United States
Environmental Protection
Agency
Office of
Research and Development
Washington, DC 20460
EPA 600/4-917013
May 1991
EPA
Agroecosystem
Monitoring and
Research Strategy
Environmental Monitoring and
Assessment Program
-------
EPA/600/4-91/013
May 1991
Environmental Monitoring and Assessment Program
Agroecosystem Monitoring and Research Strategy
by
Walter W. L k, Technical Director
C. Lee C • ell, Associate Director
Robei T -eckenridge
Gerald E. Byers
Alva L. Finkner
George R. Hess
Julie R. Meyer
Thomas J. Moser
Steven L. Peck
John O. Rawlings
Charles N. Smith
with contributions from:
Craig M. Hayes
Virginia M. Lesser
Deborah A. Neher
Gail L. Olson
This study was conducted in cooperation with
U.S. Department of Agriculture
Agricultural Research Service
Raleigh, NC 27711
U.S. Environmental Protection Agency
Office of Research and Development
Washington, D.C. 20460
Environmental Monitoring Systems Laboratory
Las Vegas, NV 89193
Atmospheric Research and Exposure Assessment Laboratory
Research Triangle Park, NC 27711
Environmental Research Laboratory
Athens, GA 30613
Environmental Research Laboratory
Corvallis, OR 97333
Printed on Recycled Paper
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, DC 20402
S/N 055-000-00417-0
-------
NOTICE
The information in this document has been funded wholly or in part by the
United States Environmental Protection Agency under interagency agreement
DW12934170 with the U.S. Department of Agriculture, Agricultural Research Service,
Contract No. 68-CO-0049 to Lockheed Engineering & Sciences Company, Contract No.
68-C8-0006 to ManTech Environmental Technology, Inc., in Corvallis, Oregon and
Contract No. 68-DO-0106 to ManTech Environmental Technology, Inc. in Research
Triangle Park, North Carolina. It has been subjected to the Agency's peer and
administrative review, and it has been approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute endorsement
or recommendation for use.
Proper citation of this document is:
Heck, W.W., C.L. Campbell, G.R. Hess, J.R. Meyer, T.J. Moser, S.L. Peck, J.O. Rawlings
and A.L. Finkner. 1991. Environmental Monitoring and Assessment Program (EMAP) -
Agroecosystem Monitoring and Research Strategy. EPA/600/4-91/013. U.S.
Environmental Protection Agency, Washington, D.C.
- n -
-------
Table of Contents
List of Figures vii
List of Tables ix
Glossary of Acronyms x
Acknowledgements xi
1. Introduction 1 1
1.1. Overview of the Environmental Monitoring and Assessment Program
(EMAP) 1 1
1.2. Definition and Importance of Agroecosystems 1 2
1.3. Overview of the EMAP-Agroecosystem Program 1 6
2. Program Approach and Rationale 2 1
2.1. Overview 2 1
2.2. Overview of Monitoring Design 2 4
2.3. Use of Existing Data 2 4
2.4. Indicators of Agroecosystem Condition 2 5
2.5. Overview of Implementation Plans 2 7
2.6. Cooperation with other agencies 2 8
2.7. Ecosystem Linkages 2 8
2.8. Relationship to Ecological Risks and Decision Making 2 9
2.9. Expected Outputs 2 10
3. Design and Statistical Considerations 3 1
3.1. Monitoring Network Design 3 1
3.2. General Statistical Sample Survey Requirements 3 1
3.3. Overview of the EMAP Sampling Design 3 2
3.4. Overview of the NASS Sampling Design 3 2
3.5. EMAP-Agroecosystem Tier 2 Sampling 3 7
4. Field Sampling 4 1
4.1. Introduction 4 1
4.2. Statistical Issues 4 1
4.3. Field Sampling Techniques 4 3
4.4. Evaluation of Field Sampling Techniques 4 6
5. Data Analysis, Integration and Assessment 5 1
5.1. Data Summaries 5 1
5.2. Interpretive Analyses 5 1
5.3. Ancillary Data 5 3
5.4. Integration of Information 5 3
5.5. Assessment Activities 5 4
in
-------
6. Indicator Development and Evaluation 6 1
6.1. Assessment Endpoints and Indicators 6 1
6.2. Indicator Categories and Conceptual Models 6 3
6.3. Framework for Indicator Development and Evaluation 6 5
6.4. Historical Overview of Indicator Development for EMAP Agroecosyste . 6 10
6.5. Indicators Addressing Sustainability 6 11
6.6. Indicators Addressing Contamination of Natural Resources 6 19
6.7. Indicators Addressing the Quality of Agricultural Landscapes 6 26
6.8. Integration of Indicator Data 6 32
6.9. Conclusions 6 33
7. Logistics 7 1
7.1. Logistics Implementation Components 7 1
7.2. Logistics Issues 7 1
7.3. Review of Logistics 7 3
8. Analytical Considerations 8 1
8.1. Analytical Methods and Coordination of Methods within EMAP 8 1
8.2. Data Quality Objectives 8 4
8.3. Ancillary Data 8 4
8.4. Laboratory and Field Support 8 4
9. Quality Assurance Program 9 1
9.1. Quality Assurance Policies 9 1
9.2. Total Quality Management 9 2
9.3. Quality Assurance Coordination Roles 9 3
9.4. Quality Assurance Objectives 9 4
9.5. Data Quality Objectives (DQOs) 9 6
9.6. Quality Assurance Project Plans 9 10
9.7. Standard Operating Procedures 9 11
9.8. The Audit Program 9 14
9.9. Documentation 9 14
9.10. Summary 9 17
10. Information Management 10 1
10.1. Overview 10 1
10.2. Relationship with USDA-NASS 10 2
10.3. Use of Existing Data 10 3
10.4. Quality of Data 10 3
10.5. Confidentiality of Data 10 5
10.6. Information Management Objectives for Pilot Program 10 9
10.7. Information Management Resource Requirements for the Pilot Program 10 10
IV
-------
11. Research and Monitoring Accomplishments and Plans (1990-1995) 11 1
11.1. Rationale 11 1
11.2. Accomplishments Through 1990 11 2
11.3. Identification and Use of Data from other Monitoring Efforts 11 5
11.4. Pilot, Demonstration and Implementation Plans 11 6
11.5. Schedule of Activities 11 11
11.6. Resources and Plans 1991 Through 1995 11 11
Literature Cited L 1
Appendix 1: Agroecosystem Resource Group Members Al 1
Appendix 2: Agroecosystem Peer Review Panel A2 1
Appendix 3: Summary of Selected Existing Databases A3 1
Appendix 4: Landscape Characterization A4 1
Appendix 5: Indicator Data Needs and Data Sources A5 1
Appendix 6: NASS Questionnaire for Pilot Survey A6 1
Appendix 7: Examples of Data Summary Tables A7 1
Appendix 8: Indicator Fact Sheets A8 1
A8.1 Crop Productivity A8.1 1
A8.2 Soil Productivity A8.2 1
A8.3 Land Use and Landscape Descriptors A8.3 1
A8.4 Habitat Linear Classification System and Habitat Layer Index A8.4 1
A8.5 Irrigation Water Quality A8.5 1
A8.6 Irrigation Water Quantity A8.6 1
A8.7 Agricultural Chemical Use A8.7 1
A8.8 Nonpoint Source Loading A8.8 1
A8.9 Biomonitors A8.9 1
A8.10 Pest Density A8.10 1
A8.ll Density and Diversity of Beneficial Insects A8.ll 1
A8.12 Socioeconomics A8.12 1
A8.13 Livestock Production A8.13 1
A8.14 Genetic Diversity A8.14 1
Appendix 9: Integration Across EMAP Ecosystems A9 1
-------
List of Figures
Figure 1.1: Conceptual model of an agroecosystem ............................... 1 3
Figure 1.2: Distribution of important areas for two major commodities a) com and b) wheat. ... 14
Figure 1.3: Process used to define expectations, goals, and objectives of the monitoring program. ... 1 10
Figure 2.1: U.S. Environmental Protection Agency Regions .......................... 2 2
Figure 2.2: Concept of a four-tiered approach in EMAP ............................ 2 3
Figure 2.3: EMAP provides a foundation fitting into the ORD's Ecological Risk Assessment Process. . . 26
Figure 3.1: The EMAP Hexagon grid ..................................... 2 15
Figure 4.1: Selection of field within segment .................................. 4 3
Figure 5.1: Analytical approach with agroecosystem indicator data ...................... 5 2
Figure 6. 1 : Agroecosystem assessment endpoints that will be addressed with a suite of indicators to
determine the status and trends in agroecosystem health ..................... 6 2
Figure 6.2: Examples of agroecosystem processes linking assessment endpoints with response,
exposure and stressor indicators .................................. 6 6
Figure 6.3: Six phases of indicator development followed in the Agroecosystem program ......... 6 8
Figure 6.4: Factors to be included in the calculation of a crop productivity index ............. 6 13
Figure 6.5: Components of a soil productivity indicator ........................... 6 14
Figure 6.6: Multi-media distribution of contaminants and potential routes of biotic exposure ....... 6 21
Figure 6.7: Development of indicators to assess the contamination of natural resources .......... 6 22
Figure 6.8: Summary of many major hedgerow functions ........................... 6 28
Figure 6.9: Cycling of land between broad categories of agricultural and other uses ............ 6 30
Figure 7.1 : Issues to be addressed in the Logistics Plan prior to implementation of agroecosystem
monitoring ............................................... 7 1
Figure 9.1: Key components of the Quality Assurance Program Plan ..................... 9 2
Figure 9.2: Direct linkages between quality assurance and total quality management ............ 9 3
Figure 9.3: Overview of the essential components for a sucessful implementation of Total Quality
Management .............................................. 9 4
Figure 9.4: Sequences of inputs and outputs in which overall quality can be enhanced by Total Quality
Management .............................................. 9 5
Figure 9.5: Representation of trade-offs in the process of developing data quality objectives (DQOs). . . 97
Figure 9.6: The multi-stage iterative process for the development of Data Quality Objectives (DQO). . . 98
Figure 9.7: Continuous feedback and improvement process for the refinement of Data Quality
Objectives ............................................... 9 9
Figure 9.8: Levels of increasing complexity (bottom-to-top) within the Agroecosystem to which Quality
Objectives (QO) can be applied .................................. 9 10
Figure 9.9: Components of the data acquisition process amenable to quality control to measure and reduce
potential sources of errors .................................... 9 13
Vll
-------
Figure 9.10: An example control chart for monitoring the Quality Assurance and Quality Control
of data 9 15
Figure 10.1: Overview of the flow of data through the Agroecosystem Information Center 10 1
Figure 10.2: Flow of data collected by NASS to the Agroecosystem Information Center 10 3
Figure 10.3: Flow of data from other EMAP sources and other agencies and institutions to the
Agroecosystem Information Center and NASS data center for integration 10 4
Figure 10.4: Data flow showing the close interaction between Total Quality and Information Management
efforts. Quality control points are shown along the data flow path 10 5
Figure 10.5: Application of QA procedures to ensure that data are of the highest quality 10 6
Figure 11.1: The agroecosystem implementation schedule 11 10
Vlll
-------
List of Tables
Table 1.1: The Environmental Monitoring and Assessment Program (EMAP) is organized into resource
and integration groups 1 2
Table 1.2: Potential users of EMAP-Agroecosystem information 1 7
Table 3.1: Hexagon Plan Sampling Schedule 3 9
Table 3.2: Rotation Panel plan sampling schedule 3 11
Table 6.1: Initial grouping of agroecosystem indicators within each indicator category 6 4
Table 6.2: Association between the Agroecosystem assessment endpoints and indicators 6 7
Table 6.3: Critical and desirable criteria for selection of indicators for an ecosystem monitoring
and assessment program 6 9
Table 6.4: Ancillary data to be used in the development and interpretation of indicator values 6 11
Table 6.5: Candidate microbial indicators of soil productivity 6 15
Table 6.6: General site characterization and pesticide model parameters 6 23
Table 6.7: Landscape descriptors currently under consideration for use in monitoring agroecosystems. . 6 32
Table 8.1: Initial list of analytical methods for soil, water and biological samples 8 2
Table 9.1: Five types of survey errors addressed in the NASS QA program 9 11
Table 9.2: Description of the required subjectareas of a quality assurance project plan for EMAP .... 9 12
Table 9.3: Categories of audits used to determine status of QA in a monitoring program 9 14
Table 9.4: Phases of EMAP planning and implementation with products 9 18
Table 10.1: Summary of confidentiality provisions of several government agencies with data of
value to the Agroecosystem Resource Group 10 7
Table 11.1: Tasks with schedules for conducting the pilot projects NC Pilot Plans (1992-93) 11 2
Table 11.2: Planned implementation of Agroecosystem monitoring and assessment across
EPA regions 11 3
Table 11.3: Planned implementation of Agroecosystem monitoring and assessment across four U.S.
mega-regions 11 4
Table 11.4: Activities in 1991 to prepare for the Agroecosystem pilot in 1992 11 12
Table 11.5: Specific research plans for the Agroecosystem Resource Group in 1991 11 13
Table 11.6: Program Tasks and Budgets for 1991 11 15
Table 11.7: Program Tasks and Budgets for 1992 11 16
Table 11.8: Program Budgets by Location Category for 1991-1993, In Thousands 11 17
Table 11.9: Implementation Schedule and Budget Needs for 1991-1995 11 18
Table 11.10: Long-Term Strategy: Technical and Administrative Personnel Needs 11 19
Table 11.11: Expected Program Products 11 19
IX
-------
Glossary of Acronyms
ARG Agroecosystem Resource Group
ARS Agriculture Research Service (USDA)
BLM Bureau of Land Management (USDI)
CDF Cumulative distribution function
CRP Conservation Reserve Program
DQO Data quality objective
EMAP Environmental Monitoring and Assessment Program
EPA Environmental Protection Agency
EPIC Environmental Photographic Interpretation Center
ERS Economic Research Service (USDA)
FAO Food and Agriculture Organization of the United Nations
FSA Food Security Act
F&W Fish and Wildlife Service (USDI)
GIS Geographic information systems
HLCS Habitat Linear Classification System
HLI Habitat Layer Index
HPLC High pressure liquid chromatograph
ICP Inductively-coupled plasma spectroscopy
IMC Information Management Committee
JES June Enumerative Survey
LCD Landscape Characterization Database
MS Mass spectroscopy
NASS National Agricultural Statistics Service (USDA)
NRI National Resource Inventory (USDA/SCS)
ORD Office of Research and Development (EPA)
PSUs Primary Sampling Units
QA Quality assurance
QAC Quality Assurance Coordinator
QAO Quality assurance officer
QAMS Quality assurance and management staff
QAPP Quality assurance program plan
QC Quality Control
SCS Soil Conservation Service (USDA)
SOP Standard operating procedures
TD Technical Director (EMAP)
TQM Total quality management
USDA United States Department of Agriculture
USGS United States Geological Survey (USDC)
-------
Acknowledgements
The members of the Agroecosystem Resource Group have interacted with many people
during the development of this Plan. Dozens of people have shared information and expertise
during various Agroecosystem workshops. Many more have responded to our written inquiries
and telephone calls. The level of interest and cooperation from other government agencies,
universities and conservation organizations is appreciated. We would like to express our
gratitude to all who have contributed to our effort.
A special note of thanks is in order for Robert Bass and those who work with him at
USDA / NASS. They have become an integral part of the Agroecosystem Resource Group,
bringing with them years of experience in developing spatial sampling techniques for agricultural
statistics. Their insight has proven invaluable and helped us avoid many of the pitfalls associated
with a program as broad and diverse as EMAP. We would also like to thank Robert Tortora, our
first contact at USDA / NASS, who really helped get things started. (Robert has since transferred
to the USDC Agricultural Census Division.)
Special thanks go to Roy Cameron, the first Technical Director of EMAP-Agroecosystems,
who passed on many valuable ideas. He has continued to share his insights with our Resource
Group and has added greatly to the continued development of the Program.
Special thanks also go to Bruce Jones for his programmatic suggestions, ideas and support,
and to Ann Pitchford for her dedicated adminstrative support to the Program.
We would also like to thank the ex-officio members of the Agroecosystem Resource Group:
Karl Hermann, Doug Lewis and Susan Spruill. These individuals have provided valuable advice,
guidance and assistance during the development of the Program.
Finally, we would like to thank Ramona Logan and Sirena Hardy, both of whom spent long
hours at the computer keyboard revising and compiling this manuscript; and our librarian, Phyllis
Garris, who never left us without reading material.
XI
-------
1. Introduction
This document is a conceptual and implementation plan for monitoring the ecological
condition of agroecosystems in the United States. It represents the combined effort of the
Agroecosystem Resource Group (ARG). The plan is viewed by the ARG as a living
document that will serve as a basis for discussion of goals, objectives, concepts, and
approaches.
1.1. Overview of the Environmental Monitoring and Assessment Program (EMAP)
The agroecosystem monitoring program described in this document is one component
of the Environmental Monitoring and Assessment Program (EMAP), a national program
administered by the U.S. Environmental Protection Agency's (EPA) Office of Research and
Development (ORD). The EPA, U.S. Congress and private environmental organizations have
long recognized the need to improve our ability to document the condition of our
environment. In the past decade, environmental scientists have been calling for more relevant
and accessible ecological data and the EPA has been encouraged to take an ecological
perspective of the environment, in which the ecosystem is the fundamental unit of research
and monitoring. In 1988, the EPA Science Advisory Board (SAB) recommended that EPA
initiate a program to provide baselines estimates of the condition of U.S. ecological resources
against which future changes could be compared with statistical confidence. In response to
recommendations by the SAB, Congress and the public, EPA is designing EMAP in
cooperation with other agencies and organizations (Kutz and Linthurst 1990).
EMAP is organized into seven broad resource categories (Table 1.1). This was done
to facilitate interagency cooperation and the use of scientific expertise, which tend to be
grouped into these types of resources. Interdisciplinary groups of scientists, called 'Resource
Groups', are responsible for developing a plan for the collection, analysis and integration of
data from each of the seven ecological resources. In addition, seven cross-cutting
coordination groups have been established to assist the resource groups and to ensure total
quality management, consistency and integration across the program (Table 1.1).
The overall goal of EMAP is to quantify the extent, magnitude and location of
changes in environmental condition. EMAP is designed to meet the following objectives
when fully implemented (U.S.EPA 1990, Kutz and Linthurst 1990):
o Estimate the current distribution, status, and trends in indicators of ecological
resources on a regional basis with known confidence.
o Monitor indicators of pollutant exposure and habitat quality and seek
associations between anthropogenic stresses and ecological condition.
o Provide periodic statistical summaries and interpretive reports on ecological
condition to the public, the scientific community, and to policy-makers.
1 1
-------
Table 1.1. The Environmental Monitoring and Assessment Program (EMAP) is organized into resource and
integration groups.
Resource Groups
Integration Groups
agroecosystems
arid lands
estuaries
forests
the Great Lakes
inland surface waters
wetlands
air and deposition
design and statistics
field sampling logistics
indicator development and evaluation
information management
integration and assessment
landscape characterization
quality assurance and quality control
An assessment of the condition of U.S. ecological resources requires a consistent data
base over large geographic regions and long periods of time. Obtaining this database by
aggregating data from many studies that are spatially and temporally fragmented is
ineffective, if the results are to be extrapolated to whole populations. EMAP is being
designed to allow extrapolation to the entire resource population, and to provide unbiased
estimates of status and trends in ecological condition by collecting data within multiple
ecosystems over large geographic and temporal scales. This approach distinguishes it from
most current monitoring efforts (Kutz and Linthurst 1990).
1.2. Definition and Importance of Agroecosystems
For EMAP, agroecosystems are defined as land used for crops, pasture and livestock;
the adjacent uncultivated land that supports other vegetation (hedgerows, woodlots, etc) and
wildlife; and the associated atmosphere, underlying soils, ground water, and drainage
networks (first and second order streams, ponds, and irrigation drainage networks).
Figure 1.1 presents a simplified conceptual model of an agroecosystem. The inputs are
both natural and anthropogenic and the outputs are composed of both desirable and
undesirable quantities. Often agroecosystems are viewed primarily as agricultural production
systems. This view is not entirely incorrect, but it is incomplete. As shown in Fig. 1.1,
agricultural commodities, crops and livestock, are one of several agroecosystem components.
The conceptual model and the definition of agroecosystems illustrate how the
EMAP-Agroecosystem program is being designed as a holistic approach that considers all
constituent components of agroecosystems. The challenge of an agroecosystem monitoring
and assessment program is to provide information on the agricultural productivity of the
system as well as with information on the ecological well-being of that system.
1 2
-------
INPUTS
OUTPUTS
Management Practices:
Tillage
Chemicals
Irrigation
Natural Environment:
Precipitation
Temperature
Humidity
Soil Processes
Pests
Beneficials
Crops and Livestock
Non-crop Vegetation
Abiotic Resources
(Soil, Water, Air)
Harvests:
AGROECOSYSTEM
Non-point Source
Loading:
Agri-chemicals
Sediments
Salts
Methane
Animal Wastes
Figure 1.1: Conceptual model of an agroecosystem.
Agroecosystems have perhaps the greatest impact on our daily lives of any of the
terrestrial ecosystems, because they provide us with food and fiber and have a large influence
on the quality of our environment. Farmers have the stewardship of .more of the global
environment than any other group (Paarlberg, 1980). Globally, agriculture accounts for nearly
20% of terrestrial net primary productivity and approximately 30% of the land area (Coleman
and Hendrix, 1988). In the United States alone, crop land accounts for approximately 443
million acres, nearly 20%, of the total U.S. land area (U.S.DOC, 1990). The extent of land
devoted to specific crops in the U.S. varies among states and regions (cf. Fig. 1.2 a,b). Total
farm land in the U.S. comprises nearly 386 million hectares which is about 43% of the total
U.S. land area (U.S.DOC, 1990). This means that there are 4-5 acres of farm land for every
man, woman and child in the nation.
1 3
-------
IMPORTANT CORN PRODUCING REGIONS
MOST IMPORTANT
LESS IMPORTANT
D
2 lo 23* «/ I*** *r«
ID corn
COLtTCTU TXTTt »HL*T HIT
*•• n«w «
IMPORTANT WHEAT PRODUCING REGIONS
S? MOST IMPORTANT
•
15 U 47t X IU< „,.
la *b*«l
LESS IMPORTANT
a
Figure 1.2: Distribution of important crop producing areas for two major commodities a) corn and b) wheat
(David T. Tingey, personal communication).
1 4
-------
Agriculture is a major component of the U.S. economy. The production, processing
and sale of food and fiber account directly for approximately 17.5% of the gross national
product or approximately $700 billion in economic activity annually (NRC, 1989). The
indirect and socioeconomic effects of agriculture are equally significant when the impact of
farms and farm-related businesses on the well-being of rural and urban communities is
considered.
Although agricultural systems have been viewed as relatively simple, even annual
monocultures such as wheat or soybean are far more complex than they initially appear
(Paul and Robertson, 1989). The periodic and chronic disturbances that are an inherent part
of agricultural management (e.g., plowing, application of agricultural chemicals, and
harvesting), place agroecosysterns among the most rapidly changing landscapes on earth.
Agricultural landscapes are disturbed by purposeful human activity which significantly
alters the original character of the landscapes. The disturbances, while economically
advantageous in the short term, can result in ecological problems. These problems include soil
erosion, dependency on fossil-fuels, contamination of natural resources with agrichemicals,
and narrowing of the gene pool of our major crops. For example,
o Soil erosion from U.S. crop land is estimated to be between 2.7 and 3.1 billion tons
annually with accompanying off-site damage between $2 and $8 billion each year
(NRC, 1989).
o More than 50% of suspended sediments discharged into surface waters nationally is
from agricultural land (NRC, 1989).
o The more than 20 million irrigated hectares in the U.S. use 83% of the total water
consumed (USDA 1985).
o Forty percent of irrigation water comes from groundwater (Pimentel 1982), and the
amount of water extracted exceeds replenishment (OTA 1983, Stokes 1983).
o Less than 50% of the water applied for irrigation is used by plants; the rest eventually
returns to rivers and streams (Postel 1985). The quality of the return flow is often
degraded by salts, sediments, fertilizers and pesticides (Soule 1990).
o Less than 0.1% of pesticides applied to crops reaches target pests; thus, over 99% of
the applied pesticides are free for multi-media transport in the ecosystem (Pimentel
and Levitan, 1986).
o Normal use of agricultural pesticides has resulted in residues of 46 different pesticides
in groundwater in 26 states (Weinberg, 1990).
1 5
-------
o Long-term declines in wildlife populations, such as the savannah sparrow, bobolink,
dicksissel and grasshopper sparrow are thought to have occurred in intensely cultivated
areas of the Midwest (Jahn 1988).
o Eighty-seven percent of wetland loss in the U.S. from 1950 to 1970 was to agricultural
use (Tiner 1984).
o As of 1980, only six cultivars made up 43% of the hybrid corn, 42% of soybean and
38% of U.S. wheat crops (Duvick 1984).
Increased awareness of ecological threats to agroecosystems and surrounding areas
have resulted in water quality initiatives (Burkart et al 1990, Wilson 1987), emphasis on input
management (Cox 1984, Odum 1989), and the design and implementation of best
management practices (Humanik et al. 1984, Pimentel et al. 1989).
The challenge to modern agriculture is to produce affordable food and fiber in an
economically viable manner, while preserving the short- and long-term integrity of the local,
regional and global environment. In a healthy agroecosystem, a balance exists between
sustainable crop and livestock production; maintenance of air, soil and water quality; and
assurance of diversity of wildlife and vegetation in the noncrop-habitats. The degradation of
any one component influences the other components in the agroecosystem and in adjacent,
linked ecosystems. As Elliott and Cole stated (1989): "The challenge, then, becomes, can we
sustain our agriculture in the face of shrinking energy supplies and increasing environmental
pollution and still maintain a profit for farmers without increasing cost to consumers too
much?"
1.3. Overview of the EMAP-Agroecosystem Program
1.3.1. Potential users of EMAP-Agroecosystem information
The agroecosystem component of EMAP is intended to meet a growing demand for
information from a wide spectrum of users. The ARG has developed an initial list of
potential users (Table 1.2) and is working on establishing communication with them to 1)
help identify and formulate essential questions to be answered by EMAP; 2) determine the
appropriate data needed; and 3) communicate to potential users the extent and limitations of
the monitoring program (NRC 1990).
1 6
-------
Table 1.2. Potential users of EMAP-Agroecosystem information.
Public
General public
Conservation organizations (e.g. The Nature Conservancy, Audubon Society)
Farmers and other members of the agricultural community
Resources for the future
Government
Cooperative Extension Service (federal, state and county, using regional information)
USDA/Soil Conservation Service (SCS) (federal, state and county, using regional information)
USDA/SCS Resource Conservation Act Appraisals and National Resource Inventory
USDA/Agricultural Research Service
Specific USDA programs: Low-Input Sustainable Agriculture (LISA); Global Change; Food Safety; Water
Quality Environmental legislation in the Farm Bills of 1985 (Food Security Act) and 1990 (Food,
Agriculture, Conservation and Trade Act): Federal Conservation and Cropland Reduction Programs,
including the Conservation Reserve Program, Swampbuster, Sodbuster and Acreage Reduction
Program
U.S. EPA: Office of Policy and Planning, Office of Research and Development, Office of Pesticide
Programs, Regional offices
U.S. Fish and Wildlife Service
State Environmental Departments (using regional information; may eventually cooperate on state-level data)
Congress; Congressional staff members
Council on Environmental Quality
Scientists
Agricultural Extension research personnel
University and industry scientists and researchers
1.3.2. Identification of relevant legislative mandates
With the passage of the 1985 Farm Bill, titled the Food Security Act (FSA) of 1985,
the federal government took a major step forward in placing concerns about resources and
the environment in the mainstream of U.S. agricultural policy (The Conservation Foundation
1986). As a result of the programs enacted by the FSA, collectively called Federal
Conservation and Cropland Reduction Programs, there will be major shifts in land use within
agroecosystems, with as yet unknown ecological implications. The 1990 Farm Bill, titled the
Food, Agriculture, Conservation and Trade Act of 1990, expanded and increased these
programs.
The EMAP-Agroecosystem program is in a unique position to provide data and
assessment of the impact of national agricultural policy programs. For example, the goal of
the conservation programs enacted in 1985 is to conserve soil by reducing crop production on
soils classified as highly erodible by wind and water. Data on soil erosion collected by the
Soil Conservation Service in the National Resource Inventor}' program can be used to assess
1 7
-------
the status and trends in the rate of soil erosion after the program was established. However,
the Cropland Reduction Programs have other important ecological goals for which no
monitoring effort currently exists (CAST 1990). These goals are to:
o protect long-term ability to produce food and fiber
o reduce sedimentation
o improve water quality
o create better habitats for fish and wildlife, and
o enhance the diversity of the ecosystem
The Conservation Reserve Program (CRP) is the largest of the long-term acreage
reduction programs, with a goal of enrolling 40 to 45 million acres (over 10% of current U.S.
cropland). The CRP is a voluntary program that places qualifying land into permanent,
soil-conserving covers such as grass and trees for a ten-year contract period. The
"swampbuster" and "sodbuster" provisions of the FSA are designed to discourage conversion
of naturally occurring wetlands or highly erodible land, respectively, to agricultural
production. The result of these programs will be an increase in the amount of rural land used
as low management habitats that can be exploited by insects, weeds, plant pathogens, and
wildlife. The program is already starting to provide benefits for many kinds of wildlife,
particularly ground-nesting birds (CAST 1990). Agricultural advantages could occur if
beneficial organisms are encouraged and reliance on pesticides is reduced. However, there is
concern that set-aside land could become islands of infestation, increasing the exposure of
surrounding cropland to insects, weeds, and pathogens (CAST 1990). Agroecological data
will be needed to effectively evaluate and guide the process of refining these national
agricultural policies.
1.3.3. Specific goals and objectives of the Agroecosystem program
The mission for the Agroecosystem Resource Group is to develop and implement a
program to monitor and evaluate the long-term status and trend of the nation's agricultural
resources from an ecological perspective through an integrated, interagency process.
The objectives of the Agroecosystem program parallel the overall EMAP program
objectives, but focus more specifically on agroecological resources. When fully implemented,
the program will meet the following objectives:
o Estimate the distribution of agroecosystems and the status and trends in indicators of
ecological condition on a regional basis with known confidence.
