_MIN        Mineralizable nitrogen, (mg/kg)
P_B1          Extractable phosphorus, (mg/kg)
S04_H2O      Extractable sulfate. (mg/kg)
CJTOT        Total carbon, (wt.%)
N_TOT        Total nitrogen, (wt.%)
S_TOT        Total sulfur, (wt.%)
FE_TOT       Total iron.  (wt%)
MN_TOT       Total manganese, (wt.%)
CU_TOT       Total copper, (wt.%)
ZN_TOT       Total zinc,  (wt.%)
B_TOT        Total boron,  (wt.%)
PB_TOT       Total lead, (wt.%)
CD_TOT       Total cadmium, (wt.%)
NI_TOT        Total nickel,  (wt.%)
CR  TOT       Total chromium,  (wt.%)
                                            33

-------
                             Appendix C.  Ancillary Data

This Appendix describes ancillary data that will probably be collected at the monitoring locations.
Although EMAP focuses on indicators of response and exposure to motivate particular measurements, it is
recognized that certain other measurements are motivated by common sense and experience, even if they
are not utilized to construct an indicator (e.g., USDA Forest Service 1985; NSEPB 1985; UN ECE 1987;
Nordic Council of Ministers 1988; Magasi 1988). Descriptions of site and stand features and observations
of biotic stresses must be made because subsequent interpretations of indicators would not be reliable
without this information. Ancillary data may also corroborate inferences about forest condition or stresses
which have been drawn from the indicators. A brief rationale for each indicator, or set of measurements, is
followed by a general description of the methods of sampling and measurement.

This appendix will include two types of "core" (standard) tables which will be prepared for each EMAP
reporting region: (1) descriptive statistics of the measured variables or any summary indices, and  (2) data
quality statistics for the measured variables. These data are not utilized in the main body of the annual
statistical summary.

In addition to the ancillary data described here, a wealth of data is available from other off-frame monitor-
ing networks, but they are not considered in the annual statistical summary. Examples include weather
data from the National Oceanic and Atmospheric Administration, pollution data from the EPA networks,
soils data from the Soil Conservation  Service's STATSGO  data bases, and forest inventory data from the
USDA Forest Service's Forest Inventory and Analysis data bases.

The listing of a particular measurement in this appendix does not imply EMAP endorsement. Rather, the
measurements that have been compiled are based on the most recent suggestions of the interagency
forest monitoring technical committees responsible for them.
                                               34

-------
                           C.1 Biotic and Abiotic Stress Indices

Rationale

Descriptions of site and stand disturbances are useful for identifying unusual local stress conditions which
are not addressed by the proposed set of exposure/habitat indicators. Many times an apparently
interesting response in growth efficiency or visual symptoms will be associated with a local abiotic event,
such as a storm, or a local biotic event, such as beaver girdling.  This is expected and does not invalidate
the concept of regional summarization of response indicators. However, these types of events must be
considered when interpreting site-by-site changes in forest condition indicators. If these events are not
recorded as they occur, it will be difficult to later sort out what has happened at particular plot locations.

The index period sampling approach of EMAP will sometimes result in not observing transient
disturbances which occur during a season other than the index sampling period.  The assumption is that, if
the disturbances result in a net change in condition, then that change will be detected by measuring the
response indicators during the index period. Many types of site and stand disturbances can occur, and it
is possible that many plots will be disturbed in one way or another. It is anticipated that regional distur-
bances would continue to be measured by ongoing Forest Service and State personnel, whatever the
season, and efforts are underway to tie EMAP data bases into that additional source of information.

Detailed measurements of biotic and abiotic stresses can be difficult to implement without trained foresters
on every field crew. It is often difficult to distinguish between different types of biotic stresses, for example,
insects versus disease, or between different types of abiotic stresses, for example, construction versus
harvesting.  But it is usually possible to distinguish between  biotic and abiotic factors, and  this is reflected
in the suggested summary indices (see below).