1 8
-------
Monitor indicators of pollutant exposure and habitat quality and seek associations
between anthropogenic stresses and ecological condition.
Provide periodic statistical summaries and interpretive reports on ecological condition
to the public, the scientific community, and to policy-makers.
1.3.4. Assessment Endpoints
Unlike indicators of human or animal health, there are no simple, integrative measures
that indicate the relative health of an ecosystem (Rapport 1989, Schaffer et al 1988).
Therefore, assessment endpoints that encompass the concept of agroecosystem health will be
used to focus the interpretation of indicator data. This is an important first step in the design
of a monitoring program (Fig 1.3). Assessment endpoints are a quantitative or quantifiable
expression of the environmental value, such as agroecosystem health, to be monitored and
assessed. Good assessment endpoints have social and biological relevance, an unambiguous
operational definition, are accessible to prediction and measurement, and are susceptible to the
environmental stressors of concern (Suter 1990). Assessment endpoints are long-term societal
values that will not change over time, even when specific stressor or specific issues do
change.
After careful consideration of important scientific, social, economic and environmental
issues concerning agriculture, agroecosystems, and their associated surroundings, three
assessment endpoints were identified that summarize the essence of the issues. These are:
o sustainability of commodity production
o contamination of natural resources
o quality of agricultural landscapes
Although members of the Agroecosystem Resource Group agree on the basic issues
addressed, they are still debating the terminology and organization of these endpoints. The
three endpoints are defined below and are used to organize the discussion of indicators in
Chapter 6.
1.3.4.1. Sustainability of Commodity Production
Sustainability refers to the capacity of a particular agroecosystem to maintain a level
of crop or livestock productivity that provides for basic human food and fiber needs, and an
economically viable livelihood for farmers without polluting or seriously depleting soil, water,
wildlife, fossil fuel or other resources. The temporal scale for sustainability is decades or
centuries, not a single growing season. The continual removal of biomass from a field
1 9
-------
Focus Scientific
Understanding and
Define Mission
Identify Relevant
Laws and Regulations
that Provide Direction
Identify Public
Concerns and
Ideological Issues
Identify what Information
is Currently Available to
Answer Questions
Identify Gaps in Information
Establish Goals and Objectives
to Cover Gaps and Answer Questions
(Adapted trom NRC 1990)
Figure 1.3: Process used Lo define expectations, goals, and objectives of the monitoring program.
necessitates inputs or adjustments to maintain or restore productivity. Long-term
sustainability of agroecosystems can be masked in the short-term by management practices.
For example, land management that results in erosion, salinization, or desertification will
eventually degrade soil quality and the sustainability of the system. Inputs, outputs,
socioeconomic factors, and the use of natural resources will be considered in the assessment
of agroecosystem sustainability.
/.3.4.2. Contamination of natural resources
Contamination is defined as anthropogenically-related stressors that have direct or
indirect effects on the sustainability, productivity, structure or function of the agroecosystem.
Contaminants can be found in the air, soil, water, and biota of agroecosystems, and may
include air pollutants, agricultural chemicals, animal and municipal wastes, water pollutants,
and genetically-altered organisms.
Contaminants can also be transported from agroecosystems. On a regional and
national scale, managed agricultural systems contribute to nonpoint source pollution through
loss of agricultural chemicals and sediments carried to streams and rivers, and from the
10
-------
atmospheric transport of pesticides and other volatile chemicals. On a local level, managed
agricultural systems can be pollution point sources, such as pesticide drift from aerial
spraying, that can impact immediately adjacent areas.
1.3.4.3. Quality of agricultural landscapes
The quality of agricultural landscapes refers to the various ways in which the
landscape matrix is modified or employed for agricultural and non-agricultural purposes over
time. Agricultural land use patterns exert a major influence on ecological processes. For
example, landscape heterogeneity may affect soil erosion, water quality, crop-pest interactions,
ecological diversity and the spread of diseases (Barrett et al. 1990). A vital characteristic of
landscape modification is the extent to which the surrounding landscape can support
populations of non-crop vegetation and wildlife. An assessment of agroecosystem health and
the development of sustainable agricultural systems must consider landscape level processes
over time and the coupling of natural and agricultural components of the landscape (Barrett et
al. 1990, Risser 1985, Risser et al. 1983).
1 11
-------
2. Program Approach and Rationale
This section provides an overview of the EMAP-Agroecosystem approach and
highlights those attributes that make it unique. An overview of the technical approach for
meeting the mission, goals and objectives outlined for the program in Chapter 1 is included.
2.1. Overview
The fundamental differences between the Agroecosystem program and existing
monitoring efforts occur in the areas of design and indicator strategy. The program is
designed to evaluate the health or condition of agroecosystems with respect to the assessment
endpoints of concern. Our discussion of these differences is not intended as a criticism of
existing programs. Rather, it is an acknowledgement that existing programs were designed
with different objectives in mind and, as such, cannot alone be effectively used to meet the
broad regional and national-scale, ecologically-oriented objectives of the Agroecosystem
component of EMAP.
2.1.1. Enhancement of Existing Efforts
The strategy of EMAP will be to use the following features to estimate, with known
confidence, the regional condition of agroecosystem populations:
o Explicit definition of target populations within agroecosystems and of sampling
units through the use of the NASS (National Agricultural Statistics Service)
area frame (Chapter 3).
o Probability sample site selection from the area frame. A stratified, random
sample of agricultural and associated ecological resources will be obtained
(Chapter 3).
o Representation of agricultural and ecological conditions in sample fields and
associated areas using ecological indicators and derived indices of system
health (Chapters 4 and 6).
o Uniform sampling and analytical methods for a suite of indicator measurements
(Chapters 5, 6, 8).
o A documented program of rigorous quality assurance and quality control
(Chapter 9).
o An information management system that will allow the acquisition, processing,
storage and analysis of data acquired from diverse sources, and the
communication of that information to researchers, administrators and
policymakers (Chapter 10).
2 1
-------
These features, coupled with an understanding of the ecological and agricultural
function of agroecosystems, and information from existing databases, will provide an
effective, integrated program capable of providing essential and increasingly valuable
information about the condition, status, extent and trends of U.S. agroecosystems.
The Agroecosystem component of EMAP is being designed as a comprehensive
agroecological monitoring program to increase our knowledge of the status and extent of our
national agroecosystem resources. It will enhance existing agricultural monitoring efforts
such as those carried out by USDA/National Agricultural Statistics Service (NASS),
USDA/Soil Conservation Service (SCS)-National Resources Inventory (NRI),
USDA/Economics Research Service (ERS), and the USDC Bureau of the Census by adding
an essential ecological dimension to current data collection, compilation and interpretation.
Because of the focus of the overall EMAP program, and the size of the United States, the
Agroecosystem program will not have the resources to make estimates for areas of the United
States smaller than EPA regions (Figure 2.1) until cooperative efforts with other federal and
state agencies allow for a sampling intensity that is sufficient to provide greater spatial
resolution.
Figure 2.1: U.S. Environmental Protection Agency Regions.
When implemented, the program will provide unique data on trends in the condition of
crop and non-crop resources which will be interpretable from environmental, ecological,
agricultural or agroecological viewpoints. The interpretations made from the integrated data
2 2
-------
on indicators of agroecosystem health will provide a more realistic and complete assessment
of resource conditions than is possible with current, agriculturally oriented data alone.
2.1.2. The Tiered Approach in EMAP
An understanding of the tiered approach to meeting the program objectives is essential
for understanding the overall EMAP approach. A tier is a level and type of activity related to
monitoring and assessing ecological condition. Four tiers comprise the general approach in
EMAP (Figure 2.2). Although all four tiers are important for an holistic understanding of the
overall ecology of agricultural systems, this document focuses primarily on activities at Tier 2
with some discussion of Tier 1 activities.
Increasing
Spatial
Coverage
Estimates of Condition
and Trends
Landscape Characterization
Estimates of Extent and Land Use
Tierl
Figure 2.2: Concept of a four-tiered approach in EMAP. Spatial coverage is maximized
in lower tiers while temporal coverage increases at higher tiers.
Tier 1 is the broadest level. Landscape characterization and estimates of extent and use
of land are performed at this level. The primary purpose of landscape characterization is to
identify the location, spatial extent, and amount of resource present on both a regional and
national basis. Most of the Tier 1 information utilized by the Agroecosystem program will be
derived from information provided by the Landscape Characterization Group of EMAP and
by NASS. Estimates will be updated periodically from new aerial photography and satellite
imagery.
2 3
-------
Tier 2 activities involve field sampling to provide estimates of the condition of
resources and to measure status and trends. A subset of Tier 1 sites will be sampled using a
probability sampling frame (Chapter 3). Environmental measurements will be made at each
of the Tier 2 sites using a suite of indicators (biological, chemical and physical; Chapter 6).
Information derived will be assembled into databases and analyzed in order to make
assessments at the regional scale. Statements about any specific sampling site will not be
made and are not appropriate.
Tier 3 studies will provide higher spatial resolution with detailed data on the sites
selected. Tier 4 studies will involve process-level research designed to complement EMAP
monitoring in Tiers 2 and 3. Activities at Tier 3 and 4 will be implemented in the future, as
resources become available. The specific objectives and approaches for Tiers 3 and 4 have
not been finalized. Activities at Tiers 3 and 4, however, will undoubtedly involve a diversity
of studies designed to supplement and complement information derived in Tier 2 studies. It is
anticipated that more intensive sampling will be conducted at fewer sites in Tiers 3 and 4,
and may involve the testing of hypotheses generated from Tier 1 and 2 data.
The majority of this document focuses on Tier 2 activities. Planning for Tier 3 and 4
activities will proceed as needed to complement the Tier 2 program. An initial priority for
Tier 3 and 4 activities may be in the area of nonpoint source loadings from agricultural lands.
2.2. Overview of Monitoring Design
The EMAP statistical design was conceived as a probability sampling frame to sample
ecosystem resources, and provide, with known confidence, statistical estimates of conditions
and trends in these resources. Two approaches are currently under consideration for selection
of sample sites. Both approaches make use of the area sampling frame of NASS.
One approach is based on a permanent national sampling framework consisting of a
hexagonal plate containing a systematic triangular point grid of approximately 12,600 points
placed randomly over the conterminous United States; a similar array is available for Alaska
and Hawaii. A defined area (40 km2 hexagon) around each point will serve as the initial
sampling domain. In this approach, the centroid of the EMAP hexagon will define the NASS
primary sampling unit (PSU) and segments to be used as the actual sampling sites. Another
approach is to use currently active NASS PSU's to identify the sampling sites. The details of
these approaches and the relative advantages of each are discussed in detail in Chapter 3.
2.3. Use of Existing Data
The ARG recognizes that a number of existing data bases can be of use to partially
fulfill our basic objectives. We have made a partial list of these databases (Appendix 3) and
have initiated efforts to contact key individuals responsible for these databases.
2 4
-------
The purpose of the database collection effort is to determine what data are available
and useful to the Agroecosystem component of EMAP. Knowledge of other databases will
help minimize duplication of effort, and identify data that could supplement agroecosystem
sampling with existing ancillary or corroborator}' data. Another aim is to develop a network
of contacts in private and government agencies who are willing to participate and assist in the
evaluation and use of data obtained from those agencies so the data are supportive of EMAP
objectives.
The identification and use of existing databases will continue for the duration of the
EMAP project. Therefore. Appendix 3 will be updated continuously as additional information
becomes available.
Whereas excellent information on agricultural productivity in agroecosystems is
available, information related to sustainability of non-crop resources (wildlife habitat,
biodiversity, etc.) and to the extent of contamination of natural resources on a national scale
is largely lacking. For example, several databases have been developed by the U.S. EPA,
USDA-ARS. and cooperating state agencies relative to chemical transport and fate at a limited
number of selected sites located in the east and midwest. These databases will be available
for evaluation by the ARG. The existing databases were designed primarily for model
development and testing of runoff and leaching models for the root zone, unsaturated and
saturated zones.
The primary emphasis in the development of many extant databases has been on
crop and livestock productivity, extent of production and the socioeconomic effects of
increases and decreases in agricultural productivity. Such databases are vital sources of
information for EMAP. but often do not fully address questions concerning the ecological
status of agroecosystem resources.
2.4. Indicators of Agroecosystem Condition
Ecosystem health is defined in the literature as both the occurrence of certain
indicators that are deemed to be present in a healthy, sustainable resource, and the absence of
known stressors or problems affecting the resource. A healthy agroecosystem is one that
balances crop and livestock productivity with the maintenance of air. soil and water integrity
and assures the diversity of vdldlife and vegetation in the noncrop habitats.
2,4.1. Assessment Endpoints
Determination of system health is based upon a series of assessment endpoints. These
endpoints are quantitative or quantifiable expressions of the actual environmental value that is
to be protected. The}- should have unambiguous operational definitions, have social or
biological relevance and be amenable to prediction or measurement (Knapp et al. 19901
Assessment endpoints. as indicated in Section 1.3.4. must relate to the environmental values
-------
identified for each resource category in EMAP. For agroecosystems, the assessment endpoints
of concern are:
o Sustainability of commodity production
o Contamination of natural resources (air, water, soil and biota)
o Quality of agricultural landscapes
2.4.2. Indicator Approach
A suite of indicators that are related to the assessment endpoints and collectively
describe the ecological health of agroecosystems will be monitored. The term indicator refers
to measurable attributes of the environment that can be monitored via field sampling, remote
sensing or compilation of existing data. EMAP proposes the use of several types of
indicators, including indicators of ecological condition, pollutant exposure, habitat quality, and
stresses impacting ecological condition (Fig. 2.3, Chapter 6).
Biotic Indicators
of Condition
t
Biological
Communities
(processes and
interactions altered
due to exposure to
modified habitats)
Policy Directed
Toward Impacting
Alteration of:
Chemical Habitat
Physical Habitat
Biological Habitat
Figure 2.3: EMAP provides a foundation fitting into the ORD's Ecological Risk
Assessment Process. EMAP is most concerned with biotic indicators of condition and the
endpoints of concern at Tier 2.
2 6
-------
Selection and development of indicators to address the assessment endpoints for the
Agroecosystem program must be considered a long-term process requiring extensive testing
and evaluation of each indicator prior to its full-scale implementation. A multi-phase process
has been identified (Knapp et al. 1990) to guide the current selection of indicators and future
development of new indicators (Fig. 6.3). It involves identification of issues, the
development, screening and qualitative evaluation of indicators, and the implementation of a
core set of indicators.
2.4.3. Indicators and Exposure Assessment Models
A variety of predictive exposure assessment models are available for use in assessing
the overall condition of ecosystems. For example, several models could provide the required
data outputs for items such as non-point source loadings integrated over the temporal and
spatial scales appropriate for interpreting data from other agroecosystem indicators and those
from other linked, adjacent ecosystems. Examples of such models include the Pesticide Root
Zone Model (PRZM) (Carsel et al. 1984), the Risk of Unsaturated-Saturated Transport and
Transformation Interactions for Chemical Concentrations Model (RUSTIC), developed by the
EPA (Dean et al. 1989), and the Groundwater Loading Effects of Agricultural Management
Systems Model (GLEAMS) and Agricultural Nonpoint Source (AGNPS) models developed by
the USDA-ARS (Leonard et al. 1987, Young et al. 1989). Information derived from
indicators such as soil characteristics, agrichemical usage and crop management practices as
well as ancillary data, such as weather records, are important model input parameters. A list
of general model input parameters for exposure assessment models that may be useful to the
Agroecosystem program is given in Table 6.6.
2.5. Overview of Implementation Plans
A pilot study is planned to evaluate the sampling design and protocols in the field
prior to full-scale implementation. The initial pilot study will involve a one state area and
will utilize five of the proposed indicators discussed in section 6.5. All facets of the overall
program developed in 1990 and 1991 will be field tested in 1992. The .second stage (1993)
will utilize results from the 1992 field pilot in a regional demonstration to evaluate all
program elements before implementation of a full regional Agroecosystem Monitoring
Program in 1994. Full national implementation could occur as early as 1995 if sufficient
funds are available. (Chapter 11).
It is essential that this multi-stage implementation program be initiated to assure the
development of a strong national monitoring program. The preliminary, critical evaluation of
important issues related to the overall program will provide information essential to the
formulation of an effective and efficient implementation plan. The progression through
planning, pilot project, demonstration project and regional implementation allows for the
orderly establishment of protocols, and for the full utilization of existing data bases and
networks.
2 7
-------
2.6. Cooperation with other agencies
The ARG has held extensive discussions with personnel of USDA/NASS both at the
operational and administrative levels. Two members of the USDA/NASS staff serve as
regular members of the ARG and as liaison between NASS and the Agroecosystem
component of EMAP.
The cooperation with NASS benefits the Agroecosystem program directly in several
ways. First, NASS has an established and well-accepted frame for sampling nationwide.
Second, NASS has been performing site visits with farmers for many years. This history of
performance has permitted NASS to develop a network of enumerators and administrators
experienced in conducting successful national-scale surveys and monitoring activities.
Growers throughout the U.S. are familiar with and have confidence in NASS personnel.
Third, a nation-wide force of trained enumerators is in place with a proven administrative
organization. Fourth, the requirements for confidentiality of data for NASS are established by
law and are well known in the agricultural community. This confidentiality requirement is
essential to the success of the program in working with growers in the U.S.
The USDA Soil Conservation Service (SCS) performs the National Resources
Inventory (NRI), a comprehensive inventory of the non-federal natural resources in the United
States. Much of the data they collect may be of use for the development of indicators by the
ARG. The Inventory has been performed every 5 years, since 1977. However, only data from
the 1982 and 1987 inventories may be directly compared. Data from 1987 have only recently
been made available and the ARG is exploring the utility of these data for use in EMAP.
The USDA Economics Research Service is also involved in the determination of
trends and status of components of agroecosystems. The ARG is exploring the possibility of
cooperative efforts with USDA/ERS.
Cooperation with other federal, regional and state agencies will be initiated as the need
arises. The ARG views such cooperation as both desirable and essential to the success of the
program. Such cooperation assures that maximum benefit is derived from existing programs
and that the concept of EMAP as a valued-added program can have the widest possible
impact.
2.7. Ecosystem Linkages
Agroecosystems consist of diverse landscapes that interact with each of the other
component EMAP ecosystems (e.g. forests, surface waters, wetlands). For this reason, an
effective information exchange and monitoring linkage with other resource groups is essential.
Linkages with other groups, both in terms of defining indicators that reach across ecosystem
boundaries and recognizing interface areas, are important. A detailed consideration of
ecosystem linkage and associations for EMAP is presented in Appendix A9 and only an
overview is presented here.
2 8
-------
One approach to the linkage issue is for each ecosystem Resource Group to identify
the linkages with other ecosystems that are key to understanding the health of their system.
Once these linkages are identified, approaches can be suggested for studying the linkages. If
one ecosystem requires limited, but important, indicator data from another ecosystem, the first
ecosystem could adopt the indicator used by the other as one of their own and obtain the data
on that indicator from the appropriate area within or adjacent to their sample sites. The
protocols utilized for sample collection, analysis, etc. would be the same for each ecosystem
Resource Group. The benefit of this approach is that the needed information is obtained on a
site specific basis, but is compatible in quality and format between ecosystems.
If a regional problem were identified that appeared to involve several ecosystems, a
Tier 2 intensive sampling protocol could be designed to sample areas where the ecosystems
are co-located. The number of sampling units would have to be determined for the area of
interest and the sampling protocol coordinated between the ecosystems. Statistical
correlations could be determined and discussion of cause-effect relationships could be
developed. However cause-effect relationships could not be proven with these data. We have
tentatively defined the co-location of the intensive sampling as a Tier 3 diagnostic approach.
For those linkage studies that require more intensive data collection and a higher level
of coordination among ecosystems, a selected suite of Tier 4 research studies aimed at
understanding specific system linkages could be initiated. The study design in these cases
would be to select sample sites and collect data across ecosystems as necessary to understand
the system. Research sites could be new or existing locations.
Plans for implementing cross-ecosystem monitoring have not been finalized in EMAP.
Linkages of ecosystems as they relate, for example, to chemical/sediment transport across
EMAP ecosystem boundaries is an important and legitimate monitoring activity. Monitoring
transport of chemicals and eroded sediment loads from and within the agroecosystem is
essential to assess the overall ecological condition of the agroecosystem as well as to provide
input loading data to other EMAP ecosystem components. Because of the intensive, high cost
nature of these measurements, it is not feasible to attempt direct experimental coverage of all
of the important agroecosystem areas in the nation. A reasonable alternative may be to select
a few test sites in representative high-intensity agricultural areas where existing data sources
confirm the existence of environmental insult. As a Tier 4 activity, semi-permanent field
installations for measuring weather variables, soil physical properties, runoff and leaching to
groundwater under typical cropping scenarios could then be established.
2.8. Relationship to Ecological Risks and Decision Making
The database developed from the program will provide essential data for conducting
ecological risk assessment for use in making environmental management decisions. Risk
assessment is a scientific process of combining data and assumptions for use in estimating
potential for adverse effects on the environment resulting from pollutant exposure and other
ecological stresses. Ecological risk assessment procedures estimate the effects of
2 9
-------
anthropogenic activities on ecological resources and allow the significance of those effects to
be interpreted with quantified uncertainty estimates. The ideal output of the risk assessment
process is an estimated probability that an event of a certain magnitude will occur (e.g. a 90%
probability of decline in commodity production in 10% of the agricultural lands in a region).
The key components of risk assessment include (Hunsaker and Carpenter 1990):
o selection of assessment endpoints,
o qualitative and quantitative descriptions of the source of the hazard (e.g.
location and emission source for pollutants),
o identification and description of the reference environment within which effects
occur or are expected to occur,
o estimation or measurement of spatial and temporal patterns of exposure, and
o quantification of the relationship between exposure in the modified
environment and effects on biota.
The primary role of the Agroecosystem component of EMAP in the overall risk assessment
process is to identify and quantify agroecological hazards, related to endpoints of concern,
which can then be used in risk communication (Figure 2.4). This identification of hazards
will occur through the use of indicators of ecosystem condition (Chapter 6). The risk
communication will allow information on ecological effects, costs, benefits, alternatives and
trade-offs to be communicated effectively to the public, to Congress, and to agency decision
makers in EPA and USDA, for example.
2.9. Expected Outputs
The ARG will produce at least four types of informational outputs (Figure 2.5). For
the information collected to be of maximum utility to decision makers and the scientific
community, reports must be made available as quickly as possible. Annual statistical
summaries and maps will be published within nine months following the collection of the last
sample information for each year. Interpretive reports will be produced at irregular intervals,
but not less often than once every three years. The interpretive reports will integrate
information on the regional and national status of agroecosystems and will identify possible
causes of changes in system condition. Emerging problems and their possible causes will
also be highlighted. Scientific reports and peer-reviewed papers will be produced periodically
as the need arises and as the scientific data warrant. An illustrative statistical summary was
prepared as a planning exercise in 1990 (Meyer et al. 1990).
2 10
-------
Ecological
Risk:
Managemeot
Ecosystem
Restoration &
Management
" / Risk Characterization
Ecological
Risk
Assessment
Ecological Effects
Ecological Exposure
Environmental Monitoring & Assessment Program
Figure 2.4: EMAP provides a foundation for the ORD's Ecological Risk Assessment
Program. Heavy lines show principal interactions of EMAP.
OUTPUT EXAMPLES
• statistical summaries
interpretive reports
• research papers
maps
Figure 2.5: Expected outputs from the Agroecosystem Program.
2 11
-------
3. Design and Statistical Considerations
3.1. Monitoring Network Design
The establishment of a monitoring network for the EMAP-Agroecosystem program
requires a thorough consideration of possible design options. The monitoring network must:
1) have a valid and acceptable statistical design; 2) utilize, wherever possible, data from
existing monitoring programs; 3) allow for the collection of essential information in an
efficient and cost-effective manner; and 4) provide regional and national estimation with
acceptable precision. The envisioned network will be coordinated and administrated by the
Agroecosystem Resource Group (ARG), but will be an interagency and interdisciplinary
effort. A stronger and more efficient monitoring network can be implemented through such
cooperation than would be possible for EMAP acting alone.
3.2. General Statistical Sample Survey Requirements
A statistically valid sampling design must have the following properties:
o A frame covering the complete universe of interest must be available or can be
developed. Ordinarily, in agricultural surveys, an area frame (or in some
instances, a multiple frame) is used to meet this requirement.
o A procedure must be established for dividing the frame into identifiable
sampling units such that no part of the frame is omitted and none is included
more than once.
o Ancillary information must be evaluated and incorporated into the design to
increase the precision of the estimates.
o The number of sampling units required to achieve a specified level of precision
at minimum cost must be determined.
o Each sampling unit must be selected with a known probability, referred to as
the inclusion probability.
o A procedure must be available for expanding the sample into estimates of the
parameters of interest over the domains of interest.
o The design must permit the assessment of its precision from the sample results.
The EMAP design and the NASS (National Agricultural Statistics Service) design,
both of which are area samples, meet the above criteria.
3 1
-------
3.3. Overview of the EMAP Sampling Design
Under the EMAP sampling design, the resources of the United States would be
sampled via a triangular point grid. The grid defines approximately 12,600 hexagons, each of
which have an area of approximately 650 square kilometers. The hexagons cover the entire
conterminous United States and, thus, the area frame is complete (Fig. 3.1). The points also
define the centroid of a smaller hexagon of approximately 40.6 square kilometers in size;
each large hexagon contains exactly 16 small hexagons. The small hexagon whose centroid
coincides with the large hexagon becomes the sample hexagon. The hexagons are
randomized by selecting a random point in space and moving the entire pattern so that the
centroid of a given hexagon is identical with the random point. The small hexagons, of
course, move accordingly and are now a probability sample with a sampling fraction of 1 in
16. The random point has been selected and the location of each of the 12,600 sample
hexagons has been identified and located on maps. There is no stratification at this stage.
The initial approach for the EMAP Landscape Characterization (LC) program was to
characterize each of the 12,600 sample hexagons. Characterization was to be carried out by
the Environmental Photographic Interpretation Center (EPIC). However, the EMAP-LC
program is currently under revision following a peer review in July 1990. Both satellite
imagery and aerial photography were to be used in the effort at an estimated cost of
$56,000,000. The sample of 12,600 small hexagons and their characterization was to
comprise EMAP's Tier 1 sample.
3.4. Overview of the NASS Sampling Design
3.4.1. The NASS Area Frame and Sample
The National Agricultural Statistics Service (NASS) has used area frame probability
sampling designs to conduct agricultural surveys since 1954. The purpose of these surveys is
to gather information on crop acreage, cost of production, farm expenditures, grain yield and
production, livestock inventories and other agricultural items. The area frame sample survey
started in 1954 with ten states, 100 counties and 703 primary sampling units (PSUs). The
current area frame completely covers the 48 conterminous states. A sample of nearly 16,000
segments is visited annually in the June Enumerative Survey (JES), the main annual
agricultural survey conducted by NASS. Much of the information collected by NASS -and the
area frame sampling design used by NASS is compatible with the major objectives of the
Agroecosystem component of EMAP. The major source of information for this summary of
the NASS area frame sampling design is a USDA/NASS publication, "Area Frame Design for
Agricultural Surveys" (Cotter & Nealon, 1987).
3 2
-------
Figure 3.1. The EMAP hexagon grid.
3 3
-------
The NASS area frame uses stratification of the land area to improve precision of the
survey estimates, and multi-step sampling to reduce the cost of defining sample units. The
NASS area frame has been constructed state-by-state; the area frame for each state is rebuilt
(the old frame discarded) every 10 to 15 years, or as needed to correct problems with the
frame that become evident from continuous quality control monitoring. The total area of each
state is divided into land-use strata reflecting intensity and type of agriculture. A primary
requirement in defining strata (and in the subsequent definition of primary sampling units and
sampling segments) is that boundaries must be permanent and easily identifiable on the
ground by an interviewer. Typical strata are:
1) General Cropland, 75% or more cultivated,
2) General Cropland, 50-74% cultivated,
3) General Cropland, 15-49% cultivated,
4) Ag-Urban, less than 15% cultivated, more than 20 dwellings per square mile,
residential mixed with agriculture,
5) Residential/Commercial, no cultivation, more than 20 dwellings per square mile,
6) Range and Pasture, less than 15% cultivated, and
7) Water.
The stratification process uses satellite imagery (multispectral scanner and thematic
mapper), National Aerial Photography Program aerial photographs (1:58000 or 1:40000),
topographic quadrangle maps (1:24,000) from USGS, Bureau of Land Management maps
(1:100,000), and county highway maps.
Each stratum is subdivided into PSUs. The target size of the PSUs vary with the
stratum and is designed to contain, on the average, six to eight final sampling segments. The
target size of the final sampling segment varies from 0.1 square mile for residential strata to
one square mile for most agricultural strata to four square miles for rangeland and other low
intensity agricultural strata. As for the original strata, permanent and readily identifiable
boundaries are a requirement for all PSUs and segments. The boundaries of the sample
segments within the PSUs are not delineated until after the sample of PSUs has been selected.
The number of final sample segments in each PSU is computed as the area of the stratum
divided by the target size of the segment for that stratum. The sum over all PSUs gives the
total number of final sample segments per stratum.
A second level of stratification is done by ordering the population of PSUs within a
state using a criterion of agricultural similarity (the initial stratification was based primarily
on percent of cultivation). This substratification is particularly effective in areas of intensive
cultivation where crops vary across the state.
Construction of the area frame is expensive, approximately $150,000 per state on the
average. Most of the cost is labor ($113,000). During the construction of the area frame
there are many quality control checks to ensure consistency in definition and integrity of the
strata and the PSUs. In addition, there is a continual review of the area frame sample to
3 4
-------
uncover problems. This includes identification of problem segments (poor boundaries or
segments too large to cover in a specified time by an enumerator), replacement of photos with
newest materials, post-survey analysis to detect problems, and maintenance of historical
information on the problems (to assess when the area frame for a state needs to be updated).