Definition and measurement
              i
The categories of stresses which will be considered include  fire, animal damage, weather,  harvest,
construction, insects, disease, and "other" (Alexander and Carlson  1988). Each plot is visually inspected
by trained observers and scored separately for each type (or sometimes the two or three most obvious
types) of disturbance. Four scores are used.

    1    no disturbance
    2   minor disturbance
    3   moderate disturbance
    4   severe disturbance

The meaning of "minor", "moderate", and "severe" depends on standards generally accepted for different
types of stresses in different forest types and regions.  Regional comparisons may have to utilize the
simple "presence or absence" of particular types of disturbances. Later, site-specific investigations will
have access to the precise definitions of standards for particular sites.

Sample selection

The scoring  system generally considers visible evidence of different stress factors in the immediate vicinity
of the plot.  Thus, sample selection is the same as was described for other plot-level indicators in
Appendices A and B.

Two types of summary indices are proposed to summarize the biotic and abiotic stress occurrences at
each monitoring  plot.  The biotic stress index is defined as the sum of the disturbance scores recorded for
animal damage,  insects, or diseases.  The abiotic stress index is similarly defined for fire, weather, harvest,
and construction. These indices assume that disturbances due to each component are scored. Other
types of summary indices should be evaluated during the early years of monitoring.  Note that the
category "other"  is not included  in these summary indices.
                                                35

-------
Core table
    Descriptive
    The following tables will be prepared on the basis of the per-plot indices.

    1.   Minimum, maximum, median, quartiles, mean, and standard deviation of abiotic stress index, by forest
        type
    2.   Minimum, maximum, median, quartiles, mean, and standard deviation of biotic stress index,  by forest
        type

    Data quality statistics
    Data quality statistics will be reported for these measurements.

    1 .   disturbance type
    2.   disturbance score
                                                36

-------
                          C.2 Injury to Pollution "Indicator" Plants

Rationale

Injury to "indicator" plant species can be used as a measure of forest exposure to certain pollutants
(Manning and Feder 1980; Posthumous 1984). Various techniques, ranging from the planting of
 "standard" gardens to the in situ observation of native plants, are possible for regional surveys
 Posthumous 1980).  In forests, techniques in use include single-species and single-pollutant surveys (e.g.,
Bennett and Stolte 1985), and multiple-species and multiple-pollutant surveys (e.g., Alexander and Carlson
1988; Magasi 1988).

Quantitative measures of the degree of pollution injury are generally not available for all pollutants, but
qualitative observations can be made with trained observers.  Individual species are differentially sensitive
to different pollutants, and they exhibit different types of injury symptoms (Skelly et al. 1989).  Different
species occur naturally in different locations, and they can exhibit symptoms at different times of the year.
Finally, the presence, absence, or degree of symptoms is usually not well correlated with either total
deposition or the overall effect on the forest.

Nevertheless, pollution injury to indicator  plants has been, in some cases, a useful supplement to measure-
ments of other abiotic stresses which can be related to other indicators of forest condition.  For example,
the presence or absence of pollution injury to  indicator plants helped to explain forest growth in a  regional
survey of loblolly pine forests (Van Deusen 1989).

Definition and measurement

Different scoring methods have been developed for different pollutants and different species in
experimental situations, but there appears to be no generally accepted definition of "pollution injury" or
how to measure different types of injury in regional surveys. In the  past,  regional forest surveys have
started with a list of species and preselected symptoms to look for; if the species are present at a
monitoring location, they are visually inspected for pollution damage (e.g., Magasi 1988; Alexander and
Carlson 1988). If they are not present, then no information is recorded. In application, lists of species and
symptoms will vary from region to region depending on the state of the science in each situation.

Sample selection

The scoring system generally considers visible evidence of different pollution injury in the immediate
vicinity of the plot. Thus, sample selection is the same as was described for other plot-level indicators in
Appendices A and B. Injuries from different pollutants are usually observed  on different species.
Techniques for estimating summary indices have not been developed.

Core tables

    Descriptive statistics
    The following tables will be prepared on the basis of the per-plot observations of pollution injury to listed
    species.

    Number of plots where listed (indicator) species was observed (with or without injury), and the number of
    plots where pollution injury to listed species was observed, by listed  species and injury type.