The selection of the sample follows a two-step procedure. A replicated sampling
scheme is used because replicated sampling has numerous advantages. For each replicate,
one PSU per substratum is selected at random with replacement and with probability
proportional to size where the measure of size is the number of segments assigned to the
PSU. Then the boundaries of the sample segments are defined, but only for the PSU's
selected which greatly reduces the cost of defining sample segment boundaries. The second
step of the sample selection is to choose one sample segment from each selected PSU with
equal probability without replacement. Because the sampling between replicates is with
replacement, there is the possibility that more than one segment may be drawn from the same
PSU. Nealon, summarizes the nature of the 1987 sample as follows: "There were 15,665
segments in the national area frame sample. This represents approximately 0.6 percent of the
total segments in the 48 states. The allocation of the sample across land-use strata naturally
concentrates the majority of the sample segments in the cultivated strata. Approximately 46
percent of the national sample is in the intensively cultivated strata, 24 percent in the less
intensively cultivated strata, 19 percent in the range or pasture strata, 7 percent in the ag-
urban strata, 3 percent in the urban and resort strata, and 1 percent in the non-agricultural
strata." (Cotter & Nealon, 1987)
3.4.2. The NASS Multiple Frame and Sample
In sampling rare items, such as inventories of cattle feeding operations, especially
where a large share of the total inventory in the country may be concentrated in a few feed
lots, NASS uses a list frame in conjunction with its area frame. It may be that ARG will
wish to take advantage of this option for some of its indicators. Therefore, the principles of
multiple frame sampling are presented here.
Samples from the list frame are usually more efficient, because size of operation data
are ordinarily available for stratification purposes. Hence, large producers can be sampled at
a much higher rate and low producers at a much lower rate with gradations in between.
However, a list frame is usually seriously flawed in that it is incomplete and can never
be kept up-to-date. Thus, any estimate made from a sample selected from such a frame is
biased; the seriousness of the bias depends upon the number and size of the operations that
are excluded from the frame and thus from probability of being included in the ultimate
sample.
The area frame can be constructed and maintained so that it is complete, provided that
all the area in a given geographic universe is included. Examples of complete area frames for
the conterminous U.S. include both the NASS and EMAP frames (and perhaps others).
3 5
-------
Hence, there is no bias from lack of completeness in samples selected from such an area
frame.
Because area frames rarely can take advantage of ancillary information the way list
frames do, estimates from area frame samples are usually less precise. NASS does stratify its
primary sampling units by the percent of cropland to increase precision and this procedure is
quite effective. EMAP does not attempt to stratify at the Tier 1 level, because it is designed
to meet the requirements of six widely disparate ecosystem groups.
A third alternative, which can take advantage of the desirable features of both frames,
is the combined or multiple frame. In essence, the list frame is sampled to increase the
precision of the estimates and the area frame is sampled to eliminate the bias of
incompleteness in the list frame. To use this procedure, all list frame units must be removed
from the area frame when making the estimates. Thus, only a member of the universe which
is not on the list frame (usually referred to as the non-overlap units) can be selected in the
area sample. In practice, the overlap is measured only on the area frame segments that are in
the sample.
The enumerator, upon reaching the segment, will interview all within the segment who
are eligible according to her inclusion instructions. Upon completion of the segment, he or
she will then check with the list frame and identify the duplicates. Any remaining member
will belong to the area frame sample. Those in the list sample, who happen to be in the
segment, also will have been enumerated but their results will be expanded with the list frame
sample. By this procedure, all those selected in the list sample, who are in the segment, will
be interviewed but the results from those who were not in the list frame sample will not be
eligible for inclusion in the list sample totals. Of course, all the sampling units in the list
sample that do not happen to fall in any of the sample segments are also interviewed. The
two samples would be expanded independently and the results added for a final estimate. In
summary, the estimate would look like this:
(list sample in segments + list sample not in segments) x (expansion factor) +
(area sampling units in segments but not in list frame) x (expansion factor)
Some concern has been expressed that in the multiple frame approach utilized by
NASS the area frame estimates might not be complete. As noted above, the complete
segment is always enumerated and the list frame duplication is removed in the office. Thus,
the sample by itself is a sample from a complete area frame.
NASS uses the multiple frame approach in a number of its surveys, including
estimates of crop production and yield per acre, production expenditures, farm labor and
livestock inventories. The multiple frame procedure is described in Nealon (1984).
3 6
-------
3.5. EMAP-Agroecosystem Tier 2 Sampling
Two options for Tier 2 sampling of agricultural resources are presented in this section,
both depend on the NASS frame. Other plans are possible, but these appear to represent the
best potential for fulfilling the objectives. The EMAP design group initially suggested that
each resource group consider a sample size of 3,200 at the Tier 2 level. The ARG will apply
the usual criteria for determining sample size (i.e. indicator variability, desired precision, and
expected cost).
For this discussion, the Tier 2 sample size suggested by the EMAP design group has
been accepted. In the first plan, the centroids of 3,200 sample hexagons are used to
determine the NASS sample segments in which they fall. This plan will be designated the
Hexagon plan. Hexagon relates to the method of selecting the NASS segments. Both plans
will use the NASS-delineated sample segments.
The second plan, denoted as the Rotation Panel plan, proposes to utilize approximately
one-fifth of the NASS sample generated for their June Enumerative Survey. Rotation Panel
refers to the fact that one replicate of the NASS June Enumerative Survey (JES) will be
rotated out of the NASS sample every fifth year. Any overlap between this sample and the
hexagon sample would be purely a part of the random process.
Some estimated cost figures are shown. They are based on NASS experience but
should be viewed only as approximations for two reasons:
o Precise specifications cannot be given at this time.
o Estimates were based on 1987 costs with a 10% inflation rate compounded
annually through 1991.
Estimates from EPIC should likewise be considered as approximate.
The selection of the plan the ARG will use depends on the results of two studies:
o The two plans will be tested in a pilot study in North Carolina (Chapter 11).
Current plans call for a minimum sample size of approximately 50 segments
each. Accurate cost figures will be maintained for each plan. Similarly, the
relative precision of the two plans will be compared although a sample size of
50 each may be too small to detect meaningful differences; a sample size of
100 each is recommended to assure detection of differences and to aid in an
accurate determination of sample number for future pilot and demonstration
projects.
o A theoretical study is being conducted at the University of North Carolina at
Chapel Hill to compare the statistical efficiency of the two designs using the
3 7
-------
variance of an estimator for year-to-year change for each design. These
estimators will account for the rotating panel feature of the NASS design and
the full overlap associated with the Hexagon design.
3.5.1. The Hexagon Plan
Under this plan it has been proposed that NASS locate the centroid of each of the
12,600 EMAP hexagons and determine the stratum classification in which that point falls.
Using the NASS stratification system, approximately 3,200 hexagons would be selected at
random within strata so as to achieve optimum allocation as closely as possible. This would
be a probability two-dimensional sample but would lose its systematic nature to some extent
as, presumably, strata with a high proportion of cropland would be sampled at a higher rate.
NASS is able and willing to provide such a sample, given sufficient lead time.
The 3,200 sampling units would be divided so as to provide four interpenetrating
samples of 800 each, with one set of 800 being enumerated each year. These sampling units
would remain throughout the life of the program. This would comprise the Tier 2 Hexagon
sample (Table 3.1).
A rough cost of identifying the location of the centroids, selecting the PSUs and
segments and delineating the segments on aerial photos and highway maps is given below.
1) Identifying the 12,600 sample centroids on NASS materials @ $8.00 per
centroid = $100,800.
2) Cost of drawing the sample.
Costs are based on those given in the NASS publication"Area Frame Design
for Agricultural Surveys" dated August, 1987. Allowing for inflation, cost was
estimated at $95 per range segment and $165 per non-range segment. It was
estimated that about 10% (approx. 320 segments) of the sample might fall in
range areas.
320 segments @ $95 $ 30,400
2,880 segments @ $165 = 475,200
Total Sampling Cost = $505,600
If EMAP should want all PSUs with any land in the hexagon to be
characterized by NASS strata, costs would be increased considerably, but no
estimate is available.
3 8
-------
Table 3.1. Hexagon Plan Sampling Schedule.
NASS
Repl.
1
2
3
4
5
6
7
8
9
10
11
12
Year in EMAP Monitoring
1
X
2
X
3
X
4
X
5
X
6
X
7
X
8
X
9
X
10
X
11
X
12
X
13
X
14
X
15 ...
X
X-Data Collected on All Indicators
-------
3) Cost of Field Enumeration
Again with sufficient lead time, NASS could undertake the field work on the
selected segments. The collected data would be subject to NASS'
confidentiality restrictions (Chapter 10). The cost of a survey similar in scope
to its June Enumerative Survey is estimated at $140 per segment.
3,200 segments @ $140 = $448,000
4) Total cost of Turn-Key Operation = $1,054,400
These estimates do not include the cost of adding EMAP questions to the NASS
questionnaire. Nor does it include the cost of revisiting the sample segments during the year
in order to obtain indicator measurements at appropriate times.
3.5.2. The Rotation Panel Plan
This plan essentially takes advantage of the NASS sample and the data collected under
its auspices. Under this plan, EMAP would use for its Tier 2 agricultural resource sample
approximately 20% (approx. 3,200) segments of the NASS June Enumerative Survey (JES)
sample. These segments would be located without regard to the location of the EMAP
sample hexagons.
The NASS sample has five interpenetrating replicates of the total sample, designed
such that each replicate rotates out of the sample after five years. Again, there are optional
plans available for using the NASS sample. The one ARG has selected for testing uses a
strategy that participates in the NASS rotation with the result that for the Rotation Panel plan
sampling units would always be a subset of the JES sample.
To be comparable with the Hexagon plan, the Rotation Panel plan would sample 800
segments per year. For example, the first sample in Year 1 would be a probability sample of
1 in 4 of the first replicate, drawn in such a way so as to provide national estimates of the
EMAP indicator being measured. Similarly, the second, third and fourth years would be
samples of 800 from the 3,200 segments from replicates 2, 3, and 4 respectively. In year 5,
EMAP would again sample the indicators on the same 800 segments measured in Year 1; in
addition, 800 from replicate 5 would also be measured in year 5. Thus, two sets of
comparisons are possible: the difference in the sample of replicate 1 between year 1 and year
5 will provide an estimate of change; and the difference between Replicate 1 and Replicate 5
(e.g. at year 5) will allow a "calibration" or relation of the differences in Replicate 5 between
Year 5 and year 9 to those observed in Years 1 and 5. Similar patterns will follow each year
with measurements of differences in replicate 2 in Years 2 and 6, replicate 3 in Years 3 and
7, etc. This sequence is shown diagrammatically in Table 3.2 where "x" represents the
estimate of a variate measured on a sample of size 800. This design is flexible enough to
3 10
-------
Table 3.2. Rotation Panel Plan Sampling Schedule.
NASS
Repl.
1
2
3
4
5
6
7
8
9
10
11
12
Year in EMAP Monitoring
1
X
2
X
3
X
4
X
5
X
X
6
X
X
7
X
X
8
X
X
9
X
X
10
X
X
11
X
X
12
X
X
13
X
14
X
15 ...
X
.
X = Data Collected on All Indicators . = Some Data Could Be Collected By NASS on Annual Visits
-------
include the 800 segments in intervening years or to add subsamples of the 800 in intervening
years in case more frequent observations of change are desired. It should also be noted that
for certain characteristics such as land use, estimates of change will be possible every year
from the entire JES sample.
3.5.3. Advantages and Disadvantages of the Plans
o The principal advantage of the Hexagon plan was that the characteristics of the
hexagon in which the sample segment falls were to be completed by EMAP-EPIC as
part of the Tier 1 procedure. If resources of two or more groups are co-located in the
Tier 2 sample, it may be possible to observe some measures of association between
variables on the different resource groups. The fact that each group will attempt to
maximize the coverage of variables of interest to its own resource may work against
having a meaningful sample size in joint distributions.
o Inclusion probabilities, although perhaps not straightforward in the case of the Rotation
Panel design will certainly not be as complex as those in the Hexagon plan since the
latter involves the overlap of two probability samples.
o The cost of the Rotation Panel plan will be much less than the Hexagon plan because
the former will be able to piggy-back on the JES and only incremental costs would be
incurred. Since under the Hexagon plan, no new samples would be required, the
sample design effort would be a one-time cost. However, in the latter plan NASS will
not be visiting the same segments, so the cost of enumeration would have to be met
every year.
o Stratification coupled with near optimum allocation increases the precision of the
NASS estimates. It is not known if the same level of increase can be expected from
the application of these principles to the Hexagon sample.
o Computer techniques are already in place for making estimates under the NASS
sample. These procedures would have to be developed for the Hexagon sample, again
at additional cost.
o Estimates of land use and changes in land use can be made annually using the JES in
conjunction with the Rotation Panel plan.
o The Rotation Panel plan has the disadvantage of losing precision (on a per unit basis)
on the estimation of time trends because each sample unit is measured only twice in a
four-year EMAP cycle. It is conceivable that this disadvantage would be more than
offset by a larger sample size for a given cost.
o Both plans are subject to response biases of two kinds but in different ways.
Conditioning, because contact with the farmer may influence management practices, is
3 12
-------
potentially present in both plans. Under the Rotation Panel plan, the farmer will be
contacted every year for five years and then will be out of the sample. In the
Hexagon sample, the farmer would be contacted on a four-year cycle but would
continue to be contacted for perhaps decades. This pattern of contacts also places a
burden on the respondent which may influence the response rate. It is not known how
these procedures may affect the plans differentially; unfortunately, the pilot test will
not be able to shed any light on this question.
3 13
-------
4. Field Sampling
4.1. Introduction
The purpose of this section is to discuss the development of protocols for obtaining
indicator data from the sample segments. This includes discussion of protocols for selection
of sample fields within segments, choice of sampling density and location of sample points
within fields, data collection procedures, and development of enumerator training manuals.
All protocols will be developed in close cooperation with NASS and, in many cases, will use
NASS protocols or variations of them. For this reason, some of the current NASS protocols
will be described.
The collection of data for each of the indicators will involve extensive planning. Most
agricultural land is owned privately and the operator's perception of what is being done is
very important. This introduces considerations such as access permission, respondent burden
and confidentiality of data, all of which must be considered. Some indicators, such as soil
chemicals, will differ greatly depending on whether the samples are taken before or after
fertilizer or pesticide application. The field sampling must be coordinated so that the data are
collected at the appropriate time. This may involve multiple visits to the field during the year
to ensure adequate coverage.
Field sampling procedures are designed to produce accurate and precise information
subject to logistic, cost and institutional constraints. The variance structure of the indicators
will be explored during the pilot (Chapter 11} to assist in the determination of sample size
and allocation of the sample to segments, fields within segments, and indicator measurements
within fields. Although this chapter concentrates on developing procedures for sampling in
agricultural fields, similar" procedures must also be developed for sampling adjacent
uncultivated areas (.e.g., biomonitors, noncrop vegetation).
4.2. Statistical Issues
4.2.1. Field Selection
A probability sample of fields will be selected from the segment. The Agroecosystem
Resource Group (ARG) will use the protocol developed by NASS for their Objective Yield
Survey, with the sampling rate modified to meet sample size requirements.
In the Objective Yield Survey, NASS chooses random fields with probability
proportional to size (.Fig. 4.1). This is accomplished by sequentially numbering all acres of
all fields in the selected segments of a particular strata from 1 to N. If n fields are to be
selected per stratum the sampling interval. K, that is required is K=N/n. A random integer
between one and K is selected. The fields sampled are those fields that contain that
numbered acre, and even" Kth acre following it. Fields may be selected more than once (.if
their acreage is larger than the sampling interval. K). In such cases, the specified number of
4 1
-------
independent samples will be drawn from the field. Data collected by NASS in the North
Carolina June Enumerative Survey (JES) in 1990 are being used to estimate the field
sampling rate to be used in the pilot study.
SEGMENT
53 acres
Farmer B
Farrier A
45 acres
39 acres
Farmer C
Fields are selected based on
their acreages.
Figure 4.1: Selection of field within segment.
4.2.2. Choosing the sampling points within the field
Current NASS protocol for selecting sample points within fields uses a random path
from the corner of the field that is most accessible. The enumerators are given a series of
rules for walking into the field that will direct them to a random point. For choosing sample
points within fields, the ARG will use an extension of the NASS protocol to allow for more
sample points per field. Starting points on the perimeter of the field will be established by
protocol and the enumerator will be given random paths (number of paces "over and down")
to identify the desired number of random points within the field.
4 2
-------
4.2.3. Sampling density within the field (soil samples, nematode samples, etc)
The primary indicators that will require sampling within fields are 1) soil fertility and
nutrient-holding capacity, 2) soil contaminants, 3) soil erosion and 4) pest density. In
addition, measurement of irrigation water quality will require sampling in or near the selected
field. The sampling density within each field will depend upon the indicator. We anticipate
that for soil analyses eight composited soil samples per field with one chemical analysis
would be adequate. However, due to the clustering nature of nematode populations, at least a
composite of 16 soil samples would be needed. Thus, the basic soil sample for the pilot
study will be a composited sample of soil collected from 16 points in the field, from which
both soil and pest density analyses will be preformed. The pilot study will include provisions
for estimating the variance associated with field sampling and with laboratory analyses. The
variance estimates will be used to improve the within-field sampling procedure.
Topographic position is an especially important factor in determining soil chemical
and physical properties, and particularly in determining temporal changes. In fields with a
slope, the soils on the lower part of the slope are nearly always more fertile than credible
soils on the upper slope. It is possible that compositing soil samples over all slope profiles
will mask the temporal deterioration of eroding slopes. The protocol for handling this
problem has not been finalized. One procedure would sample upper and lower slope strata
separately. This approach would allow estimates of population characteristics to be made
separately, allowing for a more complete picture of the soil properties. It has the
disadvantages, however, of being more logistically complex. Also, the identification and
definition of upper and lower slope may be difficult and separate acreage would need to be
determined for the two categories. An alternative procedure would sample the entire field as
described with soil cores composited only within strata. This has the advantage of allowing
simpler field sampling procedures and separate estimates of trend on upper and lower slopes.
However, the procedure still requires the identification of upper and lower slopes by the
enumerators in the field.
4.3. Field Sampling Techniques
4.3.1. Interacting with NASS Enumerators
NASS enumerators will have an integral role in data collection for the Agroecosystem
program. The enumerators are well trained by NASS in conducting agricultural surveys.
They conduct interviews and take field measurements, such as objectives yields, as needed.
The NASS enumerators are themselves members of the agricultural community and, therefore,
are people the farmers and operators know and trust.
Because NASS is already involved in a national survey effort, enumerators are in
place throughout the United States. When the ARG provides the appropriate training
materials, NASS will train enumerators in the field sampling techniques that will be needed to
collect indicator data and samples (Chapter 7). Because many of the inherent logistical
4 3
-------
problems have already been solved, this will greatly facilitate the implementation of the
Agroecosystem Resource Group's data collection process.
4.3.2. Sampling Protocol
Before field measurements can be taken, protocols for the gathering and handling of
that data must be specified. NASS and the ARG will work together on the development of
these protocols, with the ARG having the primary responsibility.
Among the more important protocols that will be finalized before field implementation
are:
Soil sampling protocols
o Specify a set of random points at which to take the sample within the
chosen field. This includes a clear understanding of special
circumstances that require contingency plans that alter the normal
sampling process.
o The mechanics of taking the sample from the soil.
Water sampling protocols
o Selection of random points in a stream or irrigation ditch. Selection of
wells for sampling ground water.
o How to take a water sample, from where, and at what depth.
o Number of samples needed to get a representative sample from a water
system.
Field selection protocols
o Number of fields needed for an adequate sample size, and the
distribution of fields among the segments. (This is currently being
explored in simulation studies being conducted on data from the 1990
NASS June Enumerative Survey).
Care and handling of samples
o Storage of the samples in the field, enroute to the lab, and in the lab.
Determine how they will be labeled and prepared for analysis while in
the field. Determine the conditions for sample storage. The procedures
4 4
-------
must insure that the sample never loses its identity in terms of where,
when and by whom it was taken.
Quality assurance interface
o Determine what measures can be taken to ensure that laboratory
analyses are providing results with acceptable accuracy and precision.
4.3.3. Training Manuals
The training manual for sampling protocols will be the joint responsibility of NASS
and the ARG. The ARG will define what samples are required and how they are to be
collected. This will include information on how to choose the random points, at which points
to collect the sample, how the samples are collected, how the samples are handled and the
process of moving them from field to laboratory. The ARG will ensure that the overall
logistics plan is complete (Chapter 7). NASS will ensure that the training manual is suitable
for use by its enumerators. It will also see that the training manual is written in language that
is clear and detailed enough that the enumerators fully understand the sampling process. The
training manual should contain information on special circumstances that will require
modifications to the sampling process. It should answer most questions that an enumerator
might have about indicator sampling.
4.3.4. Survey Errors
Throughout the entire survey process, the ARG must be alert to the kinds of errors
which may occur and take steps to control, minimize and measure them. There are two main
classes of errors: sampling and non-sampling. Sampling errors, arise from the fact that only a
small portion of the population of interest is examined. They are controlled through a
properly designed sample from which estimates of error can be made from the sample data.
Non-sampling errors are associated with data collection efforts in the field. It is
important to identify steps in the survey process where these errors may occur and plan to
minimize their impact. It is necessary to develop field protocols and training procedures that
help minimize non-sampling errors. Through years of experience, NASS has identified a
large number of potential non-sampling errors and has classified them into five main types: 1)
specification error, 2) coverage error, 3) response error, 4) non response error, and 5)
processing error. NASS has established procedures to help control these types of errors.
Specification error occurs at the planning stage of a survey because data specifications
are inadequate or inconsistent with the survey objectives. Sources of error can occur from 1)
poorly stated uses and needs, 2) changing uses or needs, 3) the population of interest not
being the same as the population surveyed, 4) poorly defined concepts, 5) ambiguous
definitions, 6) proxy data used instead of primary data, and 7) unclear questionnaire
instructions.
4 5
-------
Coverage error, which includes both undercoverage and overcoverage, is defined as
the error in an estimate that results from 1) failure to include all units belonging to the
defined population or failure to include specified units in the survey and 2) erroneous
inclusion of some units because of a defective frame, the inclusion of unspecified units, or the
inclusion of specified units more than once in the actual survey.
Response error, which occurs in the data collection phase of the survey, may be
thought of as the difference between the value collected during the survey and the correct
value. Response errors may result from 1) the failure of the respondent to report the correct
value (respondent error), 2) the failure of the interviewer to record the value correctly
(interviewer error), or 3) the failure of the instrument to measure the value correctly.
Non-response error results from a failure to collect complete information on all units
in the selected sample. Non-response produces error in survey estimates in two ways. Most
importantly, a bias may be introduced to the extent that non-respondents differ from
respondents within a selected sample. Secondly, the decrease in sample size or in the amount
of information collected in response to a particular question results in larger standard errors.
Processing error is the error in final survey results arising from the faulty
implementation of correctly planned survey methods. Sources of processing error can occur
at several different parts of the survey. Questionnaires could be designed in a way that
results in skip patterns or boxes left incomplete. Typographical errors in questionnaires could
cause a problem in collecting the correct data. Errors in data editing, data entry and data
analysis are also included in this type of error.
If a particular type of non-sampling error is systematic in nature, and its deviation
from the true value is in the same direction and of approximately the same magnitude, from
one sampling unit to another, the error is usually referred to as a bias. As noted above,
estimates made when non-response is high are usually biased since those failing to respond
are normally different from those who respond. Quite often in agricultural surveys, non-
respondents have lower values of the variable being measured, that is, fewer acres, fewer
cattle, or fewer fruit trees. Similarly, an incorrect calibration on a measuring instrument will
result in bias.
The EMAP-Agroecosystem program will rely heavily on the experience of NASS
personnel to avoid problems that might otherwise trouble a new sampling effort.
4.4. Evaluation of Field Sampling Techniques
The pilot project (Chapter 11), will be an opportunity to test many of the field
sampling protocols and identify potential problems. Among the goals for field sampling are:
1) Ensure that a probability sample both of fields and of samples within fields is
being selected;
4 6
-------
2) Test the effectiveness and thoroughness of field sampling protocols;
3) Recognize and address sources of non-sampling errors;
4) Determine that the data gathering processes from field computer dataset, are
providing the most complete and accurate information;
5) Ensure that our interaction with NASS is smooth and mutually beneficial.
4 7
-------
5. Data Analysis, Integration and Assessment
The analyses of the monitoring data will proceed along several lines. The initial
analyses will be directed toward the routine summarization of the indicators for reporting the
status and trends of the health of the ecosystem in annual statistical summaries. Interpretive
analyses will be aimed at understanding the correlation structure (associations) among the
indicators, regional spatial patterns and concordance of spatial patterns for stressor and
response indicators, time trends, and the development of health indices for the agroecosystem.
The relationship of these two types of analytical approaches can be seen in Figure 5.1. The
purpose of the annual statistical report is to report by region and nationally the status of the
agroecosystem and the observed changes in status as reflected by the indicators. It is
anticipated that the interpretive analyses will serve two purposes. In the initial phases of
EMAP. considerable effort will be devoted to understanding the correlation structure of the
indicators to identify essential indicators for the monitoring process and to develop an index
of ecological health. The second purpose of the interpretive analyses is to reveal relationships
that may suggest mechanisms for observed changes and help to define further research needed
to establish causality. Interpretive analyses may also help identify linkages to other
ecosystems.
5.1. Data Summaries
The annual statistical summaries will report the regional behavior of the indicator
variables by presenting key features of their probability distributions in tabular form or
through graphical techniques using estimated cumulative distribution functions (CDFs) with
associated measures of confidence, or box-plots. Box-plots are a compact method of
reporting key quantiles and extremes for distributions. The CDFs will be constructed to give
the proportions of land area having values for the indicator less than values specified on the
abscissa of the graph. Approximate confidence interval estimates of the population estimates
of the proportions will be superimposed on the CDFs to provide measures of confidence in
the estimates. Estimates of area of land in agricultural use and for other specific categories of
use will be reported. To facilitate interpretation of the CDFs, key quantiles from the CDFs
will be reported in tabular form. Time trends will be reported, once sufficient data have been
accumulated, by plotting key quantiles from the CDFs against time or by giving a time
sequence of several box-plots. In order to maintain confidentiality of data, regional spatial
patterns of certain key indicators will be displayed using a smoothing procedure to obscure
sample site locations.
5.2. Interpretive Analyses
Part of the interpretive analyses of indicator data will be aimed at quantification of
relationships known a priori to be important to the health of the ecosystem. Correlation and
regression analyses will be used to quantify and test these relationships. A major portion of
the analyses, however, will be exploratory in nature; the methods used will depend on the
specific questions addressed. In general, multivariate techniques (principal component
5 1
-------
Indicator 1
Raw Indicator Data
Initial Summary
Development of
Indices of
Agroecosystem Helath
1
Raw Indicator Data
Annual Statistical Reports
Regional Probability Distributions
Spatial / Temporal Trend Analyses
Multivariate Analyses to Determine
Inter-Indicator Relationships
Detection of Trend
Association Between
Agroecosystm Health
and Stressors
Indicator n
Raw Indicator Data
Initial Summary
i
Identification of
Indicator Relationships
Across Ecosystem Boundaries
Ancillary Data
Figure 5.1: Analytical approach with agroecosystem indicator data.
-------
analyses, multivariate analyses of variance, multiple regressions, path analysis, biplots, etc.)
will be used to gain an understanding of the relationships among the indicators at a given
point in time as well as over time. This will reveal redundancies among the indicators being
measured and may lead to reduction in number, or redefinition, of indicators being measured;
the analyses may suggest additional indicators. As the indicator set becomes better defined
and understood, it is anticipated that specific indices of agroecosystem health will be
developed. As trend data become available, emphasis of the analyses is expected to shift to
the detection of trend associations between ecosystem health and stressor indicators. These
associations may suggest hypotheses as to the mechanisms associated with health changes.
The hypotheses would then be independent.
Another important component of interpretive analyses will be the use and development
of spatial statistics. Spatial methods, such as kriging, are used to analyze and interpret data
that are spatial in nature or have a spatial component. Much of the data gathered by the
Agroecosystem program will be of this type.
5.3. Ancillary Data
In addition to the indicator variables themselves, extensive information will be
collected on ancillary data. These data will be included in numerous analyses (Table 6.4).
Ancillary variables are those variables which might influence the indicators independently of
what the indicator is designed to denote. The ancillary variables will be used as covariates in
the models to remove associated variation in the indicators and improve the estimates of the
indicators.
Variables such as climate, weather, socioeconomic measures and catastrophic events
such as hurricanes, tornados, and pest infestations play an important part in spatial and
temporal variations of the ecosystem. Accounting for these variations should facilitate an
understanding of the innate relationships among the indicators and the health of
agroecosystems. For example, certain stressor indicators may become important to the health
of the ecosystem only under certain environmental conditions. Judicious choice of ancillary
variables can strengthen, explain, and clarify some of the relationships that may be discovered
among the indicators. There also may be strong relationships of interest found between the
indicators and the ancillary variables.
5.4. Integration of Information
Another phase of the more interpretive analyses is to investigate specific relationships
of interest that cross ecosystem boundaries. Part of the information for this linkage between
ecosystems may come from indicators measured by the ARG in areas adjacent to agricultural
fields, such as wildlife habitat in boundary areas surrounding those fields or pesticide,
fertilizer, and silt exports to surrounding water systems. A study of the concordance of
spatial patterns for key indicators from different ecosystems may also provide some
suggestion of relationships to be studied further in Tier 3 or 4 studies. It is anticipated that
5 3
-------
most of the data to clarify linkages between ecosystems will come from specifically designed
Tier 2 sample surveys or Tier 3 or 4 research studies.
5.5. Assessment Activities
One of the primary roles of the ARG will be to assess Agroecosystem health.
Ecosystem health is a difficult concept to define and interpret. Because agricultural systems
are intensively managed, agroecosystem health encompasses several dimensions not usually
present in less disturbed systems. Concepts such as sustainability, productivity, the resources
required to produce a given yield, pest management and other agroecosystem-specific ideas
all contribute to the difficulty of defining health. In addition, the concept of health must
include information on surrounding ecosystems.
Because EMAP is designed as a longitudinal study of change in agroecosystems, the
health of those systems will be among the more important features monitored. Some measure
of baseline health will be established and then the change in health will be observed over
time. As our knowledge and understanding about the indicators grow, we expect to be better
able to detect and define changes in ecosystem health.
An important part of agroecosystem monitoring will be the use of geographical
information systems (GIS). These systems allow the digitization and manipulation of
remotely sensed data such as satellite imagery and aerial photography. GIS will allow the
monitoring of changes in agroecosystem landscapes. Measures such as patch edge-to-area
ratio, size and distribution of hedgerows and shelterbelts, changing land use patterns (e.g.
wetlands being converted to agriculture, agriculture lost to desert, etc.) and other landscape
descriptors such as the fractal dimension of patch boundaries will be most easily evaluated
using GIS technology (Chapter 6).
Because spatial statistics will play a vital role in assessing indicator data, new areas of
investigation to identify and employ new statistical techniques will likely open. In addition,
the comparison of remotely sensed and ground-truth data will require the use of statistical
techniques such as confusion matrices and other measures of correlation between ground-truth
and remotely sensed data.
5 4
-------
6. Indicator Development and Evaluation
Traditionally, environmental monitoring programs have used a "bottom-up" approach
to risk assessment that emphasizes the measurement of stressors (e.g., ambient monitoring of
ozone levels in urban areas) and use of mathematical models of transport, transformation and
fate processes to estimate exposure levels (e.g., kriging estimates of ozone levels in
croplands). To predict ecological effects, exposure-response models are used to couple
estimates of exposure with the results of laboratory lexicological research. This approach is a
reasonable way to assess the impact of a single pollutant, where transport, transformations and
effects are understood and can be modeled reliably. However, ecological resources are
affected by multiple stressors which arrive in multiple media and interact with management
actions and natural processes. The "bottom-up" approach cannot be used to detect
unanticipated pollutant interactions or stressors reliably, and often the use of this approach
does not allow modeling of the effects of stress on poorly-understood natural processes.
The EMAP strategy for monitoring and assessing the status and trends in ecological
resources is based on a "top-down, risk assessment" approach (Hunsaker and Carpenter 1990,
Knapp et al. 1990). Within a risk assessment framework, anthropogenic influences which
have the potential to adversely impact indigenous populations (e.g., pesticide export to non-
target areas) are referred to as stressors. The magnitude of the stress to which organisms are
exposed depends on the concentration and duration of the exposure, the habitat characteristics
and physical conditions prevailing at the exposure site, and the preexisting condition of the
organisms at the time of exposure. Most organisms have a variety of compensatory responses
that minimize or reduce exposure to pollutants, including avoidance and modification of
various physiological processes. If organismal responses fail to reduce stressor exposure,
impaired function, alterations in condition (i.e. from healthy to pathological), or death may
result. Such responses may result in adverse effects on higher-order ecosystem attributes that
are beneficial to society (e.g., diversity and abundance of wildlife, protection from erosion).
A top-down approach focuses on the valued attributes of the system, or the endpoints
of concern, rather than the stressors. In top-down risk assessments, the observation of an
effect stimulates efforts to identify the stressor that caused the effect. In EMAP this is
accomplished by focusing on changes in the system and associations among various indicators
of exposure and response. The top-down approach is more likely to detect cumulative
impacts of natural and anthropogenic influences on the ecosystem than are other approaches.
6.1. Assessment Endpoints and Indicators
As stated in Section 1.3., the goal of the Agroecosystem program is to monitor and
assess the long-term status and trends in the health of the nation's agricultural resources from
an ecological perspective through an integrated, interagency program. Because agroecosystem
health cannot be measured directly, a suite of indicators will be monitored that will
collectively describe the overall condition of agroecosystems. An environmental indicator is
6 1
-------
defined as a measurable surrogate for values (the assessment endpoints) deemed important by
society.
The first phase of the indicator development process is to establish a framework for
indicator selection and interpretation by identifying the assessment endpoints and
environmental stressors of primary concern for agroecosystems. Assessment endpoints are
defined as "formal expressions of the actual environmental value that is to be protected, and
as such should have unambiguous operational definitions, as well as social or biological
relevance" (Knapp et al. 1990). The three assessment endpoints which will be used to focus
the interpretation of indicator data are shown in Figure 6.1.
Sustainability
Contamination
of Natural
Resources
Status & Trends
in Agroecosystem
Health
Quality of
Agricultural
Landscapes
Figure 6.1: Agroecosystem assessment endpoints that will be addressed with a suite of indicators to determine the
status and trends in agroecosystem health.
The indicators selected must be responsive to the major environmental stressors of
concern. Stressors are considered as those physical, chemical, biological or social
6-2
-------
disturbances which affect the ecological health of agroecosystems both positively and
negatively. Major stressors or potential stressors of agroecosystems include:
o Air pollutants
o Water pollutants
o Land management practices
o Wastes
o Agricultural chemicals
o Global climate change
o Biological technologies
6.2. Indicator Categories and Conceptual Models
A key element of EMAP's top-down, risk assessment approach is the linkage of
indicators to assessment endpoints. Part of this approach is to use indicators that fall into one
or more of the following four indicator categories (Table 6.1):
1. Response indicator: an environmental characteristic measured to provide evidence of
the biological condition of a resource at the organism, population, community, or
ecosystem level of organization.
2. Exposure indicator: an environmental characteristic measured to provide evidence of
the occurrence or magnitude of contact with a physical, chemical or biological
stressor.
3. Habitat indicator: a physical, chemical or biological attribute measured to
characterize the conditions necessary to support an organism, population, community
or ecosystem in the absence of stressors.
4. Stressor indicators: a characteristic measured to quantify a natural process, an
environmental hazard or a management action that results in changes in exposure and
habitat.
Conceptual models illustrate the linkages between assessment endpoints, indicators and
environmental stressors. The development of a conceptual model is an important component
of the indicator development process. Conceptual models serve three primary purposes:
1. To define explicitly the framework for indicator interpretation (e.g., how the response
indicators relate to the assessment endpoints; the role that they play in determining
endpoint status, and how they will be used to assess that status).
2. To assist in the identification of any gaps within the proposed indicator group (e.g.,
missing indicators for the assessment endpoints, or links for which additional or new
indicators are needed).
6 3
-------
Table 6.1: Initial grouping of agroecosystem indicators within each indicator category.
Indicator Title
Crop productivity
Soil productivity
Land use
Landscape descriptors
Wildlife populations
Beneficial insect density
Pest density
Response of biomonitor species
Irrigation water quality
Irrigation water quantity
Agricultural chemical use
Non-point source loading
Foliar symptoms
Livestock production
Socio-economic factors
Genetic diversity
Habitat quality
Atmospheric variables
Response
Indicator
X
X
X
X
X
X
X
X
Exposure
Indicator
X
X
X
X
X
X
X
Habitat
Indicator
X
Stressor
Indicator
X
X
X
X
X
3.
To assist in guiding the data analysis strategy for determining potential causes of poor
condition.
Conceptual models can be constructed at many scales, ranging from simple models
which demonstrate the major components of the system (Fig. 1.2) to complex models
identifying all known linkages. At this stage of program development, the purpose of the
conceptual model of the agroecosystem is to delineate valued attributes and the major
stressors that may affect them, as is known from agroecological research.
6 4
-------
The ARG is constructing a comprehensive conceptual model that will identify
agroecosystem processes and link the indicators and stressors with the assessment endpoints.
Figure 6.2 presents a general conceptual relationship of the three agroecosystem assessment
endpoints with examples of response, exposure and stressor indicators. The relationship of
specific agroecosystem indicators to the three assessment endpoints is illustrated in Table 6.2.
6.3. Framework for Indicator Development and Evaluation
The selection and evaluation of indicators is critical to the success of the overall
EMAP program. Therefore, an indicator development framework has been constructed that
provides a set of goals and criteria and a consistent process by which indicators can be
objectively selected and evaluated across EMAP ecological resource groups (Knapp et al.
1990). This framework also provides flexibility to accommodate the addition of new
indicators and the refinement or removal of indicators which have proven inadequate to meet
the objectives of the program.
The EMAP indicator development framework consists of six phases (Figure 6.3) which
summarize the sequence of activities required to identify a candidate indicator and advance it
to the level of a core indicator where it can be implemented for agroecosystem monitoring on
a regional and national scale.
The use of clearly defined criteria in the identification, selection, and evaluation of
indicators encourages an objective perspective and a non-biased evaluation of all important
indicator characteristics prior to acceptance or rejection (Knapp et al. 1990). The ARG has
and will continue to use specific indicator selection criteria throughout the decision-making
process (Table 6.3). Although certain decisions made in the early stages of indicator
identification and selection were subjective, the selection criteria provide a way to organize
and record information about the decision-making process, so that another independent
evaluation can confirm the validity of the original decisions.
Indicator development will be an ongoing process of identifying, screening, and
evaluating indicators that are most appropriate for assessing the health of the nation's
agroecosystems. As the program develops, the emphasis will shift from indicator
identification to indicator evaluation, refinement and final selection. All phases of this
process will be revisited frequently, and indicators will be re-evaluated to ensure that the
program responds and adapts to new challenges, new knowledge, and new technology.
6 5
-------
o\
ON
Assessment
Endpoint
Response
Indicator
Exposure
Indicator
Stressor
Indicator
(a)
(b)
Contamination
of Natural
Resources
Changes in Honeybee
Populations
Honeybee Tissue
Concentrations
Agri-chemical Use
J
(C)
Quality
of Agricultural
Landscapes
Changes in
Wildlife Diversity
and/or
Populations
Changes in
Amount of
Habitat
Land Use
Figure 6.2: Examples of agroecosystem processes linking assessment endpoints with response, exposure and stressor indicators.
-------
Table 6.2: Association between the Agroecosystem assessment endpoints and indicators.
Indicator
Crop productivity
Soil productivity
Nutrient-holding capacity
Erosion
Contaminants
Microbial component
Land use
Landscape descriptors
Wildlife populations
Beneficial insect density
Pest density
Status of biomonitor species
Irrigation water quantity
Irrigation water quality
Agricultural chemical use
Non-point source loading
Foliar symptoms
Livestock production
Socio-economic factors
Genetic diversity
Habitat quality
Atmospheric variables
Sustainability
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Contamination
of Natural
Resources1
X
X
X
X
X
X
X
X
X
X
Quality of
Agricultural
Landscapes
X
X
X
X
1 Air, soil, water and biota, including transport of contaminants into, within, and out of agroecosystems.
6 7
-------
Phase 1
Identification of
Issues and
Assessment
Endpoints
Phase 2
Identification of
Candidate
Indicators
Phase 3
Selection of
Indicators for
Further
Research
ISSUES/ASSESSMENT
ENDPOINTS
Objectives
Identify indicators
linked to endpoints
- Qualitative evaluation
Methods
Expert Knowledge
Literature Review
Conceptual Models
Evaluation
Workshops
Criteria
CANDIDATE INDICATORS^
Prioritize based
on criteria
- Qualitative/Quantitative
evaluation
- reject, suspend, or
proceed
Expert Knowledge
Literature Review
Conceptual Models
Criteria
Peer Review
Phase 4
Evaluation of
Research
Indicators
Phase 5
Selection of
Core Indicators
RESEARCH INDICATORS
Evaluate expected
performance
- quantitative testing
and evaluation
Analysis of Existing Data
Simulations
Pilot Tests
Example Assessments
Conceptual Models
Criteria
Peer Review
PROBATIONARY CORE INDICATORS
Evaluate actual
performance on a
regional scale
build Infrastructure
demonstrate utility
assess logistics
Regional Demonstration
Projects
Regional Statistical
Summaries
Criteria at
Regional Scale
Peer Review
Agency Review of
Summary
Phase 6
Reevaluatlon of
Core Indicator
Set
CORE INDICATORS
Implement regional
and
National monitoring
periodic
reevaluation
EMAP Data Analysis and
Assessment
Correlate
Indicators with Proposed
Replacements
Feedback from
Peers and
Agencies
Peer Review
Assess Promising
Candidate Indicators
Revisit Assessment Endpoints
Figure 6.3: Six phases of indicator development followed in the Agroecosystem program.
6 8
-------
Table 6.3: Critical and desirable criteria for selection of indicators for an ecosystem monitoring and assessment
program.
Critical Criteria
Responsive
Regional Applicability
Unambiguous
Integrate Effects
Correlative
Important
Low Measurement Error
Interpretability
Explanation
Must reflect changes in ecosystem condition, and respond to
stressors of concern, or management strategy.
Must be applicable on a regional basis, and to a broad range
of regional ecosystem classes.
Must be related unambiguously to an endpoint or relevant
exposure or habitat variable.
Must integrate ecosystem condition over time and space.
Must directly measure or correlate with changes in
ecosystem processes or with unmeasured ecosystem
components.
Must reflect conditions that are important to overall
ecological structure and function.
Exhibits low natural temporal and spatial variability at the
sampling site during the index period to be able to detect
regional patterns and trends.
Clear interpretation or can be related through conceptual
models to meaningful changes in ecosystem structure and
function, or to change in stresses, habitat, or exposure
factors affecting the ecosystem.
Desirable Criteria
Simple Quantification
Standardized Method
Historical Data
Retrospective
Anticipatory
Cost Effective
Explanation
Can be quantified by synoptic monitoring or by cost-
effective automated monitoring.
Must have a generally accepted, standardized measurement
method.
Has a historical database or a historical database can be
generated from an accessible data source.
Can be related to past conditions by way of retrospective
analyses.
Provides an early warning of widespread changes in
ecosystem condition or processes.
Has low cost relative to its information value.
6 9
-------
6.4. Historical Overview of Indicator Development for EMAP Agroecosystems
The ARG met formally in early August 1989 to discuss assessment endpoints, major
stressors of concern for agricultural resources, indicator categories, and selection criteria. In
mid-August 1989, two workshops were convened to discuss assessment endpoints and to
identify a broad group of indicators that would be appropriate for monitoring and assessing
the health of agroecosystems. The forty workshop participants included members of the
ARG, university and government scientists, and USDA Agricultural Research Service (ARS)
scientists. Workshop participants generated 138 one-page reports on candidate indicators,
which upon examination for redundancy were reduced to 90 specific candidate indicators
(Heck et al. 1989). In follow-up meetings, the ARG critically evaluated each of the 90
candidate indicators and, based on the indicator selection criteria listed in Table 6.3,
recommended 16 candidate indicators for subsequent evaluation and development.
The ARG critically evaluated the 16 candidate indicators and prepared an extended
fact sheet for each (Heck et al. 1989). The proposed indicators were introduced in two
chapters of the overall EMAP document on ecological indicators (Campbell et al. 1990a, b in
Hunsaker and Carpenter 1990). An external review of these indicators was performed by the
EPA Science Advisory Board in May 1990. The ARG held a workshop in May 1990 to
discuss the status of the candidate indicators in relation to: (1) indicator selection criteria; (2)
indicator assessment endpoint associations; (3) index period and measurements associated
with each indicator; (4) the status of each indicator in relation to its readiness for
implementation in EMAP; and (5) prioritizing of indicators. Specific measurement and data
needs for each indicator were identified, along with the potential source for each measurement
(Appendix 5).
Indicators currently being considered by the ARG for agroecological monitoring are
presented in Table 6.2. Weather variables, ambient air pollution concentrations, crop
management practices and other important ancillary data essential to the development and
interpretation of indicator values will be collected along with indicator data (Table 6.4). The
five indicators selected for testing in the initial pilot project (Chapter 11) are crop
productivity, soil quality, irrigation water quantity and quality, agricultural chemical use, and
land use. Selection of these indicators was based on: (1) availability of existing techniques
appropriate for obtaining the measurements needed; (2) suitability of the indicator for use in a
single index period; and (3) interpretability of the data to be obtained. Data for most
indicators will be collected by NASS enumerators. An example of a NASS questionnaire
developed to collect indicator data in the pilot project and examples of data summary tables
are included in Appendix 6 and Appendix 7, respectively. The remainder of this chapter is
devoted to discussion of individual indicators. Detailed fact sheets on each indicator may be
found in Appendix 8.
6- 10
-------
Table 6.4: Ancillary data to be used in the development and interpretation of indicator values.
Data Needs
Data Source
Climate data
temperature (minimum, maximum, degree days)
total rainfall
rainfall during growing season
solar radiation (growing season and total)
relative humidity
catastrophic events (hail, hurricane, flood)
National Weather Service
NWS
NWS
NWS
NWS
NWS
Air Quality Data
gaseous air pollutant levels
dry/wet deposition rates
toxic release inventories
atmospheric dispersion models
EMAP Air and Deposition; EPA
EMAP Air and Deposition; NADP
EPA
literature
Management practices
use of manure (type , amount)
tillage practices (no. passes, type machinery)
use of sewage sludge and amount applied
use of fertilizers and micronutrients
herbicide use (type, rate, frequency)
nematicide use (type, rate, frequency)
insecticide use (type, rate, frequency)
crop variety
crop rotation history
NASS
NASS
NASS
NASS
NASS
NASS
NASS
NASS
NASS
question
question
question
question
question
question
question
question
question
Economic data
cost of chemical inputs
cost of irrigation water
literature
literature
Miscellaneous data
irrigation water temperature
severe pest or disease outbreaks
participation in CRP or other set-aside programs
NASS or EMAP sample
EMAP sample
NASS question
6.5. Indicators Addressing Sustainability
For the Agroecosystem program, sustainability is defined as the ability of an
agroecosystem to sustain adequate commodity (crop and livestock) yields over the long term,
without degradation of natural resources. Indicators addressing sustainability include crop and
soil quality, irrigation water quantity and quality, the density of beneficial insects, pest
density, foliar symptoms, agricultural chemical use, and socio-economic factors (Table 6.2).
6 11
-------
6.5.1. Crop Productivity
Crop productivity may be viewed as output from the system driven by a series of
inputs, both natural (e.g. solar energy and rainfall) and anthropogenic (e.g. tillage, seed,
fertilizer, pesticides and irrigation water). Crop yield, a major output of agroecosystems, will
be a major component of productivity assessment. Crop yield is one of the best-known
measures of function in agroecosystems, and has measurable responses to both natural (Huff
and Neill 1982, U.S. EPA 1978) and anthropogenically-generated stressors (Heck 1990,
Treshow 1984).
Yields of different crops will be converted into net primary productivity (NPP)
estimates (Sharpe et al. 1975, Turner 1987). Such estimates are needed to compare the
relative productivity of crops as diverse, for example, as wheat, cotton and potatoes.
Because the high crop yields in U.S. agroecosystems are subsidized by high produc-
tion inputs (fertilizers, pesticides, fossil fuel, irrigation water), an estimate of the efficiency of
these inputs in sustaining output is a conventional estimate of crop productivity. Crop
productivity, or efficiency indices have been calculated in different ways by different
agencies. An economic index (commercial value of the crop related to the cost of inputs) is
used by the USDA Economic Research Service (USDA/ERS 1989). However, there is no
single conventional way in which crop productivity is defined, or consistency in which inputs
and outputs are included in a productivity analysis (Tangley, 1986). Therefore, depending on
the measures one uses, agricultural systems can be considered more or less productive. Since
productivity measures are usually ratios of output to input, changes in their values may reflect
changes in input, output, or both. For example, higher productivity can mean higher yield,
but it can also mean the same yield, with fewer inputs (Tangley 1986).
The ARG is developing an approach to the assessment of crop productivity that would
estimate the efficiency of the inputs in achieving the desired output and also attempt to
account for beneficial or detrimental environmental inputs and outputs, such as the run-off of
fertilizer or changes in soil properties. This approach includes (1) assessing crop yield in
relation to the inputs; (2) determining if the quantities of specific inputs needed to maintain
yield are changing; and (3) developing a comprehensive productivity function in terms of total
system inputs and outputs (Figure 6.4). Olson and Breckenridge (1990) proposed a
sustainability index which includes pollutants from external sources, waste disposal "costs"
(e.g. the run-off of nitrates from fields into farm ponds or streams), and climatic factors, as
well as conventional inputs such as fertilizer, energy, and irrigation. Outputs include crop
yield and residual effects of management practices such as the status of critical biological and
soil factors, which could be estimated by other agroecosystem indicators.
6 12
-------
I Temperature
1 Precipitation
Climatic"^ severe Weather
Factors
Crop
froductivity
Inde^
Figure 6.4: Factors to be included in the calculation of a crop productivity index.
6.5.2. Soil quality
Soils are essential to the sustainability of terrestrial ecosystems. The extent to which
agricultural soils can tolerate intensive use without physical, chemical or biological
degradation will be a primary indicator of agroecosystem sustainability. Soil quality will be
characterized by soil structure, nutrient-holding capacity, acidity or alkalinity, the soil biota,
extent of contamination with agricultural or industrial wastes and rate of erosion (Figure 6.5).
Nutrient-holding capacity and contamination. To assess the status of the soil with regard to
desirable physical and chemical characteristics, data on soil properties such as soil organic
matter, cation exchange capacity, soil texture, bulk density, soil permeability, infiltration rate,
hydraulic conductivity, coarse fragments, and key soil fertility factors will be collected
(Appendix 8). The extent of contamination with heavy metals, persistent organics and other
contaminants will also be determined in soil samples. Data obtained for soils will be grouped
by multivariate techniques to assess the regional distribution of soils based on their structure,
nutrient-holding capacity, degree of fertility, acidity, salinization, contamination among others.
This information will be used as an essential data layer to compare with values of other
indicators, including crop productivity, land use, soil erosion, certain aspects of pest density,
management practices and agricultural chemical use.
Erosion. Erosion affects agroecosystems primarily through the loss of soil water-holding
capacity, nutrient-holding capacity and the alteration of soil structure (USDA 1981), and
presents a risk to both soil and crop productivity. Erosion is a long-term integrator of the
effects of human activity on soil; it indicates how management activities (cropping, tillage
6 13
-------
Nutrient
Holding
Capacit
Contaminants
Soil
Productivity
Index
Erosion
I
Microbial
Biomass
or
Processes
Figure 6.5: Components of a soil productivity indicator.
and conservation practices) are affected by environmental factors (topography, rainfall
intensity). Soil displaced from U.S. agricultural lands was estimated to be 7.6 metric tons/ha
in 1987 (George and Choate 1989), considerably greater than estimated rates of soil
formation (Logan 1990, OTA 1982). Accelerated erosion affects ecosystems negatively both
at the source of soil loss and at the point of sediment deposition at some other location in the
watershed (Langdale and Lowrance 1984).
Microbial component. Soil microorganisms are the driving force behind nutrient
transformations in soil and have a major role in soil fertility and ecosystem functioning
(Smith and Paul 1990). Soil microorganisms stabilize the ecosystem by providing a
"biological buffer" against plant pathogens (Baker and Cook 1974), by degrading natural and
anthropogenic waste products, and by serving as a nutrient pool, especially for nitrogen
(Domsch 1985). Soil microbial biomass is considered a sensitive measure of ecosystem
resilience because of the measurable response of this biological parameter to ecosystem
disturbance and to recovery (Smith and Paul 1990). Changes in microbial biomass have been
linked with the presence of toxic soil contaminants (Brookes et al. 1986), tillage practices
6 14
-------
(Carter 1986), and soil salinization (Malik and Azam 1980). The development of rapid
estimation procedures for soil microbial biomass (Jenkinson and Powlson 1976, Anderson and
Domsch 1978) and advances in the interpretation of biomass measurements have made the
biomass estimations useful for comparative ecosystem studies (Smith and Paul 1990) and a
good candidate for the program. There are several specific populations and functional groups
of microorganisms that are being considered as potentially sensitive indicators of
agroecosystem health (Table 6.5).
Table 6.5: Candidate microbial indicators of soil productivity.
Functional groups, specific microbial
populations, or microbial processes which reflect
soil health.
Selected references.
Soil microbial biomass
Nitrification
Populations of Nitrosomonas or Nitrobacter spp.
Vesicular-arbuscular mycorrhizal fungi
Populations of Rhizobium spp.
Soil mesofauna, e.g. soil mites
Degradation rates of soil organic matter
Trophic groups of soil nematodes
Nutrient leaching
Smith & Paul 1990
Domsch 1985, Pimentel and Edwards 1982
Domsch 1985, Pimentel and Edwards 1982
Rabatin and Stinner 1989
Domsch et al. 1983
Werner and Dindal 1990
Domsch et al. 1983
Yeates and Coleman 1982
O'Neill et al. 1977
One of the major challenges with the implementation of soil microbial indicators is in
the spatial and temporal variability of soil microbial populations and processes (Domsch et al.
1983).
Soil microbial biomass is dependent on soil organic matter as an energy source and
the percentage of organic matter in soil is correlated with microbial biomass. Soil organic
matter will be used as a surrogate indicator of biomass while other approaches are
investigated.
6.5.3. Irrigation water quantity and quality
Two aspects of irrigation water quality and quantity will be addressed: 1) the impacts
of water quality and availability on the ecological condition of irrigated agroecosystems and
2) the impacts of agroecosystems on water quality and quantity.
Irrigation for agricultural usage is widespread in the western U.S. and is increasing in
the southeast. In the western U.S., water for agricultural uses comes from surface delivery
through canal systems, groundwater pumping, or direct pumping from natural surface water
sources. In the southeast, surface water pumping is the most prevalent source of irrigation.
6 15
-------
Nationally, water is currently the limiting natural resource for crop production (OTA 1983).
Furthermore, the demand for water for all uses is increasing, and the net value of water for
irrigated agriculture will not be high enough to compete with the values of water used for
industrial or domestic purposes (USDA 1989). The availability and quality of irrigation
water will increasingly determine the sustainability of crop production in the future.
Indicators of irrigation water quantity will be measurements of the amount of water
used, the amount of land irrigated, the availability of irrigation water and the impact of
irrigation on groundwater levels, either directly as a result of pumping for irrigation, or
indirectly as a result of depletion of the recharge capacity of the system.
The quality of irrigation water is also important. Water retained in soil becomes
progressively more saline over time, especially in poorly drained soils. This process is
believed to be responsible for the failure of many irrigation projects throughout history (Carr
1966, OTA 1983). Irrigation water of poor quality can affect the permeability and aeration of
the soil and affect plants through the presence of phytotoxic substances or through the
modification of processes that limit water uptake by the plant (OTA 1983). Surveys by the
USDA before 1985 indicated that 2.9 million of California's 10.1 million irrigated acres
showed signs of salt damage (Maranto 1985).
In both arid and non-arid irrigated ecosystems, the drainwater from irrigated soils
often contains a higher salt burden than the water applied initially, and may contain toxic
concentrations of heavy metals (e.g. cadmium, selenium) and persistent organics (e.g. DDT,
PCB) which are recognized threats to both crop plants and wildlife (OTA 1983, USGS 1986).
In 1983, incidences of mortality, birth defects, and reproductive failures in waterfowl from
selenium toxicity at the Kesterson National Wildlife Refuge (California) suggested the
seriousness of the problem. The Kesterson incident served to elevate the importance of the
issue of irrigation drainwater quality, and provided the impetus for an extensive interagency
investigation within the Department of Interior. An investigation of nine western areas in
1986-87 confirmed that problems associated with agricultural drainwater exist in many areas
of the western U.S. (USGS, 1986-1987).
Data will be collected for the Agroecosystem program on the source and availability
of irrigation water and on the type and concentrations of selected chemicals in the irrigation
water and drainwater.
6.5.4. Density of beneficial insects
Insect predators and parasites play an important ecological role in regulating natural
populations of phytophagous insects (Strong et al. 1984). The ARC will be working with
entomologists to determine the feasibility of developing an indicator of beneficial insect
diversity such as the frequency of occurrence and prevalence of phytophagous insect
predators, pests and pest parasites, and insect pollinators in the agroecosystem. This indicator
would integrate multiple ecological hazards, including climate, the effects of pesticides on
6 16
-------
nontarget organisms, and the effects of other pollutants and stresses on both insects and
plants.
6.5.5. Pest density
The abundance and diversity of genera in weed and plant-parasitic nematode
communities are being considered as measures of the exposure of crop plants to a specific
biotic stress. An increased abundance of certain noxious weeds (e.g., cockle bur, bindweed
and witchweed) can result in higher stress on crop and soil resources than would be present in
a weed-free system. The negative effects of weed competition in the agroecosystem are
usually manifested in reduced crop productivity. Similarly, increased abundance of certain
plant parasitic nematodes (e.g., Meloidogyne spp. and Heterodera spp.) can signal hazards or
risk to crop productivity.
The abundance and diversity of particular genera of weeds and nematodes may reflect
the relative efficacy of pest management practices (e.g., pesticides applied, tillage, rotations).
Herbicides are often applied to attain weed-free fields. Shifts in the virulence or
aggressiveness of plant pathogen populations may demand application of increased amounts
and/or different types of pesticides, or planting different resistant cultivars. Extensive
herbicide and pesticide applications may signal eventual decreased crop productivity. The
application of certain types or amounts of herbicides and nematicides may also pose a hazard
to non-crop and natural resources, especially wildlife, water and soils. The increase in the
amount of rural land devoted to low-management, set-aside programs such as the
Conservation Reserve Program will affect populations of insects, weeds and plant pathogens.
There is concern that set-aside land could become islands of infestation, increasing the
exposure of surrounding cropland to pests (CAST, 1990).
6.5.6. Foliar symptoms
Foliar symptoms or deformation can be caused by biotic (insects or disease) or abiotic
(pollutants, nutrient imbalance or extremes of weather) stresses. In a monitoring program,
foliar symptoms are of greatest value for acute exposures to specific stressors (Laurance and
Grieteur 1984). The use of selected indicator plants for biomonitoring of specific pollutants,
including air pollution, is being addressed by the ARG. Ozone (O3) is one air pollutant of
particular interest, because it is responsible for most of the air pollutant-related losses in crop
yields on both a regional and national scale within North America, where reductions in
growth, biomass and yield of many crops can occur under current ambient concentrations of
O3 (Heck 1990). It is anticipated that an evaluation of the extent of other foliar symptoms,
particularly symptoms of severe disease, insect problems, and nutrient imbalances will be
essential ancillary data for the interpretation of crop productivity.
6 17
-------
6.5.7. Agricultural chemical use
Since the 1940's the intensification of agricultural production in the U.S. has been
extremely rapid in order to feed an increasingly urbanized society (Andow and Davis 1989).
This has resulted in a shift in agricultural practices towards large monocultures, a reduced
genetic base for plants and animals, fewer crop rotations and increased reliance on
agricultural chemicals such as fertilizers and pesticides.