    Data quality statistics
    Data quality statistics will be reported for these measurements.

    1.  species (presence or absence)
    2.  injury (presence or absence)
                                                37

-------
                                   C.3 Landscape Indices

Rationale

Mapping and determining the variety, relative abundances, sizes, shapes, and spatial correlations of forest
cover types in a landscape are basic concerns when consideration is given to both extent and change in
vegetation and associated plant and animal composition and diversity (Noss 1983). Techniques for
making such a set of measurements exist and will be implemented by the EMAP-Characterization Task
Group (Muchoney et al. 1990). Given that some data will be available, a regional monitoring system
should consider landscape-level indicators, but there is no general agreement as to what types of in-
dicators are appropriate for EMAP-Forests. It is therefore suggested that the landscape-level indicators of
forest patterns identified by EMAP (Hunsaker and Carpenter 1990) be incorporated into EMAP-Forests
annual statistical summaries as ancillary data, with the expectation that these or other landscape
measures will appear as "indicators" in the future.

Definition and measurement

The basic set of measurements is made by delineating vegetative, physical, and land use categories from
aerial images as described in Appendix A.1.  These "polygon" data are converted to a "raster" (i.e.
point-based) format for computing certain indices.

Sample selection

The sample selection procedures are described in Appendix A.1.

Hunsaker and Carpenter (1990) describe a set of landscape indices which are calculated on a per-'sample-
unit ("per-hexagon") basis.

    Habitat proportions (cover types)
    Patch size and perimeter-to-area ratio
    Contagion or habitat patchiness
    Fractal dimension
    Patton's diversity index
    Abundance or density of key physical features and structural elements

Core tables

    Descriptive statistics
    The following tables will  be prepared on the basis of the per-hexagon indices.

    1.  Minimum, maximum, median, quartiles, mean, and standard deviation of proportion of total sample area
       in a given forest type, by forest type
    2.  Minimum, maximum, median, quartiles, mean, and standard deviation of patch size and perimeter-to-
       area ratio, overall and by forest type
    3.  Minimum, maximum, median, quartiles, mean, and standard deviation of contagion or habitat patchiness
    4.  Minimum, maximum, median, quartiles, mean, and standard deviation of fractal dimension, overall and
       by forest type
    5.  Minimum, maximum, median, quartiles, mean, and standard deviation of Patton's diversity index
    6.  Minimum, maximum, median, quartiles, mean, and standard deviation of abundance or density of key
       physical features and structural elements

    Data quality statistics
   These statistics are developed for variables measured by the EMAP-Cnaractenzation team and are not
   available at this time.
                                               38

-------
                   C.4 Tree/Shrub/Herb Species Number and Variety

Rationale

The concepts of species "richness" and "diversity are fundamental considerations when evaluating the
status and trends of forest condition in terms of composition and biodiversity. Such concepts are difficult
to quantify unambiguously (Magurran 1988) and are therefore difficult to implement as indicators by
EMAP. Nevertheless, given that some data are likely to be available, it is prudent to consider several
plot-level indices of vegetation richness and diversity as ancillary data.

Definition and measurement

Vegetation is often defined and measured in terms of life-forms such as shrubs, vines, herbs, ferns,
mosses, sedges, rushes, grasses, and trees.  The precise definitions and measurements can be expected
to vary among regions, but it is usually possible to obtain average per-unrt area estimates of the frequency
or relative density or canopy coverage of different life forms. In some cases, species are identified.
Measures of species richness and diversity will be calculated for each life-form if species are identified.  If
not, then measures of life-form richness and diversity will be calculated.

A plausible set of assumptions for tree measurements permits species richness and diversity estimation on
the basis of frequency  or basal area per acre by species obtained from the measurements made to es-
timate stand volume growth (see Appendix A.3). For other life-forms, species information is less certain,
but if measured, the basic measurements usually consist of visual estimates of the percentage cover, to
the nearest 5%, of each life-form (or species within life-form) that is observed. These estimates can be
used to construct estimates of richness and diversity indices.