In most U.S. farming operations, synthetic fertilizers are used to replace the large
amount of nutrients removed from agroecosystems in crop biomass. An estimated 11.2
million tons of nitrogen, 4.3 million tons of phosphate, and 5.1 million tons of potash, were
applied to U.S. agroecosystems from June 1989 to July 1990 (USDA-ERS 1990). Pesticides,
which include insecticides, herbicides, fungicides and nematicides, protect crops from pest
damage or weed competition. Pesticide use on major field crops was forecasted to total 470
million pounds of active ingredients from June 1989 to July 1990 (USDA-ERS 1990). The
application of pesticides is not evenly distributed among all crops (Pimentel and Levitan
1986). For example, about 74% of herbicides are applied to two major crops, corn and
soybeans, and 40% of the insecticides are applied to cotton alone (Pimentel and Levitan
1986).
Data on actual chemical use in agroecosystems do not yet exist. The data that are
available are estimated values and do not include all chemical types and crops. The most
widely-used database is the National Pesticide Usage Database developed by Resources for
the Future, Inc. (Gianessi 1985).
The widespread use of agricultural chemicals has increased and stabilized agricultural
yields (Andow and Davis 1989). Because of the effects on crop yields, the type, amount and
rate of pesticide and fertilizer use will be a primary input into the calculation of a crop
productivity index.
A second aspect of agricultural chemical use important to the health of agroecosystems
are the effects of agrichemicals on nontarget sectors of the agroecosystem and on adjacent
ecosystems (Pimentel and Edwards 1982). Whereas many factors such as soil hydrologic
properties, degradation rate, rainfall patterns and application dates are known to have
substantial impacts on the likelihood of chemical uptake by wildlife, none are as important as
application rate. For example, recent Monte Carlo sensitivity tests with the Terrestrial
Ecosystem Exposure Assessment Model (TEEAM) in a simulated exposure scenario on the
American robin feeding predominately on earthworms in a south Georgia peanut field treated
with diazinon, the application rate alone accounted for about 76% of the total variation in
dose (Bird et al, 1989). Although these results are by no means inclusive and may be
extrapolated only for species with habits similar to the robin, they do illustrate the sometimes
underrated importance of application rate in the overall exposure sequence.
6 18
-------
The nontarget ecological effects of agricultural chemical use will be monitored and
assessed by combining use data with that on the status of biomonitor species such as
earthworms, honeybees and other beneficial insects; soil microbial processes; and the
non-point source loading of agricultural chemicals into adjacent terrestrial and aquatic
ecosystems.
6.5.8. Socio-economic factors
There are several issues that the ARG consider integral to the long-term health of
agroecosystems for which the development of appropriate indicators have yet to be addressed.
One of these is socio-economic factors. The ARG recognizes that factors such as population
shifts in rural areas, farm indebtedness, price of production inputs, and federal farm policies
such as commodity programs and conservation reserve programs affect daily management
decisions and, therefore, have major impacts on nearly all components of agroecosystem
sustainability, including land use, crop and soil quality and crop management practices,
(Benbrook 1990, Flora 1990a, 1990b, Madden and Dobbs 1990, Williams 1990).
6.5.9. Genetic diversity
Genetic diversity in domestic crop plants and the preservation of genetic resources is
another issue that increases in importance as national and global developments continue to
push wild crop species into extinction and the reserves of known genes are incorporated into
domestic species from wild progenitors. From a purely anthropocentric viewpoint, the
conservation of the genetic diversity in domestic crop plants and livestock is essential as a
source of raw material for continued genetic improvement. Loss of genetic diversity could
seriously hamper efforts to feed an expanding population under changing environmental
conditions.
6.6. Indicators Addressing Contamination of Natural Resources
A large variety and quantity of contaminants from anthropogenic point and nonpoint
sources are emitted into the environment. Broadly grouped, these contaminants include
synthetic industrial organics, agrichemicals, trace metals, organometallic compounds, and
non-metallic inorganics, including primary and secondary gaseous pollutants. Contamination
of natural resources air, water, soil and biota in agroecosystems can originate both from
within and outside agricultural lands. Major contaminants of concern include trace metals,
pesticides, fertilizers, pathogens and salts. Urban-industrial areas are also potential sources of
synthetic organic and trace metal contaminants as well as precursors of photochemical
pollutants (e.g., ozone). These contaminants are discharged into the air and water, and can
reach agroecosystems in trace amounts.
Contaminants within the broad categories listed above are all potential candidates for a
monitoring and assessment program such as EMAP. However, the large number of
contaminants present within agroecosystems makes the cost of monitoring all potential
6 19
-------
contaminants prohibitive. The selection criteria for choosing contaminants to be monitored
include:
o major anthropogenic discharge into the environment that results in known or suspected
regional distribution,
o persistence and mobility in the environment,
o bioaccumulation in food chains,
o short- or long-term adverse effects on biota, including humans, and
o availability of standard analytical methodology.
Based on these criteria, the following contaminants are being considered important to
monitor: arsenic, cadmium, lead, mercury, chromium, zinc, nickel, selinium, vanadium,
ozone, polychlorinated biphenyls, polycyclic aromatic hydrocarbons, dioxins, furans and
selected pesticides with widespread use such as atrazine and aldicarb.
Assessing the spatial and temporal trends in the distribution and concentration of
contaminants in agroecosystems is a complex undertaking because of:
o the existence of thousands of contaminant sources,
o spatial and temporal variability of source strengths,
o multi-media distribution of contaminants, and
o transformation reactions resulting in
products different from the parent contaminants.
Connell and Miller (1984) state that the objectives of environmental monitoring can be
realized by monitoring the contaminant(s) in different compartments of the environment, and
monitoring the effects of the contaminant(s) on biota (Figure 6.6). The physical and chemical
monitoring of air, water and soil can provide information regarding the spatial and temporal
trends of the contaminants, but monitoring of the ambient environment does not address
issues pertaining to the bioavailability and fate of a contaminant, nor their potential for
biological effects. Given these complexities, it is necessary to monitor both the abiotic and
the biotic component of ecosystems (Figure 6.7).
6.6.1. Non-point Source Pollutant Loading
Nonpoint source loading of agricultural chemicals and sediments from agroecosystems
is a measure of the efficiency of the agroecosystem with respect to resources and inputs, as
6 20
-------
Figure 6.6: Multi-media distribution of contaminants and
potential routes of biotic exposure.
well as a measure of the potential for
contamination of surrounding areas.
Nonpoint source loadings include
agricultural chemicals and their
breakdown product, animal wastes, and
eroded soils and may include genetically
engineered organisms.
Nonpoint source pollution is
characterized by highly variable
loadings, with rainfall and other
environmental characteristics dominating
the timing and magnitude of chemical
transport. Chemicals are exported from
their site of application to nearby
streams and lakes by runoff and
subsurface flow, leaching to
groundwater, drift from aerial and ground application equipment, chemical dust transport and
volatilization to, and deposition from, the atmosphere.
Aquifers underlying agricultural lands receive chemicals by leaching through the root
zone and vadose zone. Nitrates and pesticides have been detected in ground water in many
states from what appears to be normal agricultural use. During 1988, EPA conducted a
survey of drinking-water wells and detected 46 pesticides in the groundwater of 26 states.
Pesticides most frequently found were atrazine and aldicarb (Hileman 1990). Data obtained
from the Agroecosystem Research Strategy Prospectus shows many rivers and streams do not
meet water quality standards because of loadings from agricultural land (U.S. EPA 1988).
Databases on chemical residue monitoring for surface and groundwater quality are available
from STORET (EPA) and WATSTORE (USGS). Comprehensive databases on pesticide
runoff and leaching from specific field sites have been developed by EPA, USDA, USGS,
TVA and chemical companies. Some example databases are reported by Smith et al. 1978,
Johnson and Baker 1984, Ellis et al. 1977.
Irrigation practices are known to enhance leaching of chemicals from soil, including
applied chemicals, naturally occurring salts, selenium and other trace elements. Irrigation
from contaminated water sources can introduce organic chemicals, salts, and nitrates into
agroecosystems. Many of these chemicals are subsequently transported to surface water.
Chemical application in irrigation water raises similar concerns.
Increased awareness of the ecological threats from water contamination in
agroecosystems and adjacent areas have resulted in water quality initiatives (Burkart et al.
1990, Wilson 1987), emphasis on input management (Cox 1984, Odum 1989; the USDA
LISA program), and the design and implementation of best management practices (Humanik
et al. 1984, Pimentel et al. 1989).
6 21
-------
Contaminant Source Contaminants Exposure Indicators Response Indicators
Abiotic Biotic
L Synthetic Organics
Urban-Industrial I .^ Trace Metals
Discharges W~~^^ Organometallics
| Non-Metallic Inorganics
i
iR*
b Pesticides j ^^™
Management B^— — ^^- Trace Metals £
Practices • ^^ Pathogens f
L-HM^^^^^^^A Sediments /
/
Toxic Release Inventories
Transport and Deposition Models
Pesticide and Fertilizer Sales
Pesticide and Fertilizer Use
Municipal Sewage Sludge Use
Stressor Indicators
Tissue H
Concentrations 1
\
Water • j i
•"N— L!
Cll !
Biomarkers and •
Overt Symptoms I
\
iNISM |
Altered 1
Performance •
\
POPULATION 1
Occurence, Abundance,, Reproduction 1
'
COMMUNITY / ECOSYSTEM 1
Stnicture, Composition, Function 1
Figure 6.7: Development of indicators to assess the contamination of natural resources.
One of the main challenges in measuring nonpoint source loadings is measuring the
phenomenon at the time it occurs. In many cases, the insult to the environment is of a pulse
nature. Leaching, for example, is a highly dynamic process that usually involves transport
and transformation of the compound. The location of the compound in time and space is
dependent on transformation rates and transport characteristics of the media. Thus, sampling
at any particular point in time or space leads only to an "instantaneous" measurement that
changes very quickly. As a result, it is unlikely that leaching, runoff, and volatilization can
be accurately monitored in a single index period. However, the ARG will attempt to use
models being developed by the U.S. EPA and other organizations to identify the variables
driving pollution loadings and measure these variables as an estimate of the potential for
nonpoint source loadings on a regional basis (Table 6.6).
Agricultural land use patterns, by modifying the landscape, may influence the
movement of contaminants, for example, strips of remnant vegetation along stream corridors
in agricultural regions may filter mineral nutrients and contaminants that would otherwise
enter the water (Karr and Schlosser 1978; Peterjohn and Correll 1984). The effects of land
6-22
-------
Table 6.6: General site characterization and pesticide model parameters.
Weather records
- precipitation
- evaporation
- max/min air temperature
- relative humidity
- solar radiation
Soil characteristics (by depth)
- identification of soil series
- horizon depths
- infiltration rate (or percolation) for surface
hydraulic conductivity
- water content (moisture release curves)
bulk density
- texture
- porosity
- organic carbon content
- depth to water table
- runoff potential
- pH and temperature
- microbial populations
Pesticide application and other field-determined parameters
- application method
- distribution coefficient of plant-soil application
- transformation rate with depth for soil and foliage
- sorption partition coefficients (by depth)
- volatilization
- plant uptake
- pesticide concentration profile for each sampling time
Crop management
- tillage practices
- other cultural practices
(Adapted from Smith et al. 1990)
use patterns on ecological processes are discussed further under Quality of Agricultural
Landscapes.
6.6.2. Soil contamination
Municipal sludge and industrial or urban waste water are commonly applied to
agricultural soils as an organic amendment and as a waste control strategy. Although an
6 23
-------
active microflora will degrade most potentially harmful contaminants in these wastes, the
safety and desirability of waste application has been controversial because of the potential of
gradual contamination of soils with toxic and persistent contaminants such as trace metals.
The buildup of residues from persistent or frequently applied pesticides is a source of
contamination which has resulted in the banning of several major pesticides, such as, aldicarb
and ethylene dibromide. Atmospheric deposition also contributes to the presence of soil
contaminants. The combustion of fossil fuels and other anthropogenic processes, such as
waste incineration and field burning, has greatly increased the environmental burden of
polyclycic aromatic hydrocarbons in soil in the past 50-100 years due solely to atmospheric
deposition (LaFlamme and Kites 1978, Jones et al. 1989a, 1989b).
Soil contaminants pose a direct risk of toxicity to plants, soil organisms and microbial
functioning, and thereby, influence populations of microbes, interactions among species, and
plant productivity. Some contaminants are taken up by plant material and pose a risk of
accumulation in grazing livestock and in humans. Persistent soil contaminants, because they
remain in the environment for long periods of time, are likely to move with sediments into
the air or water media through erosion processes and pose a threat to aquatic species and
adjacent ecosystems.
6.6.3. Biomonitors
The use of biologically based measurements of environmental quality is referred to as
biomonitoring. Frequently there is a distinction made between the use of organisms as
biomonitors and bioindicators. Biomonitors are organisms that are used to determine the
presence of a contaminant through its incorporation and accumulation in the organism's
tissues; bioindicators are organisms that reveal a biological response that normally is
associated with a particular contaminant (Tingey 1989). A bioindicator becomes a
biomonitor, if the response is used to predict the concentration of the contaminant. For the
purposes of the Agroecosystem program, biomonitors are defined as plants and animals that
accumulate one or more contaminants and/or demonstrate a quantifiable response to a given
contaminant concentration. Accumulators may or may not reveal a measurable biological
response to increased tissue concentrations of contaminants. Indigenous organisms may
reveal a large array of biological responses to contaminants, that range from the biochemical
to the population level of organization. Advantages of using biomonitors include:
o integration of potentially wide fluctuations in contaminant exposure which results in a
more comprehensive assessment of the bioavailability of the contaminant;
o reduced cost compared to sophisticated instrumentation required to monitor air, water
and soil;
o some species occur over broad areas which allows for the development of regional
(and possibly national) networks; and
6 24
-------
o tissues can be archived (dried or frozen) for future analyses.
6.6.3.1. Honeybees and earthworms
Although several organisms could be used as biomonitors in agroecosystems, honey
bees and earthworms, because of their respective roles in pollination and soil processes and
their proven utility as both exposure and response biomonitors, are currently being examined
for contaminant monitoring. Honey bees are effective bioaccumulators of trace metals,
radionuclides, pesticides and industrial organics (Anderson and Wojtas 1986, Bromenshenk
and Preston 1986, Bromenshenk et al. 1985, Morse et al., 1987, Smirle et al. 1984,
Wallwork-Barber et al. 1982). Although various contaminants have been detected in honey,
wax and pollen, the forager bee is the recommended sampling unit for exposure monitoring,
because they are multi-media samplers of air, water, soil and vegetation in an area of
approximately 7 km2, and they often contain the highest concentrations of contaminants
(Bromenshenk 1988, 1989).
Earthworms are being considered as biomonitors of contaminants in agroecosystems
because of their occurrence in the majority of soils, importance as soil organisms, and
position in the food chain. Earthworms make a large contribution to the total biomass of soil
invertebrates, especially in temperate regions, where they also have a significant impact on
soil structure, aeration, organic matter degradation, and fertility (Edwards and Lofty 1977).
Because of their intimate contact with the soil, earthworms have been shown to be ideal
organisms for assessing soil contamination and bioavailability of persistent organics and trace
metals (Callahan 1988, Ebing et al. 1984, Ireland 1979).
Although exposure monitoring with honey bees and earthworms will provide spatial
and temporal patterns for contaminants that are environmentally persistent, many
contaminants are short-lived in the environment, but nevertheless capable of adverse
biological effects. Honey bee colonies may present the best opportunity to conduct
multi-dimensional testing from the biochemical to the population level of organization.
Exposure to some contaminants often results in adverse responses such as dysfunction or
death of individual bees, but only in severe cases does the bee colony perish. Although the
colony is relatively rugged and long-lived, exposure to pollutants is reflected by alterations in
reproduction, population size, life span, mortality, behavior, productivity (e.g., stores of
pollen, nectar, and honey), enzyme induction (e.g., metallothionen), and enzyme depression
(e.g., cholinesterases) (Bromenshenk, J.J. personal communication).
6.6.3.2. Lichens, mosses and ladino clover
Non-vascular plants such as lichens and mosses have been used as biomonitors of
atmospheric pollution for over a century. Lichens are long-lived perrenials without a known
differential sensitivity to oxides of sulfur and various fluorine compounds. Because they
6 25
-------
absorb most of their nutrients (and contaminants) directly from the atmosphere lichens are
ideal organisms to assess exposure to airborne chemicals.
Lichens and mosses have been used as accumulators of trace metals around point
sources (Hawksworth 1973, Pilegaard 1979), and in regional surveys for area source
contamination by trace metals and persistent organics (Carlberg et al. 1983, Glooschenko
1989, Thomas 1979, Thomas 1986, Villeneuve and Holm 1984 ). Although lichen and moss
species are distributed widely, it is not likely that a common species will occur at all
monitoring sites, even within a single region. If multiple lichen species were used for
monitoring, correction factors would have to be developed to ensure comparability among
regions because of the differences in uptake, and possibly in retention, of atmospheric
contaminants (Ross 1990) among lichen species. Monitoring atmospheric contaminants with
lichen transplants (Hawksworth 1973, Strachan and Glooschenko 1988) has been successful
on a local scale, but has not yet been tested on a larger scale.
An inherent problem with the use of lichens or mosses as biomonitors of atmospheric
deposition is that they do not provide information on the absolute magnitude of deposition
(Ross 1990). However, they are a relatively inexpensive means of detecting contamination.
They allow for qualitative comparisons of atmospheric contamination across locations and
provide an early warning of contamination. Epiphytic lichen species have been used in
numerous floristic surveys where changes in lichen species composition or impoverishment of
lichen flora resulted in various indices which reflect ambient air quality, such as the Index of
Atmospheric Purity (Hawksworth 1973). The ARG is currently investigating the merit of
using lichens and/or mosses as passive or active biomonitors on a regional scale.
In addition, current research with ozone-sensitive and resistant ladino clover selections
shows promise of using plants of white clover as a diagnostic, active biomonitor (Heagle,
A.S., personal communication). Ladino clover is a potentially valuable biomonitor species,
because it can be propagated vegetatively to produce genetically identical cuttings of selected
clones and it is widely adapted in most temperate regions.
In order to adequately assess spatial and temporal trends and the fate of particular
contaminants in agroecosystems, it is necessary to monitor both the abiotic and biotic
compartments. The long-term monitoring of chemical contaminants in water and soil, as well
as selected biomonitoring organisms can be viewed as a connecting link between the abiotic
and biotic compartments of the environment, providing information on contaminant exposure,
fate and bioavailability.
6.7. Indicators Addressing the Quality of Agricultural Landscapes
Concepts from the field of landscape ecology, which recognizes the dynamic role of
humans in the landscape (Naveh and Lieberman 1984) and focuses on the interaction of
landscape patterns and ecological processes (Forman and Godron 1986), provides a
framework in which to address the quality of agricultural landscapes. Forman and Godron
6 26
-------
(1986) describe three stages of agricultural development from a landscape perspective: 1)
traditional, 2) combined traditional and modern, and 3) modern with remnants of traditional.
These stages are a subset of the 'cultivated landscape', part of a broader gradient of human
impact on the landscape and the plant and animal species which inhabit it (Godron and
Forman 1983). Traditional agriculture is characterized by scattered, fairly small,
irregularly-shaped fields, whereas modern agriculture is characterized by large homogeneous
fields with small and widely scattered patches of traditional agriculture and remnant
vegetation. The cycling of land over time between agriculture and other uses, as well as the
changing use of land within the agricultural segment, is affected by, and also affects,
ecological and socioeconomic conditions. Any development of sustainable agricultural
systems must consider landscape level processes and the coupling of non-crop and agricultural
components of the landscape (Barrett et al. 1990, Risser 1985, Risser et al. 1983).
Hedgerows and shelterbelts provide an excellent example of the coupling of non-crop
and agricultural components in an agroecosystem (Figure 6.8). Forman and Baudry (1984)
state that hedgerows "exert a major control on many major landscape fluxes. Such fluxes
include animal populations, wind speed, evapotranspiration and soil desiccation, soil erosion
and nutrient runoff, species movement along network lines, and movement of field species
across the network." Other landscape elements within agroecosystems include woodlots (e.g.,
Gottfried 1979, Helliwell 1976, Henderson et al. 1985, Lynch and Whigham 1984, Weaver
and Kellman 1981), wetlands (e.g., Bedford and Preston 1988. Weller 1988), grasslands, and
corridors (MacClintock et al. 1977. Noss and Harris 1986, Schlosser and Karr 1981).
Agricultural landscapes exhibit the characteristics of a shifting mosaic: a pattern of
long-term change with short-term internal spatial fluctuations. Socioeconomic factors often
outweigh ecological factors in short-term, local, land use decisions. In the long-term,
however, patterns of agricultural land use will be dictated by ecological factors (Auclair
1976). Analyses of agroecosystems from a landscape perspective must consider the disparate
temporal and spatial scales at which processes occur: from annual changes in agricultural land
use to decades-long successional processes as land cycles between agriculture and other uses;
from the influence of a hedgerow on an adjacent field, to the cumulative effects of
agricultural systems on the habitat of far-ranging wildlife species (Allen and Starr 1982,
Harris 1988, O'Neill et al. 1986).
6 27
-------
DEPOSITS
SNOW, AIR-uw,,.,
SOIL PARTICLES
AEROSOLS
WINDBREAK
PREDATORS FEEDING
ON PREY IN HELD
CORRIDOR EFFECT FOR
MOVEMENT OF SPECIES
ACROSS LANDSCAPE
MICROCLIMATIC ZONAT1ON IN FIELD, AND EFFECT
ON CROP PRODUCTION
EVAPOTRANSPIRATION AND SOIL DRYING
FLUXES BETWEEN HABITATIONS AND HEDGEROW
DECREASES EROSION
AND NUTRIENT RUNOFF
DEEP ROOTS MAY INCREASE
PERCOLATION OF WATER
SOIL ACCUMULATION
FROM RUNOFF
CONTROL OF
STREAM QUALITY
FLUXES BETWEEN WOODS
AND HEDGEROWS
Figure 6.8: Summary of many major hedgerow functions. Arrows from upper left indicate predominant wind direction. (Forman and Baudry 1984)
-------
6.7.1. Land Use
Changes in land use patterns may have significant effects on agroecosystems. For
example, an increase in acreage planted in tobacco, cotton, or other chemical-intensive crops
might affect water quality in surrounding areas; removal of hedgerows and shelterbelts may
lead to increased erosion of soil by wind and water. Within the agricultural landscape matrix,
we define several broad categories of land use, including the land area 1) under cultivation; 2)
in each crop, in permanent pasture, in set-aside programs, or fallow; 3) in use for managed
animal production acreage, such as feed lots; 4) in farm ponds; 5) in border areas, such as
hedgerows and woodlots; 6) devoted to grassed waterways; and 7) devoted to buildings and
paved areas (Figure 6.9). Annual data collected by NASS, for example, may be used to
follow short-term changes in the use of agricultural land. These changes will most likely
reflect farmer perceptions concerning market opportunities and conservation practices,
changing federal programs, and current socioeconomic conditions. Land use data may be
utilized to analyze the movement of land between broad use categories within the agricultural
segment, as well as changes within each category. The correlation between land uses and
other indicators will also be explored.
Long-term changes in land use patterns over large geographic areas may be addressed
with the aid of databases such as the proposed EMAP Landscape Characterization Database
(LCD) and the USDA Soil Conservation Service's (SCS) National Resource Inventory (NRI).
The frequency at which the EMAP LCD will be updated is presently undetermined. The NRI
has been performed in a fairly consistent fashion in 1982 and 1987; future NRIs are
scheduled for 1992 and every five years thereafter. Because both of these efforts are intended
to monitor changes in the condition of given parcels of land over time, the data may be used
to study the shifting allocation of land to various uses. These long-term trends will reflect
changing ecological conditions, moderated by socioeconomic factors. Changes in soil quality,
water quality, incidence of plant pathogens and insect pests, and climatic conditions are
examples of ecological forces that might cause major long-term changes in agricultural land
use patterns.
6.7.2. Landscape descriptors
The spatial structure of a landscape affects the flow of energy, materials, and
organisms among its components (Forman and Godron 1986; Turner 1989). Agricultural land
use patterns, by modifying the landscape, may exert a major influence on ecological
processes. For example, strips of remnant vegetation along stream corridors in agricultural
regions may serve several beneficial purposes. They keep some animals from damaging the
stream banks; they directly reduce stream bank erosion by intercepting and slowing flowing
water, they filter mineral nutrients and contaminants that would otherwise enter the water; and
6 29
-------
AGRICULTURAL LAND USES
Land Under Cultivation
Specific Crops
Permanent Pasture
Federal Conservation and Crop Reduction Programs
Idle Land
Farm Ponds
Grass Waterways
Border Ares (e.g. woodlots)
Confined Animal Lots
Buildings and Paved Areas
OTHER LAND USES (examples)
Forests
Wetlands
Wildlife Refuges
Parks
Cities
Roads
Figure 6.9: Cycling of land between broad categories of agricultural and other uses.
they act as wildlife habitat and transportation corridors (Fortran and Godron 1986, Karr and
Schlosser 1978, Omernik et al. 1981, Peterjohn and Correll 1984, Schlosser and Karr 1981).
Another effect of agricultural land use patterns on the landscape is habitat
fragmentation. As the gradient of agricultural landscapes from traditional to modern is
traversed, fewer and smaller patches of the native vegetation remain. The changing
composition, number, size, shape, amount of edge, vertical structure, spatial distribution, and
connectivity of these patches are critical determinants of their value as habitat for plant and
animal species (Forman and Baudry 1984, Harris 1988, MacArthur and Wilson 1967, Noss
1983, Noss 1987, Noss and Harris 1986, Roth 1976, Schroeder 1986, Short and Williamson
1986, Yahner 1988).
The proposed EMAP Landscape Characterization Database, which will contain
remotely sensed land use and land cover data stored in a geographic information system
(GIS), will provide a unique opportunity to evaluate spatial patterns. The use of GIS
technology permits data to be stored, displayed, and analyzed in its spatial context. Thus, a
6 30
-------
description not only of the amount and type of agricultural land can be developed, but also of
how it relates to other elements of the landscape. By incorporating data from the other
indicators described in this chapter, as well as data from other sources, into the GIS database,
the relationships between various agricultural processes and the pattern of the landscape may
be further explored. For example, landscape heterogeneity may affect soil development, soil
erosion, crop-pest interactions, and the spread of diseases (Barrett et al. 1990). Techniques
based on information theory (Shannon and Weaver 1949) may be used to quantify the
heterogeneity of a landscape at varying spatial scales (Foreman and Godron 1986, Barrett et
al. 1990). Various measures of spatial pattern have been applied in landscape ecology,
including diversity, dominance, patchiness, nearest neighbor analysis and contagion (Turner
1989). Measures generated from the techniques of fractal geometry (Lam 1990, Mandelbrot
1977) may also be used. The fractal dimension of agricultural regions tends to be inversely
related to the intensity of cultivation (DeCola 1989). Fractal dimension might be used to
determine the spatial scale at which processes are occurring (Krummel et al. 1987), and to
describe the relative amounts of patch-edge to patch-area in order to indicate potential
sustainability of an area as wildlife habitat.
A suite of landscape descriptors can be developed (Table 6.7) that describe habitat
characteristics important to various plant and animal species. A comparison of these
measures with known habitat requirements of wildlife (e.g., birds, small mammals, large
mammals) will provide information about the ability of the habitat to support them. The
abundance of species sensitive to fragmentation has been correlated with the size and shape of
remnant woodlots (Freemark and Merriam 1986, Gottfried 1979, Lynch and Whigham 1984,
Temple 1986). Several measures of size, density, edge, isolation and connectivity have been
used to describe the pattern of woodlots (Bowen and Burgess 1981) and could also be applied
to descriptions of wetland, grassland and other patches in agroecosystems. The amount of
edge relative to area of patches may be compared using a Dissection Index adapted by Patton
(1975) and further defined for an entire landscape (Bowen and Burgess 1981). The spatial
relationship between patches, which is an important determinant of the ability of species to
move from patch to patch, can be described by measures of isolation, connectivity,
accessibility and contagion (Bowen and Burgess 1981; Forman and Godron 1986; Turner
1989). The vertical complexity of vegetation can be described using the Habitat Linear
Classification System (Short 1990) or the Habitat Layer Index (Short 1989, Short and
Williamson 1986, USDA 1989).
Some combination of these or other measures will be used to describe the agricultural
landscape and to explore correlations between land use and landscape patterns and other
indicators. Landscape ecology is a relatively young discipline, and a great deal of basic
research into the effect of landscape pattern on ecological processes remains to be completed
(Rissen et al. 1983, Turner 1989). It is clear, however, that sustainable agricultural systems
must be managed to coexist with the ecological processes of the larger landscapes in which
they are developed.
6-31
-------
Table 6.7: Landscape descriptors currently under consideration for use in monitoring
agroecosystems.
Measure
Fractal dimension
Diversity indices
Nearest neighbor analysis
Contagion index
Dissection index
Habitat Layer Index
Describes
Broad-scale measure of pattern
Spatial complexity
Diversity of landscape
Fragmentation
dumpiness
Edge-to-area relationship
Vertical complexity of habitat
6.8, Integration of Indicator Data
The integration of indicator data into an overall assessment of agroecosystem health is
a challenging and complex task. The integration must be agroecological and multifaceted.
Agricultural productivity and preservation of non-crop and natural resources must be balanced
in a general assessment of agroecosystem health.
From an agricultural perspective, system health can be evaluated on the basis of
productivity and efficiency of production. These are seemingly single and straight-forward
measures; however, government policies and incentive program have large influences on both
measures. Socioeconomic factors must be an integral part of any interpretation of
agroecosystem health.
From an ecological perspective, system health can be evaluated on productivity as well
as on the preservation and status of natural resources (air, soil, water) and the health of
non-crop resources. The difficulty in judging ecological health of agroecosystems is
compounded by a lack of definitive standards or pristine sites against which comparisons may
be made. The perspectives of the new focus on sustainable agriculture in agriculture research
and policy may provide useful direction.
The integration of indicators to provide an overall, meaningful environmental index of
agroecosystem health is a long-term research goal of the ARG. An environmental index
refers to a mathematical aggregation of more than one indicator into a single number (index)
that retains environmental meaning. An index has both strengths and weaknesses. If properly
generated, an index can provide a relative ranking (healthy to unhealthy) of agroecosystems
over spatial and temporal scales. The weakness of an index is that useful information is lost.
6 32
-------
However, in EMAP, the raw data and algorithms used to generate the index will be
retrievable.
6.9. Conclusions
The indicators currently under consideration for monitoring agroecosystems must be
scientifically acceptable, statistically valid and compatible with the overall EMAP monitoring
design. Information derived from the response and exposure indicators may be correlative
with apparent stresses (e.g. pollutants), but should not and can not be taken to indicate causal
relationships between specific stresses and specific responses in the ecosystem. Rather, the
correlations between system perturbation (either positive or negative) and changes in response
or exposure indicators should be used as guides for further research activities designed to
establish causal relationships. However, correlative data will be used to suggest associations
for use in risk assessment strategies.
To provide assessments of agroecosystem condition, the indicators must be selected
and implemented so that the relatively simple measures provided by individual indicators can
be integrated with confidence into a multivariate index, or several indices, of system health.
These indices must be sensitive to changes in system condition due to imposition or removal
of stresses. The interpretation of system condition provided must be balanced between
commodity production and the preservation of natural resources and noncrop communities.
It is probably not possible or economically feasible to develop a suite of indicators
that will be all-inclusive for assessing the condition of agroecosystems. Compromises must
inevitably be made. Indicators that are surrogates of important ecosystem functions or
structure, or which can serve to represent several important ecosystem components, must be
chosen to fit within the monitoring and assessment goals of EMAP and, realistically, to fit
within a fixed budget.
The decisions that are made concerning indicator selection and implementation for
monitoring the condition of agroecosystems must be based on the best scientific information
available, including research data from studies designed to specifically evaluate indicator
efficacy, efficiency and interpretability. The process of indicator development is not a rapid
one. Rather, it is an iterative and continuing process. Some of the information needed to
evaluate potential indicators is available in the scientific literature; other critical data must be
derived from empirical studies. Implementation of the resulting suite of indicators will
provide information vital to the future of agriculture and to the preservation of agroecosystem
resources in the United States.
6 33
-------
7. Logistics
7.1. Logistics Implementation Components
Implementation of the Agroecosystem program
will require detailed, comprehensive logistics planning.
Logistics considerations include coordination and
oversight of all support and data collection activities
(Figure 7.1). Overlooking or ignoring apparently
minor issues or details could jeopardize the success of
the program. In accordance with EPA policy (US EPA
1990), a logistics plan addressing these issues must be
developed prior to implementing field activities to
assure that these activities are consistent with program
goals.
Staffing
Training
Sampling Schedule
Reconnaissance
Site Access
Field Operations
Sample Handling and Transport
Communications
Laboratory Operations
Waste Disposal
Safety Plan
Procurement and Inventory Control
Figure 7.1: Issues to be addressed in the
Logistics Plan prior to implementation of
agroecosystem monitoring.
As the first actual field monitoring operation
the one-state pilot is not scheduled until 1992, a
comprehensive logistics plan is not presented at this
time. Rather, a brief discussion of issues is provided
including how the Agroecosystem Resource Group (ARG) and the USDA National
Agricultural Statistics Service (NASS) will work cooperatively to develop the overall
program. Prior to initiation of the pilot, a detailed logistics plan will be presented. The plan
will make full use of current NASS logistics in areas of mutual cooperation.
7.2. Logistics Issues
The current intent of the ARG is to develop an inter-agency cooperative agreement
with USDA / NASS under which NASS enumerators will collect data required for the
Agroecosystem indicators. These enumerators, operating within the NASS organization, will
use procedures selected and developed jointly by the ARG and NASS. From the standpoint
of logistics, working with NASS has several benefits. Based on the integrity and reliability of
their personnel, NASS has developed a relationship over time with the agricultural community
which will greatly facilitate the collection of data. Additionally, NASS has a fully developed
infrastructure for the collection of agricultural data, including well-developed logistical
procedures and strict quality controls. Use of this infrastructure greatly reduces the
expenditure of resources that would be needed for the ARG to develop similar procedures.
The Agroecosystem pilot program (Chapter 11) is designed to define more completely
the interactions between NASS and the ARG and to develop and refine logistics procedures.
A brief summary of the current perception of logistics procedures follows; the ARG
recognizes that a great deal of effort will be required to fully develop this aspect of the
program.
7 1
-------
7.2.1. Staffing
NASS enumerators will collect data for the ARG. Enumerators are employed in each
state on a part-time basis through the National Association of State Departments of
Agriculture (NASDA). Most enumerators are local farmers, members of a farm family,
retired rural residents, or other persons with an interest in agriculture. They are located
throughout each state which expedites data collection and minimizes travel time and expenses.
NASS will be responsible for hiring and supervising enumerators, as well as payroll and other
administrative functions.
The ARG will maintain a scientific and statistical staff for the analysis and synthesis
of the information collected. Responsibility for the development of indicators and indices of
agroecosystem health will reside with the ARG staff.
7.2.2. Training
NASS enumerators are part-time employees with a wide range of educational
backgrounds. Prior to participating in any data collection efforts, they undergo an intensive
training program for sample and data collection methods. Training of enumerators will be a
joint responsibility of the ARG and NASS. An enumerators' manual for the collection of
EMAP data will be developed by the ARG with input from NASS. The manual will include
information on the background and objectives of the Agroecosystem program and will define
specific interview and sampling procedures.
NASS will be responsible for enumerator training for survey questionnaires. A three-
day training school is held in mid-May for the June Enumerative Survey. Training includes
background information and a question-by-question review of the survey instrument. Group
and one-on-one practice exercises are conducted to strengthen the enumerators' knowledge of
the questions. Members of the ARG will be responsible for training enumerators in irrigation
and ground water sampling techniques during these sessions.
An additional one-day training session will be held in November for the December
survey. This session will cover the collection of post-harvest data and soil samples. NASS
will be responsible for training enumerators for the collection of post-harvest data; the ARG
will train enumerators in soil sampling techniques.
7.2.3. Sampling Schedule
Data will be collected by NASS enumerators for the ARG during the June
Enumerative Survey, and during additional field visits in July and December. NASS will be
responsible for the development of detailed sampling schedules within each survey or
sampling period.
1 2
-------
7.2.4. Site Access
NASS has an excellent record of national experience and long term contact with the
agricultural community at the national, state, and local levels. Site access is rarely a problem.
During survey periods, enumerators attempt to locate and interview all farm owners or
operators in sampled segments. During the interview, enumerators will solicit permission to
collect soil and water samples.
7.2.5. Procurement and Inventory Control
The ARG will provide equipment and supplies for the collection of soil and water
samples and for the transportation of samples to contract analytical laboratories. NASS will
provide all survey instruments and supplies associated specifically with the questionnaire.
7.2.6. Sample Handling and Transport
Procedures for labeling, handling and tracking samples will be developed by the ARG
with input from NASS. The ARG will coordinate the delivery and processing of soil and
water samples. The NASS enumerators collecting the samples will be responsible for the
handling and transportation of samples to an agreed-upon destination.
7.2.7. Field Operations
NASS will be responsible for all data and sample collection activities during both the
June Enumerative Survey and the December survey. The ARG will be responsible for all field
activities involved in the development of new indicators during the initial stages of testing.
7.3. Review of Logistics
During the pilot project, regional demonstration, and the initial phases of the national
implementation, semi-annual reviews of logistics plans and procedures will be conducted.
Members of the ARG, NASS, and the EMAP Technical Coordinator for Logistics will
participate in each review. The purpose of the reviews will be to discuss problems associated
with the monitoring program and procedures for their resolution, and to re-examine all phases
of the logistics plan.
7 3
-------
8. Analytical Considerations
Analytical considerations are defined here as analytical methodology (analysis of
samples, in the laboratory or field), data quality objectives, ancillary data, personnel training,
and communication between the analytical arm of the Agroecosystem program and other
functional groups of EMAP.
8.1. Analytical Methods and Coordination of Methods within EMAP
To ensure that data are usable to the ARG and other resource groups of EMAP, it is
essential that the selection of analytical methods be coordinated between and among resource
groups. Comparable methods will be agreed upon, unless there are important reasons to
deviate. A thorough examination of each analytical technique will be conducted by each
ecosystem resource group and a comprehensive analytical methods manual will be prepared
for the entire EMAP program.
Communication must be maintained between the laboratory or field personnel
conducting analyses and other EMAP personnel. If, for example, it is necessary to analyze a
sample within three days of collection, it will be necessary for the sample collectors to know
this. Similarly, it will be necessary for appropriate supervisors to communicate with the
shippers and laboratory personnel to ensure that samples can be analyzed within the
appropriate time frame. Some primary considerations include, but are not limited to:
o detection limits
o sample handling, preservation and storage procedures by type and analytes of interest
o holding times
o sample containers, sources and cleaning requirements
o calibration and standardization requirements
Preliminary proposed analytical methods presented in Table 8.1 have been selected
based on previous monitoring experience. The summary table outlines the methods, contract
required detection limits, and required support as known at this time. The table will be
revised as more information is collected. In some cases, a general type of method (e.g.,
titration or ion electrode) is identified where a specific method is not known. A major ARG
activity in 1991 will be to identify specific analytical methods to be tested and to identify
appropriate laboratories for the analyses. The field pilot in 1992 (see Chapter 11) will be
used to compare and evaluate the methods. It may be possible to use field or laboratory
screening procedures to reduce the need for more expensive analyses if the limitations of the
technique are known and the data will meet the data quality objectives (DQOs).
8 1
-------
Table 8.1 Initial list of analytical methods for soil, water, and biological samples.
INDICATOR
1. Soil quality
a. Soil Fertility
pH
% Organic Matter
Total carbon
Specific conductance
Cation exchange
capacity
NO3-N
Phosphorus
(extractable)
Sulfates (H2)
(extractable)
exchangeable acidity
Total S (%)
Bulk Density
Coarse Fragment
Texture (% silt,
sand and clay)
b. Metals
Cd
Pb
Cr
Ni
As
Se
Zn
V
Hg
METHOD
Electrode w/meter
Calculated from Organic C
Oxidation/IR or TC
Meter
FIA or Titration
Ion chomatography
Titration
Combustion/IR
Core/gravimetric
Gravimetric
Sieve and pipet
HNO3 and H202 digestion -
ICP (inductively coupled
plasma spectroscopy)
Acid digest - potassium
permanganate and
potassium persulfate
CONTRACT REQUIRED
DETECTION LIMIT
0.01 pH unit
N/A
0.01 wt %
0.15 - 0.45 mg/1
0.5 mg s/kg
0.0005 meg
0.001 % wt
N/A
N/A
0.005 mg/1
0.02 mg/1
0.002 mg/1
0.002 mg/1
0.08 mg/1
0.001 mg/1
0.01 mg/1
0.05 mg/1
REQUIRED SUPPORT.^
(approximate S/Sample)
Field Lab
560
All from same
site with
multiple
samples.
$10
N/A
$-5
S10
S10
S5
510
$10
$10
$5
$8
.$30
$30
$60/sample
8 2
-------
Table 8.1 (Continued)
INDICATOR
2. Biomonitors
a. Foragers Bees
b. Earthworms
c. Lichens/Mosses
Pb
Cd
As
F
PCB
HCB
HCH
Atrazine
3. Water Quality
a. Irrigation
Electrical
Conductivity
pH
Na, Ca, Mg, SAR
Class, boron,
bicarbonate, chloride,
sodium carbonate, SO4
b. Nonpoint source loading
to surface or
groundwater
Inorganic
Nitrate
PH
Organic
atrazine
aldicarb
METHOD
ICP/MS
ICP/MS
ICP/MS
Fluoride specific electrode
GC with Electron Capture
GC with Electron Capture
GC with Electron Capture
HPLC
Electrode
Electrode
ICP
Tritration
ICP
Auto-analyzer
Electrode
Gas chromatograph with
Hall detector
Gas chromatograph
CONTRACT REQUIRED
DETECTION LIMIT
0.005 mg/1
0.005 mg/1
0.08 mg/1
0.1 ppb
1.0 ppb
1.0 ppb
1.0 ppb
0.5 ppm
Ca, B 0.07 mg/1
Mg 0.05 mg/1
SO4 0.5 mg/1
0.05 ppm
0.01 unit
low ppb
low ppb
REQUIRED SUPPORT.^
(approximate S/Sample)
Field Lab
45
45
45
45
45
45
45
45
45
90
$40 all
100
150
100
100
100
20
20
20
20
150.2
150^
Assumes labor cost for sample collection at $45/hour.
Analysis costs depend on pesticide type, number of samples, and compatibility for multiple pesticide analysis. The total
cost for analysis of both atrazine and aldicarb per sample would be about $150.00. Pesticides such as atrazine and
paraquat would not be compatible, separate analysis required.
8 3
-------
8.2. Data Quality Objectives
Data quality objectives must be firmly established in order to select the appropriate
analytical techniques. The required levels of precision, accuracy and the detection limits will
be evaluated in the pilot study based on the anticipated use of the data. Detection limits and
levels of precision and accuracy will be evaluated in relation to field sampling errors. It is
not cost effective to strive for the best analytical data when field sampling errors are high, or
the values of concern are orders of magnitude above the detection limits. These issues will
be assessed during the pilot.
8.3. Ancillary Data
It will be necessary to clearly identify the ancillary data needed for data interpretation.
For example, carbonate clays, pH, organic matter and 15-bar water data are necessary for the
interpretation of the particle size analysis on soil samples. Similarly, electrical conductivity
measurements run in the field on water should also include measurements of water
temperature. Supplementary data that will aid in interpretations will be identified for each
analyte and each environmental media. A list of ancillary data to be used in the interpretation
of indicator values is presented in Table 6.4.
8.4. Laboratory and Field Support
For each required analysis, a qualified laboratory that participates in EPA-approved
analytical testing will be selected. Adequate training will be provided for NASS and EMAP
personnel collecting samples in the field. These personnel will have a clear understanding of
how to take and record data in the field. The pilot program will evaluate the quality of the
training program to identify areas that need improvement; appropriate changes will be
incorporated into subsequent pilot and demonstration projects.
8 4
-------
9. Quality Assurance Program
Decision makers, the public, and other users of EMAP data must have a high degree
of confidence in the data and statistics generated by the program. The purpose of the quality
assurance program is to ensure that the data will yield sound and unbiased conclusions related
to the principal hypotheses being addressed. This chapter describes the quality assurance
(QA) program, an integrated program for assuring the reliability of measurements, that will be
implemented for the Agroecosystem component of EMAP.
9.1. Quality Assurance Policies
9.1.1 EPA Policies And Programs
The QA policy of the EPA requires every monitoring project to have a written and
approved Quality Assurance Project Plan (Stanley and Verner, 1983). The goal of the policy
is to ensure that all decisions made by EPA personnel are supported by a sound data base.
An important step toward achieving this goal is to design quality assurance procedures as an
integral part of all data collection and processing activities. QA consists of multiple steps
that are taken to ensure that all data are suitable for the user's intended purpose. In the
Quality Assurance Program Plan several key components of QA, including data quality
objectives, standard operating procedures, QA project plans, audits, QA annual reports, and
work plans are defined. (Figure 9.1)
9.1.2. Quality Control at MASS
Because the Agroecosystem program is being developed as a cooperate effort between
the EPA and USDA/NASS, the ARG will take advantage of QA procedures already employed
by NASS. NASS views quality control as the process to eliminate as many survey errors as
possible. To limit survey errors, every survey process must be associated with some type of
quality control procedure. The ARG group intends to use all of NASS's established quality
control procedures in each survey process. The major survey processes in the Agroecosystem
pilot program amenable to quality control considerations include:
Area sampling frame Survey software
Construction Training schools
Maintenance Survey management
Sampling Questionnaire handling/processing
Survey specifications Manual data review and coding
Questionnaire design Data edit and review
Preparation of manuals Summarization
Interviewers Post survey evaluations
Supervisors Survey research
Editing
9 1
-------
Data Quality
Objectives
Standard
Operating
Procedures
Annual Reports
and Workplans
Project Plans
Audits
Corrective Action
Memoranda
Figure 9.1: Key components of the Quality Assurance Program Plan.
9.2. Total Quality Management
Total Quality Management (TQM) is a process of continuous improvement and
innovation led by top-level management in which management philosophy, planning, and
operating methodology focus on the needs of the client. TQM is a new philosophy of
operation which concentrates on maximizing value to the client by continuously improving
the systems by which all work is performed (Figure 9.2). For EMAP, data are the products
and its users our clients.
TQM is based on "doing the right thing, the right way, the first time". It is
participatory in nature and aimed at achieving a total commitment to quality. This
commitment must span all levels of the program, from management through to the most basic
implementation and support activities. This empowers the individual closest to each process
to focus on continuous improvement and to take corrective actions when needed (Figure 9.3).
All aspects of the product production process may be viewed as a series of inputs and outputs
9 2
-------
Quality
Assurance
QA Program Plan
Data Quality Objective
QA Project Plan
Audits
Reports
Total
Quality
Management
Establish principles and
guidance for EMAP personnel.
Establish customer needs.
Establish the process for
satisfying customer needs.
Evaluate the process to improve
it and to judge its ability to satisfy
customer needs.
Product to meet customer needs.
Figure 9.2: Direct linkages between quality assurance and total quality management.
where quality is directly affected by the TQM philosophy (Figure 9.4).
9.3. Quality Assurance Coordination Roles
Direct responsibility for quality assurance lies with the Technical Director (TD) as the
manager and coordinator of several activities. Quality assurance functions for which the TD
is responsible include:
o Provide adequate resources
o Direct development of Data Quality Objectives
o Direct development of QA Project Plans
o Implement Total Quality Management
o Ensure adequate training of personnel
o Ensure the auditing process
o Support the QA program
9 3
-------
Culture
Vision
Values
Goals
Objectives
Management
Output
Customer Focus
Training
Accountability
Resources
Priorities
People Sensitive
Teamwork
Quality Improvement
Quality Design
Rewards & Recognition
Leadership
Employee Empowerment
Involvement
Quality Data
Quality Reports
Quality Studies
Figure 9.3: Overview of the essential components for a successful implementation of Total Quality Management.
The ARG Quality Assurance Officer (QAO) will serve as an advisor to the Technical Director
in matters relating to QA programs, and is responsible for developing the QA policies for the
program. For the Agroecosystem Resource Group, the QAO will also interact with
USDA/NASS personnel in matters relating to quality control and assurance. The QAO also
conducts mandated QA assessments (e.g., audits) and prepares documents (e.g., QA Project
Plans). The QAO also acts as an advisor to the ARG and has responsibility for the overall
Agroecosystem QA program and its implementation.
9.4. Quality Assurance Objectives
The following are specific QA program objectives applicable to the Agroecosystem
program. In part, these objectives will be fulfilled through well established procedures used
by USDA/NASS.
9 4
-------
Sample
Collection
PROCESS
Transportation
& Storage
PROCESS
Output
(Sample at Lab)
Input
(Lab Sample)
Sample
Preparation
PROCESS
Output
(Sample Extract)
_ Analysis
Input • ^ Output
(Sample Extract) | ^ (Data)
PROCESS
Figure 9.4: Sequences of inputs and outputs in which overall quality can be enhanced by Total Quality Management.
Comparability of data
Common data generated by different EMAP Resource Groups must be comparable.
Data describing the same measurement activity must also be comparable over time. Data
generated within and across ecosystems must be stored in data bases which are compatible.
Documentation
The scope and duration of EMAP will require re-examination of data, methods, and
conclusions as the program matures. Thorough documentation is vital in order to follow the
trail of information, alternatives, decisions, and conclusions which 1) form the basis for new
activities or 2) require the revision of previous conclusions based on new methods or
information.
Validation
Data will be evaluated systematically to assure that they are adequate for the users'
needs. Validation is accomplished by comparing data at each level of processing, (raw,
composited, enhanced) against established data quality criteria.
9 5
-------
Verification
Verification will confirm that measurements, processes, and products are accurate.
Verification is accomplished through a program of audits of technical and 'management
systems, as well as technical and management outputs.
9.5. Data Quality Objectives (DQOs)
Data quality refers to the magnitude of error and the probability of its occurrence in a
particular data set. DQOs are statements of the quality of data needed for a particular
purpose and reflect the level of uncertainty a decision maker is willing to accept. They are
definitive, quantitative or qualitative statements developed jointly by data users
(environmental program managers and decision makers) and data producers (scientists and
technicians). They should define acceptable levels of accuracy, precision, representativeness,
comparability, and completeness in measurement data and provide estimates of non-
measurement error such as naturally-occurring spatial and temporal variability. The Quality
Assurance and Management Staff (QAMS) of EPA has developed an approach for
establishing data quality objectives (DQOs). The DQO process is an iterative approach,
balancing costs against uncertainty to achieve a desired or acceptable level of quality (Figure
9.5). This information is used to allocate resources in order to generate data of sufficient
quality to support decisions or answer specific questions.
The Agroecosystem Resource Group is committed to the DQO process as a means of
optimizing the allocation of limited resources and assuring that data collected provide the
information needed to meet program goals. The ARG intends to utilize the DQO process
starting with the planning stage and continuing through implementation.
9.5.1. The Data Quality Objective Process
The first stage of the DQO generation process (Fig. 9.6) involves defining the major
issues of concern to the users. The focus is on information needs, resource and time
constraints, and the consequences of Type I and Type II errors. Type I errors (so-called false
negative) indicate no adverse ecological effects when such effects actually exist. Type II
errors (so-called false positive) indicate the presence of adverse ecological effects when none
actually exist.
The second stage of the process involves a definition of the information needed to
address the issues identified in Stage 1. This includes development of pertinent questions,
definition of populations of interest, identification of specific design constraints, and
examination of existing data.
The third stage of the process involves a determination of a scientific approach to data
collection and the data quality requirements for the approach. This includes various
approaches to collecting the data, levels of data quality required to meet the constraints in
9 6
-------
Increasing
Increasing
Data
Quality
Objectives
t
Decreasing
Decreasing
Uncertainty
Costs
Figure 9.5: Representation of trade-offs in the process of developing data quality objectives (DQOs).
Stage 2, and identification of research activities needed to meet the requirements of Stages 1
and 2.
Throughout the process, there is a continuous feedback and communication at all
levels which includes decision makers, scientists involved in identifying specific information
needs, and those individuals who will collect the necessary data (Fig. 9.7).
9.5.2. Data Quality Hierarchy
Data quality exists at several levels (Fig. 9.8). Measurement quality objectives include
attributes such as precision, accuracy, representativeness, comparability, and completeness
associated with the measurement of environmental variables. At the next level, data quality
includes the uncertainty associated with the procedures used to assimilate this measurement
data into an assessment (i.e., provide information from the data). For the ARG, this can be
perceived as "indicator quality objectives" because the indicators are those tools used to
provide information about specific aspects of ecosystem condition. Factors include
9 7
-------
Stage 1: Problems of Concern
o What is the purpose of the data?
o What are the resources and time constraints?
o What are the consequences of Type I and II errors?
Stage 2: Information Needs
o What is the population of interest?
o What level of confidence must attend results?
o Does pertinent, usable data currently exist?
o What new data are needed?
Stage 3: Scientific Approach
o What approaches to data collection are available?
o Which approaches provide data quality commensurate with
Stage 2 requirements?
o What R & D activities are needed to meet Stage 2 requirements?
Figure 9.6: The multi-stage iterative process for the development of Data Quality Objectives (DQO).
measurement data quality, sampling design, and statistical data analysis. At the next level,
these indicators are aggregated into an overall assessment of the system condition. Here,
uncertainty associated with each indicator is an additional factor in the indicator total
uncertainty. This uncertainty will then be compared to ecosystem-level quality objectives.
Finally, EMAP hopes to integrate information for across ecosystems quality objectives in
order to make regional-scale assessments of ecological condition. Here, the uncertainty
associated with each ecosystem component must be included in the overall uncertainty of the
assessment. The DQO process requires that sources of variability be identified at each level
and that all relevant sources be considered in generating estimates of uncertainty at any level
of the hierarchy.
9.5.3. The Role of DQOs in EMAP-Agroecosystems
The need for environmental information has been stated in very qualitative terms
during Stage 1 (Fig. 9.6). Once the tools necessary to measure and assess the condition of
agroecosystems are developed, quantitative requirements for data quality can be formulated.
9 8
-------
Ill iT f 'i' 1 T> L -^
^_ r, i ,i r> L
Problem Solving •
Change the Process I
Collect Data ^
1
Evaluate Performance |
INu i VPS
Npw Pr^^ss Superior • — « i T,T
1JT1 o 1 ^" Adopt N
Continuous
Improvement
Process
;w Process |
Figure 9.7: Continuous feedback and improvement process for the refinement of Data Quality Objectives.
The program is now in Stage 2, with extensive feedback to Stage 1. During Stage 2 a series
of indicators are being developed that, in aggregate, allow for overall assessment of
ecosystem condition. Quantitative "logic statements" must be developed describing the data
to be collected for each indicator and the way in which these data will be used to assess
agroecosystem condition. These statements should relate each indicator to assessment
endpoints of societal concern so that the ramifications of changes in system condition can be
understood and appreciated by a variety of user groups.
In addition to developing these logic statements, a series of error constraints must be
developed that identify known sources of error or uncertainty associated with each indicator.
Where possible, estimates should be provided for each source of uncertainty. In this way,
factors that contribute significantly to the overall variability of the indicator are identified and
the efficacy of various means to reduce error may be evaluated. For example, if spatial
variability is the major factor in the overall uncertainty associated with an indicator and
measurement error is small in comparison, it may be judicious to use a less precise and less
costly method of analysis and invest more resources to increase the sampling density within a
9 9
-------
Societal / Environmental
Endpoints
Cross-Ecosystem Assessment
Aggregate Indicators
\
Indicator QO
Indicators of
Ecological
Interest
Measurement QO
Baseline
Data
Figure 9.8: Levels of increasing complexity (bottom-to-top) within the Agroecosystem to which Quality Objectives
(QO) can be applied.
region to reduce the overall uncertainty in the data. (Fig. 9.5). NASS has identified five
types of survey errors which must be addressed by the NASS QA program. These errors are
summarized in Table 9.1.
9.6. Quality Assurance Project Plans
Quality Assurance Project Plans for the Agroecosystem Resource Group will outline
the policies, organization, objectives, and functional activities for specific projects. Each plan
will also describe the QA and QC activities and measures that will be implemented to ensure
that the databases will meet or exceed all criteria for data quality established for the project.
The Project Plans serve as guidebooks to all personnel by describing each individual's role
and identifying specific procedures which will be followed to generate and evaluate data. The
Project Plans will be revised frequently to reflect changes in procedures as they are optimized
over time; used as a formal means of communicating with other program areas; and serve as
a standard for technical audits. All project personnel should be familiar with the policies and
9 10
-------
objectives outlined in the pertinent plan to assure proper interactions among the various data
acquisition and management components.
Table 9.1 Five types of survey errors addressed in the NASS QA program.
Error Type
Specification
Coverage
Response
Non-Response
Processing
Cause
Data Specification is inadequate or inconsistent
with objectives
Sample frame does not probably represent
proportion being sampled
1) Failure of respondent to report correct value
2) Failure of enumerator to record correct
value
3) Failure of instrument to measure value
correctly
Failure to collect complete information on all
sample units
Faulty implementation of a correctly planned
survey
The information to be incorporated in a Quality Assurance Project Plan is summarized
in Table 9.2. Further details on the operation and preparation of a QA project plan can be
found in the U.S. EPA Publication QAMS/005/80.
There are several QA-related requirements during the sequential processes comprising
an environmentally-related measurement, from sample collection through to the processed
data. (Fig. 9.9). By selection of an appropriate suite of QC tools, it is possible to isolate the
error contribution and set control criteria for each of these sequential processes based upon
the overall measurement quality objectives derived from the hierarchical DQO process. This
approach is essential for providing diagnostic information for out-of-control conditions so that
timely corrective action can be taken.
9.7. Standard Operating Procedures
Any activity that is to be performed repeatedly over time is a candidate for a Standard
Operating Procedure (SOP). Environmental monitoring SOPs are devised for sampling,
analysis, data management, QA, QC, reporting activities, accounting, project finance and
contracts, and in the analysis and integration phases of the project. The use of written SOPs
helps to assure 1) consistency in planning, implementation, and analysis activities and 2)
consistency over time and among personnel for routine activities within an organizational unit
(but not among units).
9 11
-------
Table 9.2. Description of the required subject areas of a Quality Assurance Project Plan for EMAP"
Subject Area
Description
Project Description
Project Organization
and Responsibility
Quality Assurance
Objectives for
Measurement Data
Sampling Procedures and
Sample Handling
Sample Custody,
Transportation,
and Storage
Calibration Procedures and
Frequency
Experimental Design and
Analytical Procedures
Data Reduction, Validation,
and Reporting
Internal Quality Control
Checks and Frequency
Performance and Systems
and Frequency
A brief introduction containing relevant background information. This section should also contain a
general statement of project goals and a discussion of how these goals will fulfill the EPA's
objectives for the project.
Summarize the overall project organization and the responsibilities of cooperating organizations.
A figure illustrating the organizational structure is usually included.
Specify the intended end use of the data, questions to be answered or decisions to be made based
on analyses of the data. This section should also spell out the Data Quality Objectives for five
aspects of data quality: representativeness, completeness, comparability, accuracy, and precision
for each indicator.
Provide specific guidelines and protocols regarding preservation, holding times, labeling, and
collection of samples for each major indicator. A description of the site selection rationale can also
be included.
Sample tracking is needed to ensure the integrity of the samples and to provide a system to trace
samples in the event that the samples are lost or damaged in transit. Since all samples collected at
the field sites will be labeled according to date of collection, sample type, sample location, and
sample class, each sample should have its own unique identification. For projects generating data
to be used in policy decisions, it is possible that the data could be challenged in a court of law.
Therefore, for litigation purposes, specific chain-of-custody forms would be required and protocols
developed by EPA should be used.
Describe instrument maintenance and calibration, and performance (QA) checks on instruments.
Performance checks should be done on a regular, specified basis, and results should be recorded.
This section may also include a trouble shooting table for common instrument malfunctions.
Detail the analytical methods to be used for each indicator. If standard methods are used, standard
and operating procedures (SOPs) can be referenced. This section also discusses changes in
methods (if necessary) as the project progresses.
This section should include the criteria that will be used to validate the quality of data; the methods
to be used for the treatment of outliers; equations for calculation or value of the indicator to be
measured; the reporting units to be used; and a description of data verification and validation
phases for the project.
A description of internal quality control, to include a description of the QC sample design and
samples should be given in this section. If control charts are used, they should be described here.