Sample selection

Trees are selected in accordance with the sampling rules as described in Appendix A.3. The measure-
ments of other life-forms generally consider the occurrences of visible species on smaller subplots within
or adjacent to the tree  sampling plots. Thus, sample selection is generally the same as was described for
other plot-level indicators in Appendices A and B. A plethora of candidate indices of richness and diver-
sity, and related measures, are available in standard textbooks.

Core tables

   Descriptive statistics
   The following tables will be  prepared on the basis of the per-plot indices of richness and diversity for any
   or all life-forms.

   1.  Minimum, maximum, median, quartiles, mean, and standard deviation of richness (e.g.,  number of
       species), by forest type
   2.  Minimum, maximum, median, quartiles, mean, and standard deviation of diversity (e.g., Shannon's
       [1948] index),  by forest type

   Data n"a|»v statistics
   Data quality statistics will be reported for the measurements made. Under a plausible scenario, these
   measurements include the following.

   1.  species (if trees, shrubs, or herbs)
   2.  dbh (of trees) or percent cover (of other life-forms)
   3.  life-form
                                               39

-------
                                   C.5 Other Ancillary Data

Rationale

This category of ancillary data includes site and stand descriptors that will be listed without justification so
that reviewers may see that these types of data are being considered as part of monitoring.

Site and stand descriptors

    Plot and tree locations and numbers
    Tree sizes and crown classes
    Tree injuries and locations of injuries
    Specific signs and symptoms of specific insects and diseases
    Visual evaluation of roots
    Tree cores (to provide historic information)
    Physical site descriptors such as slope, elevation, and aspect
    Standard mensurational features
    Selected soil and foliage measures which are not included in Appendix B
                                                40

-------
                        Appendix D. Uncertainty Estimation

Data uncertainty estimation is an integral feature of monitoring design. It is useful to know when a change
in a given indicator is so small, in relation to the uncertainty about the components of the indicator, that
conclusions on the basis of the available information would be erroneous. Thus, the quality and
uncertainty of the data which are collected and reported will be documented in the annual statistical
summary.

Uncertainty is partly due to imprecision or bias in the measurement system arising from, for example,
inconsistent field instrument readings, missing data, unstable analytical laboratory 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 quality assurance program during all phases of measurements.  A quality assurance program
provides the information needed to quantify the measurement uncertainty obtained from the monitoring
data. Uncertainty of extrapolation, on the other hand, is controlled and estimated through application of
statistical principles for sampling and aggregating data to describe populations.

A major data reduction issue is the fact that each additional level of sample aggregation adds a degree of
uncertainty to the results which is dependent upon the appropriateness of the aggregation scheme.
Hence, the actual ability to discern actual changes in indicators over time or space also depends partly on
the classification 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 appendices of the annual statistical summary will provide estimates or indications of measurement
and population uncertainties for each of the individual variables which are measured. At a minimum, this
information will consider the effects of detectability (e.g., system detection limits), sampling error, and
(where appropriate) error propagation as a result of data aggregation. The appendices also attempt to
 relate the overall uncertainty estimates to the project-level data quality objectives that are defined prior to
data collection.

The statistics to be reported  for each variable could include the following.

    1.  Measurement precision, sampling error, and standard errors for population-level estimates, which could
       be reported as measurement uncertainty, sampling uncertainty, and extrapolation uncertainty
   2.  Percent of samples above the measurement system detection limit
   3.  Percent of planned sample size actually obtained  (a measurement of completeness of the sample)
   4.  System detection limits and inherent precision at different magnitudes of measurement
   5.  Ratios of various components of uncertainty
                                               41

-------
  Appendix E. Classification and Aggregation Schemes for Reporting

EMAP-Forests Reporting Regions

An EMAP-Forests reporting region is an area of the United States which is a geographic unit for reporting
EMAP-Forests monitoring data. These regions could be defined in terms of natural features, political
boundaries, or administrative regions.  Because the audience for the annual statistical summary is
primarily the public and the Administrator, the reporting regions have been defined in terms of admini-
strative boundaries. Although this does not take advantage of some efficiencies which could be gained by
using a more natural grouping scheme, those efficiencies can be realized in other types of EMAP-Forests
reports where data can be aggregated differently.