Describe the performance systems audits (both internal and external) used to monitor the Audits
performance of the measurement systems being used for the project. If laboratories will be
expected to participate in a performance evaluation program of any sort, it should also be included
here.
9 12
-------
Table 9.2. (continued)
Subject Area
Description
Preventative Maintenance
Specific Routine Procedures
be Used to Assess
Data Quality
Corrective Action
Quality Assurance Reports
to Management
Preventative maintenance to be performed on instruments on a scheduled basis, and any critical
Procedures and Schedules parts that should be kept on hand should be included in this section.
Specific procedures to be used for the assessment of accuracy and precision of the data for each to
indicator, including confidence limits, central tendency dispersion, bias, and the five aspects of data
quality should be detailed in this section.
The limits for data acceptability, the point at which corrective action should be initiated, and a
description of the corrective action to be taken for each indicator. Corrective actions can also be a
result of other QA activities, such as performance audits, system audits, and laboratory comparison
studies.
Describe the types and time frames for documents reporting on data accuracy, completeness, and
precision, the results of performance or system audits, and any significant QA or methods problems
and the corrective action taken for resolution of problems.
a Based on requirements of the U.S. Environmental Protection Agency (Stanley and Vemer 1985).
ProcessedSampte
Response/Observation
Data Reduction & Processing
Field Sample |
Duplicates
Dynamic Blanks
Sample Stability Studies
-1
npfe 1
Lab Reagent Blanks
Lab Fortified Blanks
Lab Fortified Samples
Internal Standards
Lab Sample Duplicates
Calibration 1
Check Standards 1
Notebook Audits
Data Audits
Data Base QC
Figure 9.9: Components of the data acquisition process amenable to quality control to
measure and reduce potential sources of errors.
The Technical Director is responsible for determining which activities require SOPs
and for insuring that they are developed, reviewed, and implemented. The development of
9 13
-------
SOPs is delegated to personnel responsible for the activities which requires them. The
Quality Assurance Officer should identify the status of all SOPs in the organization in
periodic audits.
9.8. The Audit Program
The role of audits in the overall QA program is to verify that the work prescribed for
measurements and assessments is being conducted properly. The ARG will conduct
managerial and technical audits and reviews at the program and project level. Audits will aid
in determining if the QA plans are being fully implemented and if they are adequate for the
objectives of the program. Audits were classified into four categories (Table 9.3) in a 1987
guidance document prepared by the Quality Assurance Management Staff for the EPA-Office
of Research and Development (ORD).
Table 9.3. Categories of audits used to determine status of QA in a monitoring program.
Type
Management Systems
Reviews
Audits of Data Quality
Technical Systems Audits
Performance Evaluation
Audits
Purpose
Assess the effectiveness of the implementation
of the approved QA Program Plans.
Check data accuracy and determine if sufficient
information exists within the data set to support
assessment of its data quality, and to verify that
DQO requirements have been achieved.
Verify conformance to the QA Project Plans and
that good laboratory and field practices have
been used.
Assess. laboratory and field analyses based on
results achieved in the analysis of blind
samples; serve as a check on the comparability
of data between ecosystems.
Frequency
Annually
Semi-annually
Semi-annually
Semi-annually
Processes for monitoring the QA/QC process are summarized in Figure 9.10.
9.9. Documentation
9.9.1. Specific Documents
The following documents and information must be current, and must be available to
all laboratory personnel and cooperating organizations:
o
Quality Assurance Program Plan A document addressing all items in Table 9.2
which includes clearly defined laboratory protocols, including personnel
9 14
-------
Develop
Re-Analyze Sample |
}M
L
| Correct Problem |
t
Develop a Flowchart I
for the Activities |
Identify Critical Processes |
Duality Control Materials & Techniques]
1
Establish Statistical Control |
[aintain Quality Control Charts |
Evaluate Performance |
|
Report Data |
Figure 9.10: An example control chart for monitoring the Quality Assurance and Quality Control of data.
responsibilities and the use of QA/QC protocols.
o
o
Laboratory Standard Operating Procedures (SOPs) A document containing detailed
instructions related to laboratory and instrument operations.
Field operations manuals Documents containing detailed instructions, including
field forms, for all field operations.
9.9.2 Documentation of the attributes of data quality
Data quality objectives must be established for the five aspects of data quality:
representativeness, completeness, comparability, accuracy, and precision, as well as for
detectability. Representativeness, completeness, and comparability are difficult to quantify
(Taylor, 1987). They relate primarily to the research design, the selection of sampling and
analytical methodologies, and the resulting data base. Precision, accuracy, and detectability
are quantifiable criteria that are developed for the different collection and measurement
9 15
-------
systems (and the individual components within those systems) being used in the project.
These data quality aspects are listed below, along with definitions and a description of the
type of information generally used to document each.
Representativeness
Representativeness is defined as "the degree to which the data accurately and precisely
represent a characteristic of a population parameter, variation of a property, a process
characteristic, or an operational condition" (Stanley and Verner, 1985). Representativeness
can apply to the location of sampling or monitoring sites, to the collection of samples or field
measurements, to the analysis of those samples, and to the types of samples being used to
evaluate certain aspects of data quality. The quantifiable aspects of representativeness relate
to the accuracy and precision of a measurement system.
Completeness
Completeness is defined as "a measure of the amount of data collected from a
measurement process compared to the amount that was expected to be obtained under the
conditions of measurement" (Stanley and Verner, 1985). Completeness can be expressed in a
variety of ways for various measurement programs, but is generally expressed as a percent of
the number of samples expected to be collected that are actually collected.
Comparability
Comparability is defined as "the confidence with which one data set can be compared
to another" (Stanley and Verner 1985). Comparability should be addressed in a qualitative
sense by the use of identical or similar methodology in the various measurement programs.
The implementation of similar rigorous quality assurance programs for all projects will also
maximize the comparability of the various data bases. The use of standardized methods, as
outlined in the QA plan, field sampling manuals, and training manuals will provide a basis for
generation of data with a high degree of comparability.
Accuracy and Precision
The term "accuracy" is defined as the difference between a measured value and the
true or expected value, and represents an estimate of systematic error or net bias (Kirchner,
1983; Hunt and Wilson, 1986; Taylor, 1987). Precision is defined as the degree of mutual
agreement among individual measurements, and represents an estimate of random error
(Kirchner 1983, Hunt and Wilson 1986, Taylor 1987). Collectively, accuracy and precision
provide an estimate of the total error or uncertainty, associated with an individual measured
value, which has been termed "accuracy" by Taylor (1987) and Hunt and Wilson (1986).
Accuracy and precision of a measurement system are documented by analyzing various types
of quality control and quality assessment samples and providing an interpretation of those
results. Samples used to control data quality at the laboratory level are known to the analyst
and are reviewed continually by the laboratory technician and the Quality Assurance Officer.
Types of samples used for this purpose include blanks, internal duplicates, matrix spikes, and
quality control check samples. Results of these sample analyses are plotted on control charts
for a continuous record of laboratory performance. Quality assessment samples are submitted
9 16
-------
to the analyst as double blind samples (also known as a performance check sample). These
sample results are assessed by an independent agency and provide the basis for overall
measurement error documentation.
Detectability
Detectability is operationally defined as the lowest concentration that can be measured
above a specified value (either zero or some background value) with a specified level of
confidence. Objectives will be established for two aspects of detectability: analytical limits of
detection and the level of tolerable contamination due to collection, handling, processing, and
measurement (operationally defined as "background"). The limit of detection of a particular
analytical method represents the lowest level or quantity that can be reliably detected with a
specified level of confidence. Related to the limit of detection is the limit of quantitation,
which is the lowest level or quantity where a method yields quantitative results (i.e., the
uncertainty associated with a measured value is less than 30 percent; Taylor, 1987). The limit
of quantitation thus defines the lower limit of the method in terms of providing interpretable
results. Background levels in samples will be minimized by careful adherence to sampling,
handling, and processing protocols, and by establishing stringent control limits for these
measurements.
9.10. Summary
In summary, quality assurance is an integral part of the process of planning for the
Agroecosystem Resource Group. The process for the ARG can be summarized in 10 steps, as
shown in Table 9.4.
9 17
-------
Table 9.4. Phases of EMAP planning and implementation with products. Specific QA products are included.
PHASE
1.
2.
3.
4.
5.
6.
7.
Initial preparation
Objectives of the program
Data required
End use of data
Total allowable error
Resource allocation
System design
Data processing
Field operations
Analytical and laboratory operations
Data reduction/analysis/validation
Audits
8. System certification
9.
10.
Conduct study
Data reporting
PRODUCT
DQOs
Budget request
EMAP and Agroecosystem Information Centers
QA Project Plans and Logistics
QA Project Plans
EMAP Information Center and Statistics Group
Audit Reports and Corrective Action Memos
Pilot and Demonstration Projects
Implementation
Statistical Summaries and Interpretative Reports
9 18
-------
10. Information Management
10.1. Overview
The Agroecosystem program will require that data be obtained, stored, manipulated,
integrated and analyzed. These new and existing data will come from many sources, including
joint ARG-NASS data collection efforts and from other EMAP Resource Groups, other
government agencies, cooperating non-government organizations (NGOs), and academic
institutions (Figure 10.1). The information collected and created by the program must also be
available, at some level, to researchers in these same agencies and institutions. Researchers
must be able to determine what data are available, where they are located, and how they can
be accessed. Information about methods used to collect data, including details about data
quality (Chapter 9), must also be readily available.
Agroecosystem
Data Collected
With MASS
I
f
Other
Existing
Data
Agroecosystem
Information
Center
Other
EMAP
Data
Agroecosystem Resource Group
PRODUCTS
o Reports
o Data
Figure 10.1: Overview of the flow of data through the Agroecosystem Information Center.
A major emphasis of the Agroecosystem pilot program scheduled for 1992 is the
development of close working relationships with USDA-NASS (Chapter 11). As discussed in
Chapter (3.4), NASS uses an area frame to gather data on crop acreage, cost of production,
farm expenditures, crop yield, specialty crops, livestock production, chemical usage, irrigation,
water quality and other items of interest to the agricultural community. Statistics are
compiled and reported annually from the June Enumerative Survey using some 16,000
Primary Sampling Units.
10 1
-------
Confidentiality of data, and consequently data security, are particularly critical issues
for the ARC. Meeting the program objectives requires that data be collected from individual
farmers and corporations. Because EMAP data are to be widely available, there must be a
policy and mechanism which protects the privacy of the individual respondents.
Success in the information management effort depends upon the development of an
infrastructure of computational capability; a communications facility which can support the
movement of large volumes of data; a coordinated effort across all resource groups;
interaction with many different agencies and universities; and support from a central
organization within EMAP to support and coordinate these efforts. The EMAP Information
Management Committee (IMC) has been established to address these and other needs. The
Resource Groups will supply input to the IMC regarding needs, and look to the IMC for
advice, standards, and assistance.
10.2. Relationship with USDA-NASS
The current approach within the ARG is to develop a cooperative agreement with
USDA-NASS under which NASS enumerators collect all or most of Agroecosystem indicator
data or samples for analysis (Figure 10.2). These enumerators will operate within the NASS
organization, using procedures selected and developed jointly by the ARG and NASS.
Microdata will be entered, verified, validated, and stored on NASS computers; aggregated
datasets will be transferred to the Agroecosystem Information Center within constraints of
NASS confidentiality agreements.
From the standpoint of information management, working with NASS is attractive for
the following reasons:
1) Over time, NASS has developed a relationship with the agricultural community
which will greatly facilitate the collection of data.
2) NASS provides confidentiality of data to individual farm operators. This allows
the organization to collect data which farmers might otherwise be reluctant to
supply for fear of legal or regulatory action against them.
3) NASS has a fully developed infrastructure for the collection of agricultural
data, including strict quality controls. Use of this infrastructure greatly reduces
the expenditure of resources on the development of new logistics and QA
procedures.
4) NASS has developed the computer resources to organize, analyze, and quickly
report on, large volumes of data. Use of these resources may reduce the overall
need for data processing within the ARG.
10 2
-------
Survey Design 1
„
V
T
F)ata Collprtion I
byNASS 1 ^ ^ ^
^^^^^^^^^™
1
Data Integration
and Analysis at
„ „, ... NASSbyARG
Confidentiality
Screening ./ ^_ 1
M
Agroecosystem
Information
Center
Field Sampling Prodcedures I
j
Sample Collection 1
by NASS 1
1
1
T
Lab Analyses 1
Figure 10.2: Flow of data collected by NASS to the Agroecosystem Information Center.
10.3. Use of Existing Data
The ARG is committed to the use of existing data whenever possible (Figure 10.3).
Although there may be some effort required to transform existing data to conform with
EMAP data standards, this effort is usually substantially less than that required to collect new
data. In many cases, ancillary data will be required for the development and interpretation of
our indicators (See Table 6.4). For example, meteorological data is critical to the development
and interpretation of any yield or productivity measures. The use of existing data also makes
analysis of historical trends possible. In this way, it may be possible to validate indicators by
attempting to predict present conditions using historical data. The ARG will import data as
needed and appropriate from other EMAP efforts as well as other agencies and organizations.
10.4. Quality of Data
To ensure that data are of the highest quality, carefully designed procedures for the
movement and manipulation of data, from field collection through analysis, are essential (See
10 3
-------
Chapter 9). Quality control considerations are critical throughout these efforts, and quality
control procedures should be incorporated from the outset. Data quality ponsiderations are
also important during the evaluation of existing data. The Agroecosystem Information
Manager, Quality Assurance Officer, Indicator Development Teams, and NASS Quality
Control personnel must work cooperatively in order to attain this goal (Figure 10.4). Data
collected by the ARG will be grouped into at least three levels (Fig. 10.5).
EMAP
Landscape
(GIS) ]
Confidentiality_
Screening
Other EMAP
Resource- Group
Data
Other Data
at EMAP
Information Center
Agroecosystem
Information
Center
NASS
Data
Center I
T
Other Existing Data (eg):
National Resource Inventory (SCS)
Agricultural Census (USDC)
STORE! (EPA)
Meteorological Data (NWS)
N.C. Center for Geographic Information
Figure 10.3: Flow of data from other EMAP sources and other agencies and institutions
to the Agroecosystem Information Center and NASS data center for integration.
Raw data, which is data as collected in the field.
Verified data, which have been analyzed for consistency on an item by item basis, and within
single records.
Validated data, which have been subjected to consistency checks across records.
A fourth level of data, called Enhanced Data, is also being considered by several of the
EMAP Resource Groups. Enhanced data are obtained from validated data that have missing
values filled in using established statistical procedures. No decision has been reached
regarding the creation of enhanced data by the Agroecosystem Resource Group.
10 4
-------
Data Collection Procedures
Data Collection
Raw Data
Verified Data
I
Validated Data
IS
Data Analysis I.
Quality
Asssurance
Checkpoints
Other
Existing
Data
Figure 10.4: Data flow showing the close interaction between Total Quality and Information Management efforts.
Quality control points are shown along the data flow path.
10.5. Confidentiality of Data
In order to protect the rights of individual respondents, strict confidentiality provisions
apply to all data collected by NASS. NASS will not release microdata; data are currently
available at the county level. The USDA Soil Conservation Service's (SCS) National
Resource Inventory (NRI) and the United States Department of Commerce's (USDC) Census
of Agriculture are also subject to confidentiality provisions. Table 10.1 summarizes some of
these policies. The rationale behind such assurances is fairly obvious. Confidentiality laws
protect individual respondents from prosecution which might otherwise result from their
participation in a data collection effort. Without such assurances, respondents may be hesitant
to comply with any survey or data collection efforts, either voluntary or legally required.
Also, without the assurances, respondents may be more likely to falsify information on
surveys. Violation of this confidence would result in loss of NASS's credibility with survey
10 5
-------
Raw Data
Verified Data
Check Values for Reasonableness
eg.: Range Check
Duplicate Entries
Format Check
Inappropriate Codes
Internal Consistency
Check Data for Reasonableness
eg: Comparison with Historical Data
Statistical Analyses
Validated Data
Figure 10.5: Application of QA procedures to ensure that data are of the highest quality.
respondents and seriously hamper future data collection efforts. Hence, NASS is very serious
about maintaining this confidence.
The current view of the ARC with respect to these confidentiality provisions is
positive. In a review of confidentiality in EMAP, Franson (1990) writes that the EPA is
presently unable to issue a blanket statement of confidentiality for EMAP data. Requests for
EMAP data from the regulatory arm of EPA, or from other agencies, corporations, and
individuals, must be reviewed on a case-by-case basis. Farmers are very unlikely to provide
data without a confidentiality agreement. In fact, the question of whether farmers would
cooperate with any team identifying themselves with the EPA, even with promises of
confidentiality, should not be lightly dismissed.
10.5.1. Working Within Confidentiality Guidelines
Although the microdata from agencies employing confidentiality provisions are not
available, there are solutions which allow EMAP to make use of the data collected by NASS.
10 6
-------
Aggregated data: While many agencies will not release their microdata, they will all release
data aggregated at various levels. The goal is to aggregate the data in such a way that
individuals cannot be identified. Obviously, for EMAP, the lower the level of aggregation,
the better (i.e., county level data is better than state level data.)
Analysis requests: The owning agency may accept requests for tabulation and analysis from
another agency. The analysis would be performed by the owning agency's personnel, the
results returned to the requesting agency. Release of any confidential material would be
strictly avoided.
Deputization: It is possible, at least with some government agencies (including NASS), for
someone to be deputized by that agency. Deputization requires completion of a non-disclosure
agreement. A deputized individual is permitted to access the data for the purpose of
performing analyses. Typically, the analysis would have to be performed at the owning
agencies facilities. Only aggregated results may be removed, and are subject to
confidentiality screening.
All of these possibilities will be explored by the ARG with NASS, and any other
agencies with such provisions, during the pilot program. Several ARG members have already
been deputized by NASS and have access to microdata at NASS headquarters in Raleigh,
North Carolina.
Table 10.1: Summary of confidentiality provisions of several government agencies with data of value to the
Agroecosystem Resource Group.
Organization
USDA National Agricultural
Statistics Service
USDC Census of Agriculture
USDA Soil Conservation
Service (National Resource
Inventory)
US Environmental Protection
Agency
Policy
Public Law 99-198.
No release of data with identity of individual
respondent.
US Code Title 13.
No release of data with identity of individual
respondent.
No release of exact location (Primary Sampling
Unit) at which data are collected.
Freedom of Information Act requests handled
on a case-by-case basis.
Lowest Level of
Aggregation Normally
. Available
County
Zip Code (5 digit)
County
Varies
10 7
-------
10.5.2. Data Integration
The integration of EMAP data with data subject to confidentiality provisions presents
a unique challenge which can be resolved only through close interagency cooperation. NASS
data are used for economic forecasts which have the potential for effecting the livelihood of
many people. They are closely guarded and access is severely restricted and carefully
monitored. No microdata may be removed from NASS facilities, and computers containing
microdata may not be connected to foreign networks. During the pilot program NASS will
allow the ARG to install a workstation at the NASS facilities in Raleigh, NC. ARG members
will be able to examine and analyze NASS microdata using the workstation. Integration of
data from other sources may be accomplished by loading that data onto the ARG workstation
at the NASS facilities and performing the required analyses on that workstation (Figure 10.3).
Although this mode of operation will suffice during the pilot program, other
approaches will be explored for national implementation.
10.5.3. Computer Data Security
We cannot guarantee confidentiality to our respondents, or any agency from which we
request data, if our computer data systems are not secure. (Of course, we must also provide
proper security for any sensitive paper documents as well.) A system of access levels and
password protection should be designed. Additionally, it may be desirable to keep sensitive
data off-line except when it is needed for analysis. When needed, the data could be made
accessible to authorized users. This issue should be addressed by the EMAP Information
Management Committee (IMC).
10.5.4. Recommendations Regarding Confidentiality
The issue of data confidentiality has been discussed, with several examples given.
There are ways to work within these agreements, but, microdata are not available for use
outside of NASS. These issues may place constraints on integration at the microdata level.
When evaluating a database for potential use, the ARG will inquire about the owning
organization's confidentiality policies. Most agencies are not accustomed to requests for
microdata and may not understand that microdata are being requested. If an organization has
a confidentiality policy, we will determine how members of the ARG may work within their
policy. It is also very important to inform the other agency that, under current EPA policy, all
microdata is subject to release under the Freedom of Information Act.
As a program, EMAP must address the issue of providing data confidentiality.
Because the public sees EPA as a regulatory agency, respondents are likely to be
uncooperative unless provided with legal guarantees prohibiting regulatory action against them
based on data collected for the EMAP program. This is an important and complex issue
which should be addressed in the very near future. Franson (1990) has written an excellent
10 8
-------
summary of this issue, and recommended that the EPA Office of General Counsel be involved
immediately.
10.6. Information Management Objectives for Pilot Program
The stated goals of the pilot program scheduled for 1992 that are relevant to
information management include the following:
1) test concepts relating to data management
2) test concepts relating to data analysis
3) full utilization of existing data bases
In the course of the pilot program, the ARG will be dealing with data from:
1) other government agencies and NGOs
2) data analyses from databases of other government agencies and non-
government agencies
3) sampling and monitoring data from the Agroecosystem program
4) sampling and monitoring data from other EMAP Resource Groups
Specific objectives for the information management component of the pilot program
include:
1) Work cooperatively with NASS in the definition of data elements for the ARG.
This task includes the construction of a data dictionary.
2) Work cooperatively with NASS in the collection, entry, and analysis of pilot
program data.
3) Work cooperatively with NASS to develop protocols for the transfer of
aggregated data to the ARG.
4) Work cooperatively with the IMC and NASS to develop a data management
system to support the receipt and further analysis of aggregated data from
NASS at the Agroecosystem Information Center (AIC).
5) Work cooperatively with the IMC and NASS to develop procedures to support
the transfer of EMAP data to the NASS computer center for any integrated
analyses at the microdata level.
10 9
-------
6) In conjunction with 4) above, work cooperatively with the IMC to develop
EMAP standards for database design. This includes input into the process of
selecting a standard EMAP database.
7) In conjunction with 4) above, work cooperatively with NASS to develop a
protocol for the transfer of aggregated data from NASS to the AIC.
8) Work cooperatively with the IMC to develop a protocol for the transfer of
Resource Group data to the EMAP Information Center.
9) Work cooperatively with NASS and the IMC to complete resolution of
confidentiality issues.
10) Support efforts to obtain, convert, and analyze existing data bases.
11) Establish and support the Agroecosystem Resource Group Information Center.
This task includes the evaluation, acquisition, installation and maintenance of
required resources; system administration; network administration; training of
Agroecosystem Resource Group members; and coordination of activities with
the IMC.
10.7. Information Management Resource Requirements for the Pilot Program
The anticipated information management needs for the pilot program (1990-1992) are
discussed in Chapter 11. Information management needs beyond the time frame of the pilot
program are dependent upon the outcome of that program. There is a vast difference in
information management requirements between a data collection effort mostly carried out by
NASS, and one in which the ARG must collect the data. As progress is made on the pilot
program, we will be better able to predict information management requirements for and
beyond 1992.
10 10
-------
11. Research and Monitoring Accomplishments and Plans (1990-1995)
The Agroecosystem Resource Group (ARG) has developed a five year program (1991-
1995) for national implementation of a suite of indicators for monitoring agroecosystem status
and trends. This five-year period includes time to test concepts relating to design, indicators,
QA/QC, data analysis, and information management at the pilot, demonstration and
implementation stages. A primary emphasis is the development of close working relations
between personnel from NASS and the ARG, so that issues relating to design, sampling,
logistics, information management, QA/QC, and data analysis can be identified and addressed.
The first stage of the program (1990) encompassed the evaluation of: 1) statistical
designs, 2) existing monitoring programs (i.e., NASS, SCS, ERS), 3) assessment endpoints
and associated indicators (availability, validity, variability, cost), 4) data management and
analysis techniques and 5) derived outputs. During 1990, a conceptual national monitoring
plan was also developed. The second stage will involve the design and execution of pilot and
demonstration projects prior to implementation. The first pilot, planned for the state of North
Carolina, will test all aspects of the monitoring program for a selected suite of five indicators
(Table 11.1). Results will be utilized to develop a regional demonstration of all program
elements in the Southeast (SE) followed by a full implementation in the SE. Additional pilots
will be undertaken in selected states and one or two additional demonstration projects will
occur in one or two midwestern or western regions (Table 11.2 and 11.3). Assuming the
pilots and regional demonstrations are successful, we anticipate being fully ready for a
national implementation by 1995; funding levels will determine degree of regional and
national implementation. The pilots and demonstration studies will address specific concerns
of the different geographic areas of the country.
This chapter addresses specific research objectives and tasks planned for 1991, and for
undertaking pilot and demonstration projects in subsequent years. The chapter does not
address details as to how the various tasks and plans will be accomplished. Many of the
details are contained in earlier sections of this document. Timelines for accomplishing tasks
are shown, as are projected budget levels through 1995 (Table 11.2). Although details of the
implementation plans are not included, it is expected that these plans will be similar to those
used for the demonstration.
11.1. Rationale
It is essential that this multi-stage program be followed to assure successful regional
and national implementation of the Agroecosystem program. The multi-stage program will
permit critical evaluation of the monitoring design, indicators, data analyses and integration,
logistics, QA/QC, and information management. This evaluation is essential to the
development of a successful, effective and efficient implementation plan for monitoring the
status of agroecosystems on a national scale. The program, as designed, allows for the
orderly establishment and rigorous evaluation of preliminary protocols and for the full
utilization of existing data bases and networks.
11 1
-------
Table 11.1. Tasks with schedule for conducting the pilot projects NC Pilot Plans (1992-
93)
Tasks
1. Supply NASS with procedural manuals for their training programs (to meet NASS
schedule).
2. Use suite of five indicators identified in the planning stages.
3. Assure that logistic, QA/QC (TQM), and information management plans are updated
and in place.
4. Participate in the NASS Enumerator Training Schools: a) May - procedures for
sampling irrigation water; b) October - procedures for sampling soils.
5. Obtain all necessary equipment and materials for the pilot.
6. Sample 100 to 200 NASS segments using NASS personnel, logistics, QA/QC and
data management protocol.
7. Work with NASS on data management and data analysis.
8. Send soil and water samples to contract laboratories for analysis and data return.
9. Compare the two design approaches through sampling of units in the field; cost,
variance, biases; determine covariance structure to refine DQOs.
10. Data analysis: provide statistical summaries, compare cumulative distribution
functions, explore spatial distribution patterns, and examine statistical properties.
1 1 . Develop data summary to derive initial indices to classify agroecosystems as
"healthy" or ''unhealthy1'.
Schedule (1992)
January (for May)
July (for October)
Developed in 1991
February
May
October
April
June
July-August
December
(NASS Survey Dates)
June-December
March 1993
(NASS time periods)
August/October
December 1992/
February 1993
December 1992
March 1993
March 1993
April 1993
11.2. Accomplishments Through 1990
Accomplishments to date focused on four primary areas of effort: 1) statistical
designs (see Chapters 3-5) and the development of an example statistical summary document
(Meyer et al. 1990), 2) development of candidate and research indicators (See chapter 6), 3)
establishment of strong working relationships with NASS and formulation of information
necessary for our cooperative operations, and 4) development of this national Agroecosystem
Research Plan.
11 2
-------
Table 11.2 Planned implementation of Agroecosystem monitoring and assessment across
EPA regions.
EPA Regions Years (Funds in thousands)
1 - ME Boston
2 - NE New York
3 NE Philadelphia
4 - SO Atlanta
5 NC Chicago
6 - SO Dallas
7 NC Kansas City
8 NC Denver
9 - WE San Francisco
10 - WE Seattle
1991 (300)
....
....
PI
— -
-—
1992 (800)
....
Pil
....
PI
....
1993 (2440)
—
De
....
Pil
PI
1994 (4060)
PI
—
Im
PI
PI
De
- — .
Pil
1995 (6200)
PI
Pil
PI
Im
Im
Im
Im
PI
De
PI
PI Planning (write a peer-reviewed Project Plan)
Pil Pilot Project (1 or 2 states)
De Demonstration Project
Im Implementation in EPA Region (full implementation in megaregion)
As discussed in Chapter 3, the statistical design effort involved the identification of a
probability sampling frame that would most effectively and efficiently meet the goals and
objectives of the Agroecosystem program. Our design team spent considerable time
discussing the advantages and disadvantages of the EMAP hexagon area frame and the NASS
area frame. The team recommended that the NASS area frame best met our goals and
objectives. Currently, we are evaluating the NASS PSU sampling sequence (rotational plan)
compared to using the NASS PSU encompassing the hexagon centroid (hexagon plan). In
either case, the NASS area frame will be utilized and NASS enumerators will collect the
information.
Most of the effort from the ARG has been devoted to the development of candidate
and research indicators. Several workshops have been held and outside expertise brought in
and consulted. The indicator development strategy was discussed in detail in Chapter 6 and
supplementary data information is presented in Appendices 5-8. A list of 16 candidate
indicators was identified and fact sheets on each were developed by ARG members (Heck et
al. 1989). From this group, five research indicators (see Chapter 6) were identified for use in
11 3
-------
Table 11.3 Planned implementation of Agroecosystem monitoring and assessment across
four U.S. mega-regions.
Mega-Regions
Southern
(EPA Regions 4, 6)
North-Central
(EPA Regions 5, 7, 8)
West
(EPA Regions 9, 10)
Northeast
(EPA Regions 1, 2, 3)
Implementation Plans (1991-1995)
Stage of
Implementation!'
State pilot (NC)
Demonstration project (4)
Implementation (4)
Full implementation (4, 6)
State pilot (KS)
Demonstration project (7)
Implementation (5, 7)
State pilot (CA)
Demonstration project (9)
State pilot (NY)
Resource^
Field Crops
Field Crops
Field Crops
Field Crops
Field Crops
Field Crops
Field Crops
Field Crops
Field Crops
Field Crops
Year
1992
1993
1994
1995
1993
1994
1995
1994
1995
1995
I/ State pilot 100-200 samplings units to permit statistical testing of design options and five indicators.
Demonstration project - 200-400 sampling units to identify minimum number of sampling units required for regional
implementation. Five indicators will be monitored initially. Depending upon funding, a reduced number of segments
may be monitored.