Suggestions for the EMAP-Forests reporting regions (Figure E-1 and Table E-1) are based largely on EPA
Regions (Figure E-2). Exceptions were made so that each reporting region contained mainly either
"eastern" or "western" forest types. The suggested regions generally contain a sufficient number of
sampling units to enable reporting by major forest type (see below). They are also generally similar to the
administrative regions recognized by the USDA Forest Service. The suggested definitions of
EMAP-Forests reporting regions should be evaluated in light of the final sampling design and suggestions
made by the other EMAP task groups.
                                             42

-------
                                                             SCALE uo^mooo
                                                     ALBERS-CONIC  EQUAL-AREA PROJECTION
                                                                                                           EMAP  Reporting   Regions

                                                                                                           NORTHEAST
                                                                                                                Connecticut, Malk, Massachusetts,
                                                                                                                New Hampshire, Rh^de Island, Vermont
                                                                                                                NewUersey, New YoNNc, Delaware,
                                                                                                                Distrrcfxaf Columbia, i^ennsylvonic
                                                                                                                Maryland,  wt^inia,  and %st  Virgin^
     Alabama, Norida, Gexc
     Kentucky, Mississippi,
     North Core I ma,  South  Carolina,
     Jenji/ssee. Arkansas,
     Louisiana, Oklahoma, and Texas

NORTH  CENTRAL
     Illinois,  Indiana, Michigan,
     Minnesota, Ohio, Wisconsin, Iowa,
     Kansas,  Missouri, and  Nebraska,

ROCKY  MOUNTAIN
     Colorado, Montana, North Dakota,
      South Dakota,  Utah,  and Wyoming

SOUTHWEST
     Arizona, California, Hawaii
     New  Mexico, and Nevada

NORTHWEST
     Alaska, Idaho, Oregon, and
     Washington
FIGURE   E-1.    EMAP   REPORTING   REGIONS

-------
                                                                  SCALE  1:20,000,000
                                                          ALBERS CONIC  EQUAL-AREA  PROJECTION
                                                                                   SCALE 1:20,000,000
                                                                            9
     o
                            Alaska
Hawaii
      EPA  Regions

1   Connecticut, Maine,
   Massachusetts,  New Hampshire,
   Rhode Island, and Vermont
2   New Jersey  and  New York
3   Delaware, District of  Columbia,
   Maryland, Pennsylvania,
   Virginia, and West Virginia
4   Alabama, Florida,
   Georgia, Kentucky,
   Mississippi,  North Carolina,
   South Carolina, and  Tennessee
5   Illinois,  Indiana,
   Michigan, Minnesota,
   Ohio, and Wisconsin
6   Arkansas, Louisiana,
   New Mexico, Oklahoma,
   and Texas
7   Iowa, Kansas,
   Missouri, and Nebraska,
8   Colorado, Montana,
   North Dakota,  South Dakota,
   Utah, and Wyoming
9   Arizona, California,
   Hawaii,  and  Nevada
10 Alaska,  Idaho,
   Oregon,  and Washington
FIGURE   E-2.     EPA   REGIONS

-------
                      Table E-1. EMAP-Forests reporting regions.
EMAP Reporting Region
Northeast
South
North Central
Rocky Mountain
Southwest
Northwest
EPA Regions Included in Reporting Region
Regions I, II (excluding PR), and III
Regions IV and VI (excluding NM)
Regions V and VII (including eastern halves
of ND and SD, and excluding northwest
portion of NE)
Region VIII (excluding eastern halves of
ND and SD, and including northwest portion
ofNE)
Region IX (including NM)
Region X
EMAP-Forests Forest Type Classifications

Within any given reporting region, monitoring data will be aggregated by major forest types (Table E-2).
The forest types are defined after the classification scheme described by Eyre (1980). This is a relatively
gross classification scheme that is useful for regional summaries such as the annual statistical summary,
but it may not be appropriate for other types of EMAP-Forests reports. An advantage of this classification
scheme is that it is directly comparable to the national classification scheme utilized by the USDA Forest
Service for national assessments (e.g., USDA Forest Service 1982).
                                              45

-------
       Table E-2.  Forest types to be reported, by EMAP Reporting Region.
 Eastern Forest Types
EMAP Reporting Region

White-red-jack pine
Spruce-fir
Longleaf-slash pine
Loblolloy-shortleaf pine
Oak-pine
Oak-hickory
Oak-gum-cypress
Elm-ash-cottonwood
Maple-beech-birch
Aspen-birch
Northeast South
X
X
X
X X
X
X X
X

X
X
North Central
X
X



X

X
X
X
Western Forest Types
                     Rocky
                     Mountain
      South West
Northwest
Douglas-fir X
Hemlock-Sitka spruce
Redwood
Ponderosa pine X
White pine
Lodgepole pine X
Larch X
Fir-spruce X
Western hardwoods X
Pifiyon-juniper X
Alaskan interior
Hawaiian Ohia
X

X
X

X

X
X
X

X
X
X

X
X
X
X
X
X

X

                                       46

-------
                             Appendix F. Literature Cited

Abrahamson, D.E. (ed). 1989. The Challenge of Global Warming. Island Press, Washington, DC.

Alexander, S. and Carlson, J.  1988. Visual damage survey pilot test project manual.  Department of Plant
Pathology, Physiology and Weed Science, Virginia Polytechnic Institute and State University, Blacksburg,
VA.

American Forestry Association.  1987. White Paper on the Forest Effects of Air Pollution. American
Forestry Association, Washington, DC.

Bennett, J.P. and Stolte, K.W.  1985.  Using vegetation biomonitors to assess air pollution injury in national
parks: Milkweed survey. Natural Resources Report Series No. 85-1, National Park Service, Air Quality
Division, Denver, CO.

EPA. 1989a. Environmental Monitoring and Assessment Program - EMAP. U.S. Environmental
Protection Agency, Office of Research and Development, Washington, DC.

U.S. EPA. 1989b.  Design report for the Environmental Monitoring and Assessment Program. U.S.
Environmental Protection Agency, Corvallis, OR. (Draft)

U.S. EPA. 1990. Draft integrated assessment strategy.  U.S. Environmental Protection Agency, Office of
Research and Development, Environmental Monitoring and Assessment Program, February 1990. 24 pp.

Eyre, F.H. (Editor). 1980. Forest Cover Types of the United States and Canada.  Society of American
Foresters, Washington, DC.

Hansen, M.H., Frieswyk, T., Glover, J.F., and Kelly, J.F.  1990.  The eastwide forest inventory data base:
Data base description and user's manual. USDA Forest Service, General Technical Report WO-GTR-	
(In press). Washington,  DC.

Hazard, J.W. and Law, B.E.  1989.  Forest survey methods used in the USDA Forest Service.  Report
EPA/600/3-89/065, U.S. Environmental Protection Agency, Corvallis, OR. NTIS Number PB89 220 594/AS.

Hunsaker, C.T. and Carpenter, D.E. (Editors).  1990. Ecological indicator report for the Environmental
Monitoring and Assessment Program. U.S. Environmental Protection Agency, Atmospheric Research and
Environmental Assessment Laboratory, Research Triangle Park, NC.

Lund, H.G. 1986. A primer on integrating resource inventories. USDA Forest Service, General Technical
Report WO-49.  64 pp. Washington, DC.

Magasi, LP.  1988. Acid Rain National Early Warning System. Manual on Plot Establishment and
Monitoring.  Inf. Rep. DPC-X-25, Canadian Forestry Service, Ottawa.

Magurran, A.E.  1988.  Ecological Diversity and Its Measurement. Princeton University Press, Princeton,
NJ.

Manning, W.J. and Feder, W.A. 1980. Biomonitoring Air Pollutants with Plants.  Applied  Science
Publishers, London, UK.

Matson, P.A. and Boone, R. 1984. Natural disturbance and nitrogen mineralization:  Wave form dieback
of mountain hemlock in the Oregon Cascades. Ecology 65:1511-1516.
                                              47

-------
Mclaughlin, S.B. 1985. Effects of air pollution on forests - A critical review. Journal of the Air Pollution
Control Association 35:512-534.

Muchoney, D.M., Kisner, D.E., Stout, K.K., and Norton, D.J. 1990. Interim report: An ecologically-based
land use/ land cover classification for landscape characterization.  U.S. Environmental Protection Agency,
Las Vegas, NV.  (Draft)

NAPAP (National Acid Precipitation Assessment Program).  1988.  Effects on Forests. Chapter 7 in Volume
4 of Interim Assessment: The Causes and Effects of Acidic Deposition, National Acid Precipitation Assess-
ment Program, Washington, DC.

MAS (National Academy of Sciences).  1983. Changing climate report of the carbon dioxide assessment
committee. National Academy Press, Washington, DC.

Nilsson, S. 1989. Forest decline in  Europe attributed to air pollutants (data to 1987). United Nations,
Economic Commission for Europe and the Food and Agriculture Organization, Report Number
ECE/TIM/46, United Nations, New York, NY.

Nordic Council of Ministers. 1988. Guidelines for Integrated Monitoring in the Nordic Countries.  Steering
Body for Environmental Monitoring, Nordic Council of Ministers, Copenhagen, Denmark.

Noss, R.F. 1983. A regional landscape approach to maintain diversity. BioScience 33:700-706.

NSEPB (National Swedish  Environmental Protection Board). 1985. Monitor 1985: The  National Swedish
Environmental Monitoring Programme (PMK). National Environmental Protection Board, Research and
Development Department,  Environmental Monitoring Section, Solna, Sweden.
Oren,  R., Schulze, E.D., Werk, K.S, Meyer, J., Schneider, B.U., and Heilmeier, H. 1988.  Performance of
two Picea abies (L) Karst.  stands at different stages of decline.  I. Carbon relations and stand growth.
Oecologia 75:25-37.

Posthumous, A.C. 1980. Elaboration of a communitive methodology for the biological  surveillance of air
quality by the evaluation of the effects on  plants. United Nations, Economic Commission for Europe.
Report EUR 6642 EN, Brussels, Belgium.

Posthumous, A.C. 1984. Monitoring levels and effects of air pollutants.  Chapter 5 in Treshow, M. (ed.), Air
Pollution and Plant Life, John Wiley, New York, NY.

Shannon, C.E.  1948. A mathematical theory of communication. Bell System Technical Journal 27:379-
423, 623-656.

Skelly, J.M., Davis, D.D., Merrill, W.,  Cameron, E.A., Brown, H.D., Drummond, D.B., and Dochinger, LS.
(Editors).  1989. Diagnosing injury to eastern forest trees. USDA Forest Service, National Vegetation Sur-
vey, Research Triangle Park, NC.

UN ECE.  1987.  Manual on methodologies and criteria for harmonized sampling, assessment, monitoring,
and analysis of the effects  of air pollution  on forests. United Nations, Economic Commission for Europe,
Convention on Long-Range Transboundary Air Pollution, International Cooperative Programme on Assess-
ment and Monitoring of Air Pollution Effects on Forests.

USDA Forest Service. 1982. An analysis of the timber situation in the United States 1952-2030.  USDA
Forest Service, Forest Resource Report Number 23, Washington, DC.

USDA Forest Service. 1985. Forest resource inventory: An overview. USDA Forest Service, Forest
Resource Economics Research Staff Report, Washington, DC.
                                              48

-------
Van Deusen, P.O.  1989.  Indicator plants in forest health surveys. In Proceedings of the IUFRO
Conference on Advanced Forest Inventory Methods, Syracuse, NY.

Waring, R.H. 1983. Estimating forest growth and efficiency in relation to canopy leaf area. Advances in
Ecological Research 13:327-354.

Waring, R.H. and Schlesinger, W.H. 1985. Forest Ecosystems: Concepts and Management. Academic
Press, Orlando.  340 pp.
                                             49

-------