Implementation - Full implementation of the design option accepted for the agroecosystem program. The number of
sampling units to be determined from the demonstration; does not include all EPA regions in the mega-region.
Full implementation - implement all EPA regions (states) within the mega-region.
2/ Resource Classes for EMAP Agroecosystems
a) Field Crops (Agronomic, Vegetable and Forage)
b) Fruit and Nut Crops
c) Managed Pasture and Grazing Lands
d) Confined Feedlots
e) Non-Managed Resources "Natural Areas" (i.e., shelter belts, ditch banks, woodlots, etc.). Please note that the
extent of these natural areas will be monitored with the Land Use indicator in all pilot and demonstration
projects.
Note: Indicators for use with other resources are being developed. As they are developed and tested, they will be
added as funds are available.
11 4
-------
the pilot and demonstration projets: crop productivity, soil quality, irrigation water quality
and quantity, land use and agricultural chemical use.
Selected research was partially funded to support development of bioindicators that
may help interpret changes in response indicators. Support for the statistical analyses of
biological monitoring data at the state level (Minnesota) was used to show that plant injury
was associated with ozone pollution in most areas of the state, within a given year. Research
was partially supported to determine the value of sensitive vs. tolerant clones of white clover
in interpreting the potential effects of ozone on crop resources. Manuscripts have been
submitted for both of these efforts.
A close working relationship with NASS at both the national and state (North
Carolina) level, has been achieved. NASS personnel are excited about the possibilities of the
EMAP program and have worked closely with us in the development of indicator questions
for the enumerator questionnaires. NASS enumerators will take these questionnaires to the
field for interviewing farm operators. A series of questions has been completed for two
NASS surveys (June, December) and have been put into questionnaire form for approval by
OMB (a requirement for any interview survey). Details are found in Appendix 6.
11.3. Identification and Use of Data from other Monitoring Efforts
The ARG has allocated a portion of the time of two members in identifying existing
data sources. Information on other data sources relevant to agroecosystems is briefly covered
in sections 2.3 and 10.3, with more detail shown in Appendix 3. Selected data from NASS
were used in the "Annual Statistical Summary Report on Agroecosystems" (Meyer et al.
1990). Some of the information used was summary data from NASS state-level statistical
reports and some was from the SCS/NRI; both were used in the report for illustrative
purposes only. Additional data have been obtained from NASS for evaluating statistical
aspects, such as appropriate sample size, of the pilot programs. We have plans to utilize
additional data from NASS as we explore approaches to data analysis.
We have met with both national and state (NC) SCS-NRI staff members and plan to
initiate more in-depth discussions with them. We intend to have a completed list of their
monitored parameters, and to determine how they handle their data. We will request access
to some of their microdata to see if we can use it as complementary data for the
Agroecosystem program.
We plan to contact other agencies that are potential sources of monitoring data to
determine the usefulness of the data in design, analysis and interpretation of our program. A
number of organizations have been contacted (Appendix 3), but we do not yet know the value
of their data for our purposes. It is, however, imperative that we make maximum use of
other data where appropriate.
11 5
-------
11.4. Pilot, Demonstration and Implementation Plans
This section addresses the developmental stages of the Agroecosystem program. The
rationale for specific pilot and demonstration efforts, as well as the tasks needed to
accomplish the pilot and demonstration projects, are highlighted. Several levels of full
implementation are discussed. It should be noted that all three levels of monitoring (pilot,
demonstration, implementation) are limited to five indicators focused on the production health
of the agroecosystem (field crops resource, Table 11.3). We are developing indicators to
address the remaining resources. These will be field tested and added to planned pilot
projects. However, current funding projections may not permit implementation of additional
indicators in the agroecosystem monitoring program, in the near term.
11.4.1. Pilot Projects
The ARG spent considerable time determining the appropriate sample size, including
the minimum sample size needed to meet specific statistical goals. The minimum number of
sample units was determined to be 100 units 50 hexagons and 50 current NASS sample
units. For the first pilot the ARG is recommending 200 sample units (100 hexagons plus 100
NASS PSU's) as a reasonable sample size to address a number of specific objectives of long-
term importance to EMAP and the Agroecosystem program. It is anticipated that subsequent
pilots will require fewer than the 200 sample units (i.e., 100 to 150). We believe the money
spent on this initial effort will have a long-term benefit, because it will allow us to expand to
regional and national implementation based upon a firm statistical foundation. We have
briefly listed our objectives and reasons for recommending the 200 sample units and
conducting the initial pilot in North Carolina in the following three sections:
Design/Statistical
1. We consider the pilot to be a small regional design. We plan to determine the
variability across the diverse landscapes (physiographic regions) found in the
Southeast, i.e. coastal plain, piedmont, and mountains. The minimum sample size of
100 would permit an overall estimate of variability, but would not provide adequate
resolution to address DQOs (data quality objectives) and variability within the
physiographic regions. A sample size of 200 would provide sufficient information for
the development of DQOs and to estimate variability within each of the three major
physiographic regions which are representative of the Southeast.
2. We need to test the efficiency of the stratification built into the NASS sampling design
(rotational plan) against the post-stratification anticipated from sampling the hexagons
(hexagon plan).
3. We want to determine within-sample-unit variability for our indicators as well as
variability across sample units within each of the three physiographic regions.
11 6
-------
4. The larger sample size will permit reasonably good estimates of the correlation
structure among the indicators, even within each of the physiographic regions.
5. The use of 200 sample units for the initial pilot, with the statistical strength this
implies, will permit a better estimate of the number of sample units required in a full
regional demonstration and subsequent implementation to assure that we reliably
establish and meet our DQO goals for EMAP.
Logistics
1. North Carolina was selected as our preferred small region for a number of reasons:
a. The three physiographic regions (coastal, piedmont, mountain) associated with
the entire Southeastern region are well represented in North Carolina.
b. NASS is organized on a state-by-state basis and enumerator training is done in
each state. By staying within a single state (i.e. North Carolina), we only need
to work with a single NASS state organization for the pilot. This simplifies the
resolution of problems and concerns during the development of logistics,
design, and implementation of the pilot.
c. We will lose nothing by starting in one state because NASS enumerator
manuals and procedures will transfer to each new state as states are added to
the program.
d. The State of North Carolina is interested in our program and is following it
very closely. They have GIS data and capability as well as other data which
they are sharing with us. Doug Lewis of the NC Department of Environment,
Health and Natural Resources is a member of our Resource Group.
e. The core staff of the Agroecosystem Resource Group is located in Raleigh.
For the first pilot study, this greatly facilitates the logistics activities.
2. We need to use sufficient sample units (minimum of 100) to permit a realistic
evaluation of the logistic problems we will encounter in subsequent demonstration and
implementation projects.
3. We need to test the logistics associated with the measurements we have planned for
each indicator (100 sample units are adequate).
4. We estimate the cost of 100 sample units to be about $550,000 ($200,000 for NASS)
above our base funding of $300,000. The cost of an additional 100 sample units is
only $450,000 ($125,000 for NASS) above the $850,000 level. These are minimal
cost estimates.
11 7
-------
Supplementary
1. Many states are interested in EMAP; we believe many state environmental and
agricultural groups want to develop programs to assess the "health" of their own
systems. The 200 sample units should provide sufficient information for us to
recommend the number of sample units needed to give status and trend information on
a state or lower level. The potential ability of state organizations to utilize EMAP
data (or the EMAP monitoring design) could be another strong point for the program.
2. The use of 200 sample units, with the costs projected by NASS, may permit us to
seriously consider the value of sampling a full 3200 sampling units on an annual basis
rather than using an interpenetrating design with 800 sample units per year. This
would permit the detection and interpretation of regional trends earlier than would be
possible using 800 sampling unit per year in an interpenetrating design.
Specific tasks for the pilot project are shown in Table 11.1. Although the tasks are
scheduled to cover the North Carolina pilot, they would be similar for other state pilots.
Results from each pilot will be used for planning of the demonstration project in the same
EPA region (Table 11.2, Figure 11.1).
11.4.2. Demonstration Projects
Demonstration projects are undertaken on a larger scale than the state pilot projects,
but the rationale is overlapping. The primary purpose is to test the validity of the pilot results
on a regional scale. Even with the initially recommended number of sampling units (400 per
EPA region) the number per state is much reduced from the recommended number for the
first pilot study.
Before each demonstration project is initiated, techniques and data from their
associated state pilot must be reviewed thoroughly. The following areas are essential:
1. Critical and empirical evaluation of indicators to: a) evaluate the ability of an
indicator to address assessment endpoints of interest; b) establish an initial range of
values for the indicator; c) assess spatial variability; d) examine the reliability and
information content of each indicator; e) identify the usefulness and sensitivity of the
indicator in determining ecological condition; and f) determine the cost-effectiveness
of the indicator.
2. Compare the cost of the Hexagon Plan compared to the Rotational Plan for the
statistical design option.
11 8
-------
3. Resolve issues with regard to: a) sampling techniques; b) sampling logistics; c)
quality assurance procedures; d) data and information management; e) data analysis
and summarization; f) reporting methods, and g) development of health indices.
4. Develop and refine the logistic, QA/QC, and information management plans for initial
regional implementation of agroecosystem monitoring and assessment.
Tasks related to the demonstration project are essentially those for the pilot (Table
11.1), only on a larger scale. The primary purpose of the demonstration is to verify that the
procedures tested in the pilot, with recommended changes, will operate on a regional scale. A
successful demonstration project should insure a successful regional or multi-regional
implementation.
11.4.3. Implementation
Implementation is shown in Table 11.2 on the basis of specific EPA regions. The
implementation schedule can also be based on four mega-regions which include 2-3 EPA
regions. Table 11.3 shows the 4 mega-regions and associated EPA regions, and plans for full
implementation within each mega-region. In each case, implementation should not be
attempted without a successful demonstration project. The map depicts plans for
implementing the mega-regions and each EPA region (Figure 11.1). The ARG currently
plans to implement the program in four EPA regions with major agricultural areas by 1995
with a demonstration in region 9 (CA and AZ) and a state pilot in NY. Thus, the primary
agricultural areas will be included in the Agroecosystem program by 1995.
Because of the realistic limitation of funding, agroecosystems have been divided into
resource classes that will facilitate the initial implementation of the program (Table 11.3,
footnote 2). The only resource to be monitored in the initial pilots is field crops (Table 11.3).
The field crops resource includes agronomic, vegetable and forage crops. The extent of
natural areas such as woodlots found in the agricultural landscape matrix will also be
evaluated in pilot and demonstration projects. The ARG is starting to develop indicators for
the other four resource classes. As the indicators are developed, they will be tested and
added when funds are available.
We are recommending 200 sample units for the initial pilot (NC, 1992) and 400
sample units for the initial demonstration (SE-Region 4, 1993). We expect the variability in
the SE to be as large as in any other region of the country. Thus, we anticipate from these
initial monitoring efforts a recommendation for sample unit numbers in subsequent pilots and
demonstrations in the remaining EPA regions and, thus, in the four mega-regions. From the
regional demonstrations we should be able to determine the minimal sample number for use
in the regional implementations. With the current funding projections, we anticipate that we
could sample 1600 units by 1995.
11 9
-------
Region?
#1992
@1993
+ 1994
• 1995
Region 4
# 1991
@ 1992
+ 1993
• 1994
* 1995
Agroecosystem Implementation Schedule
# Planning
@ Pilot
+ Regional Demonstration
* Implementation
Figure 11.1: The Agroecosystem implementation schedule.
11 10
-------
11.5. Schedule of Activities
The activities for 1991 will focus on detailed preparation for the first monitoring pilot
in 1992. These efforts are listed in Table 11.4. More detailed tasks and research efforts for
1991 are shown in Table 11.5. Program tasks with budgets are shown in Tables 11.6 (1991)
and 11.7 (1992), while program budgets by location and category for 1991 through 1993 are
shown in Table 11.8. The implementation schedules with number of states, number of sample
units and estimated costs are shown in Table 11.9 for 1990 through 1995. The technical and
administrative personnel needs from 1991 1995 are listed in Table 11.10 and the expected
program products for 1991 1995 are shown in Table 11.11.
11.6. Resources and Plans - 1991 Through 1995
to1
Programs plans with tasks and costs are shown in Table 11.6 for 1991. Program plans
with tasks and costs are shown for approved funding in Table 11.7 for 1992. Approved
funding for 1991. 1992, and 1993 are shown by location and category in Table 11.8. The
cost of a full pilot project in 1992 is also shown. Table 11.9 shows our proposed
implementation schedule with estimated costs and approved or projected funding (in
parenthesis) for 1991 thorugh 1995. This table also shows the number of states for the
planned monitoring and the number of sampling sites for the various pilot, demonstration, and
implementation projects.
Personnel on the project for 1991 and for currently approved funds in 1992 are shown
in Table 11.10. Personnel needs for 1993 represented a reasonable increase to do the two
projects shown in Table 11.10. Personnel needs for 1993 represented a reasonble increase to
do the two projects shown in Table 11.9. The needs for 1994 and 1995 are general estimates
based on the costs for 1993. Expected program outputs are shown in Table 11.11.
11 11
-------
Table 11.4. Activities in 1991 to prepare for the Agroecosystem pilot in 1992.
Design and Sampling
1. Develop a detailed implementation plan for discussion with NASS. Work with NASS to
finalize and implement the plan.
2. Request NASS to characterize PSUs and hexagons; Ground truth these PSUs. Determine
the sample number needed to compare the two approaches.
3. Identify the NASS PSUs to use in the NC pilot.
4. Determine what data is accessible to the ARG and what data needs to be handled by
NASS - develop future protocol for working with NASS on data analysis.
5. Develop cost estimates for design and sampling options based on the Hexagon and
Rotational Plans.
6. Determine characterization cost for EPIC Tier I and NASS characterization of PSUs at
centroid.
Indicators
1. Finish background work on the five indicators identified for the pilot in 1992.
2. Identify the next indicatftrs to be added and develop detailed background information
needed for each of these indicators.
3. Further identify indicators that can be handled by NASS and those that might have to be
done by EMAP-Agroecosystem.
4. Evaluate the five indicators with regard to reporting format, statistical analyses and an
initial interpretive summary (preliminary determination of indices of system condition).
5. Utilize field data to further improve logistics and to provide estimates of variance, sample
size, precision, and other considerations for the 1992 pilot.
6. Construct preliminary indices, such as crop and soil quality indices, that will be
components of agroecosystem health indices related to the assessment endpoints.
Data and Information Management
1. Identify data management needs and prepare a data management system.
2. Develop QA/QC plans for all areas of the program.
Implementation
1. Develop field protocols and logistics plan.
2. In cooperation with NASS, develop additional EMAP training documents for NASS
enumerators.
3. Develop specific plans for the 1992 pilot.
11 12
-------
Table 11.5 Specific research plans for the Agroecosystem Resource Group in 1991.
Task
Scheduled
Completion
1. Finalize plans with NASS for the NC Pilot.
a. Develop complete list of needed tasks.
b. Write manual for procedures to take soil samples.
c. Write manual for taking soil erosion data.
d. Develop protocol for collecting land use data from NASS aerial photographs.
e. Finalize output formats for pilot data.
f. Final review of questions for incorporation in the NASS questionnaire.
g. Complete logistics and QA/QC protocols.
h. Complete Information Management protocols.
i. Develop contract arrangements for analytical laboratories to process soil samples (for pilot and for long-term).
j. Determine range of values expected for each new EMAP question on the NASS pilot questionnaire.
September
2. Address outstanding statistical and design issues.
a. Obtain data from NASS to determine appropriate sample size for field sampling.
b. Ask NASS to delineate and stratify all PSUs within the EMAP hexagons in NC. Data will be used to compare design
options.
c. Analyze existing data on climate, soil, yield and pesticide use.
d. Develop statistical techniques for interpretive analysis of indicator data.
March
September
Continuing
December
3. Deliverables.
a. Complete 1990 Research Plan peer review.
b. Prepare journal manuscript on agroecosystem overview.
c. Prepare Pilot Project Plan.
March
June
September
4. Continue development of soil indicators.
a. Work out procedures for obtaining soil erosion data (USLE): length and slope in the field by NASS enumerators; other
variables from SCS field offices and climate maps.
b. Undertake selected testing of indicators in the field to evaluate variability.
c. Develop background information on soil microbial processes.
September
December
Continuing
-------
Table 11.5. (continued)
Task
Scheduled
Completion
5. Continue development of pest density indicators.
a. Undertake selected testing of indicators in the field to evaluate variability (nematodes, weed diversity).
b. Develop background information on pest density.
c. Develop background information on beneficial insects.
December
Continuing
Continuing
6. Continue development of land use & landscape descriptor indicators.
a. Develop procedures for calculation of landscape descriptors using characterization data from the EMAP 10-hexagon pilot.
b. Determine availability and utility of bird and other wildlife population data; relate to landscape descriptors.
c. Continue evaluation of the Habitat Linear Classification System (HLCS); coordinate with other terrestrial resource groups.
d. Explore the use of remote sensing to estimate crop condition (e.g., foliar symptoms, water stress). Develop a memorandum
to the Remote Sensing Group requesting support in these areas.
September
September
Continuing
December
7. Develop productivity indices.
a. Collect harvest index and rootrshoot ratios from the literature and determine a protocol to calculate net primary
productivity.
b. Identify an appropriate source for climate and air pollution data; obtain the data.
c. Evaluate proposed indices of system health with regard to actual field data (e.g. crop productivity, sustainability index, soil
productivity index).
September
September
Continuing
8. Miscellaneous
a. Develop interfaces with other ecosystems; examine indicator selection, databases and reporting.
b. Develop socio-economic indicator; hold workshop with economists, sociologists, etc.
c. Continue development of water quality / quantity indicators, including development of sampling procedures.
Continuing
December
Continuing
-------
Table 11.6. Program Tasks and Budgets for 1991
Budget
(Thousands')
Finalize plans with NASS for the NC Pilot $ 75
Address outstanding design, statistical, and data analysis issues 91
Develop productivity indices for pilot 35
Develop soil indicators for pilot 25
Develop irrigation water protocol for pilot 12
Develop agricultural chemical use protocol for pilot 10
Develop land use for pilot & landscape descriptors 30
Develop pest density indicators 30
Deliverables 50
Miscellaneous (Found Data, Inf. Man., TQM, Logistics, Other Indicators, 94
Linkages/Integration)
Total EMAP Costs 452
Other (ARS, NCSU, EPA) Costs 255
Total $ 707
11 15
-------
Table 11.7. Program Tasks and Budgets for 1992
Objectives Budget
(Thousands)
Conduct Southern pilot $ 275
Manage and analyze data from pilot 85
Evaluate design and sampling options 60
Evaluate pilot indicators (integration) 90
Evaluate and update data management protocol 50
Evaluate and update QA/QC protocols 25
Collect and analyze data from existing data bases - Determine 50
applicability to EMAP
Develop plans for 1993 Regional Demonstration 75
Strengthen linkages with other ecosystems 40
Evaluate additional indicators 50
Total EMAP Costs $ 800
11 - 16
-------
Table 11.8. Program Budgets by Location Category for 1991-1993, In Thousands
Budget bv Category
1.
2.
3.
Locations (Supportive)
a) Athens
b) Corvallis
c) Idaho
d) Las Vegas
e)RTP
NASS (DC/NC)
USDA/ARS/NCSU
a) Personnel
b) Travel
c) Supplies/Service
d) Advisory Com.
e) Equipment
f) Sample Costs
g) Utilities/Space
h) Indirect
EMAP
(452)
„
-
-
-
-
25
288
33
18
8
30
-
-
50
452
1991
Other
(255)
10
30
10
10
45
-
82
18
30
-
5
-
15
-
255
1992
Approved
(800)
30
10
10
-
25
200
302
40
35
10
45
30
-
63
800
Full Prog.
(1,300)
50
50
30
20
45
325
450
48
44
15
60
60
10
93
1,300
1993
Prel. Approved
(2,400)
100
100
60
40
50
550
830.
80
130
35
75
140
30
180
2,400
-------
Table 11.9. Implementation Schedule and Budget Needs for 1991-1995
YEAR/STAGE
No. of
States
No. of
Sites
EstimatedCost
(Budgeted)
$Millions
1990 Planning
1991 Final Planning/Field Research
1992 Southern Pilot
$ 0.45
0.40
200 (100) 1.3 (0.80)
1993
1994-
1995
Soutem Regional
Demonstration
North Central Pilot
Southern Implementation
- North Central Regional
Demonstration
Western Pilot
Southern Full
Implementation
North Central
Implementation
- Western Regional
Demonstration
Northeastern Pilot
8
1
8
4
1
13
9
3
1
400
100
400
200
100
400
300
200
100
2.4 (2.40)
5.5(4.16)
7.0 (6.2)
11 18
-------
Table 11.10. Long-Term Strategy: Technical and Administrative Personnel Needs i;
Personnel 1991 1995
91 92 (approved) 93 94 - Est. 95 Est
Technical Director
Deputy TD (EPA)
Associate TD
Scientific Staff
Research Associates
Statisticians
Information Managers
QA/QC Staff
Logistics Staff
Technicians
Support Staff
Total
1.0
0.5
.35
1.25
1.75
2.0
0.5
1.5
1.0
9.85
1.0
1.0
.35
2.0
1.5
1.3
1.0
1.0
1.5
10.65
1.0
1.0
.35
4.0
6.0
3.5
2.0
0.5
0.5
3.0
3.0
21.85
1.0
1.0
.35
7.0
5.0
5.0
2.5
1.0
1.0
10.0
4.0
37.85
1.0
1.0
.35
9.0
8.0
5.0
3.0
2.0
2.0
18.0
7.0
56.35
This does not include staff of the NASS who actually carry out the surveys.
Table 11.11. Expected Program Products
Product Years
91 92 93 94 95
Annual Statistical
Summaries (A)
Interpretative Reports (A)
Pilot Studies (A)
• Field Methods Manual (B)
QA/QC Reports (B)
Implementation Plans (B)
Journal Articles (B)
Presentations/Proceedings
(B)
1
1
Example
X
X
X
2
2
X
X
X
X
X
3
5
X
X
X
X
X
X
6
5
X
X
X
X
X
X
10
5
11 19
-------
LITERATURE CITED
Alfoldi, L. (1983) Movement and interaction of nitrates and pesticides in vegetation cover-soil
groundwater-rock systems. Environ. Geology 5: 19-25.
Allen, T.F.H. and Starr, T.B. (1982) Hierarchy: Perspectives for Ecological Complexity. U.
Chicago Press, Chicago.
Anderson, J.F., and Wojtas, M.A. (1986) Honey bees (Hymenoptera: Apidae) contaminated
with pesticides and polychlorinated biphenyls. /. Economic Entomology. 79:1200-1205.
Anderson, J.P.E. and Domsch, K.H. (1978) A physiological method for the quantitative
measurement of microbial biomass in soils. Soil Biology and Biochemistry 10: 215-221.
Andow, D.A. and Davis, D.P. (1989) Agricultural chemicals: Food and environment. In:
Food and Natural Resources ed. D. Pimentel and C.W. Hall. Academic Press, San Diego, pp.
191-234.
Auclair, A. (1976) Ecological factors in the development of intensive management ecosystems
in the midwestern United States. Ecology 57: 431-444.
Baker, K.F. and Cook, RJ. (1974) Biological Control of Plant Pathogens. American
Phytopathology Society, St. Paul, MN.
Bailey, R.G. (1983). Delineation of ecosystem regions. Environmental Management 7: 365-
373.
Ball, V.E. (1985) Output, input and productivity measurement in U.S. agriculture, 1948-79.
Am. J. Agr. Econ. (August): 475-486.
Barrett, G.W., Rodenhouse, N., and Bohlen, PJ. (1990) Role of sustainable agriculture in rural
landscapes. In Sustainable Agricultural Systems, ed. C.A. Edwards, R. Lai, P. Madden, R.H.
Miller and G. House. Soil and Water Conservation Society, Ankeny, Iowa, pp. 624-636.
Bedford, B.L. and Preston, E.M. (1988) Developing the scientific basis for assessing cumulative
effects of wetland loss and degradation on landscape functions: status, perspectives, and
prospects. Environmental Management 12: 751-771.
Benbrook, C.M. (1990) Society's stake in sustainable agriculture. In: Edwards, C.A., Lai, R.,
Madden, P., Miller R.H. and House, G. Sustainable Agricultural Systems. Soil and Water
Conservation Society. Ankeny, Iowa. pp. 68-76.
Beyer, W.N., and Gish, C.D. (1980). Persistence in Earthworms and Potential Hazards to Birds
of Soil Applied DDT, Dieldrin, and Heptachlor. J. Appl. Ecol. 17: 295-307.
L- 1
-------
Bowen, G.W. and Burgess, R.L. (1981) A Quantitative Analysis of Forest Island Pattern in
Selected Ohio Landscapes. Publication ORNL-TM-7759, Oak Ridge National Laboratory,
Environmental Sciences Division, Oak Ridge, TN.
Bird, S.L., Cheplick, J.M and Brown, D.S.. (1989). Preliminary testing, evaluation and
sensitivity analysis of the Terrestrial Ecosystem Exposure Assessment Model (TEEAM). U.S.
EPA, Athens, GA.
Bromberg, S.M. (1990) Identifying ecological indicators: An environmental monitoring and
assessment program. /. Air Waste Management Assoc. 40: 976-978.
Bromenshenk, J.J. (1988) Regional Monitoring of Pollutants with Honey Bees. In: Progress in
Environmental Specimen Banking., ed. S.A. Wise, R. Zeisler R, and G.M. Goldstein. U.S.
Department of Commerce, National Bureau of Standards, NBS Special Publication 740.
Bromenshenk, J.J. (1989) Terrestrial Invertebrate Surveys. In: Ecological Assessment of
Hazardous Waste Sites: A Field and Laboratory Reference, ed. W. Warren-Hicks, B.R.
Parkhurst, and S.S. Baker Jr. EPA/600/3-89/013, U.S. EPA, Washington, D.C.
Bromenshenk, J.J., and Preston, E.M. (1986) Public participation in environmental monitoring:
A means of attaining network capability. Environmental Monitoring and Assessment 6:35-47.
Bromenshenk, J.J., Carlson, S.R., Simpson, J.C., and Thomas, J.M. (1985) Pollution monitoring
of Puget Sound with honey bees. Science 221: 632-634.
Brooks, P.C., Heijen, C.E., Me Grath, S.P., and Vance, E.D. (1986) Soil microbial biomass
estimates in soils contaminated with metals. Soil Biology and Biochemistry 18: 383^388.
Burkart, M.R., Onstad, C.A. and Bubenzer, G.D. (1990) Research on agrichemicals in water
resources. EOS (Trans. Amer. Geophysical Union) July 17,1990, pp. 980, 981, 988.
Callahan, C.A. (1988) Earthworms as ecotoxicological assessment tools. In: Proceedings of
the International Conference on Earthworms in Waste and Environmental Management ed. C.A.
Edwards, and E.F. Neuhauser. Cambridge, England. SPB Academic Pub., The Hague, The
Netherlands, pp 295-301.
Campbell, C.L., Heck, W.W., and Moser, T.J. (1990). Indicator Strategy for Agroecosytems.
In: Carolyn T. Hunsaker and Dean E. Carpenter. Eds. Environmental Monitoring and
Assessment Program: Ecological Indicators. EPA/600/3-90/060. U.S. EPA, Office of
Research and Development, Washington, D.C., pp.8-1 to 8-11.
Campbell, C. L., Heck, W.W., Moser, T.J., Breckenridge, R.P., Hess. G.R. and Meyer, J.R..
(1990). Indicator Fact Sheets for Agroecosystems. In: Carolyn T. Hunsaker and Dean E.
Carpenter. Eds. Environmental Monitoring and Assessment Program: Ecological Indicators.
EPA/600/3-90/060. U.S. EPA, Office of Research and Development, Washington, D.C., pp. F-
1 to F-23.
L-2
-------
Canter, Larry, 1986. Environmental Impacts of Agricultural Production Activities. Lewis
Publishers, Inc. Chelsea, MI. 382 pp.
Carlberg, G.E., Ofstad, E.B. and Drangsholt, H. (1983) Atmospheric deposition of organic
micropollutants in Norway studied by means of moss and lichen analysis. Chemosphere 12:
341-356.
Can, D.E. (1966) Death of the Sweet Water. W.W. Norton, New York, NY.
Carsel, R.F., C.N. Smith, L.A. Mulkey, J.D. Dean and P.P Jowise. 1984. Users Manual for the
Pesticide Root Zone Model (PRZM): Release 1. U.S. Environmental Protection Agency,
Athens, GA. EPA-600/3-84-109.
Carter, M.R. (1986) Microbial biomass as an index for tillage-induced changes in soil biological
properties. Soil and Tillage Research 7: 29-40.
Coleman, D.C. and Hendrix, P.F. (1988) Agroecosystem processes. In Concepts of Ecosystem
Ecology: A Comparative View, ed. Pomeroy, L.R. and J.J. Alberts. Springer-Verlag, New York,
pp. 149-170.
Connell, D.W., and Miller, L.G.J. (1984) Chemistry and Ecotoxicology of Pollution. John
Wiley and Sons, Toronto.
Conway, G. R. (1986). Agroecosystem Analysis for Research and Development. Winrock
International Institute for Agricultural Development, Bangkok.
Corwin, D. L., Sorensen, M., and Rhoades, J.D. (1990). Delineating Areas of Salinity
Development on Irrigated Agriculture. Salinity Laboratory, Riverside CA.
Cotter, J. and Nealon, J. (1987) Area Frame Design for Agricultural Surveys. Area Frame
Section, Research and Applications Division, National Agricultural Statistics Service, USDA.
Council for Agriculture and Technology (CAST). (1990). Ecological Impacts of Federal
Conservation and Cropland Reduction Programs. Task Force Report No. 117. Ames.IA.
Council on Environmental Quality. 1981. Environmental Trends. Washington, DC.
Cox, G.W. (1984) The linkage of inputs to outputs in agroecosystems. In Agricultural
Ecosystems: Unifying Concepts, ed. R. Lowrance, B.R. Stinner and GJ. House. John Wiley and
Sons, New York, pp. 187-209.
Dean, J.D., Voos, K.A., Schanz, R.W., and Popenuck. B.P. 1989. Terrestrial Ecosystem
Exposure Assessment Model (TEEAM). EPA-600/3-88/038. U.S. EPA, Athens, GA.
Dean, J.D., Huyakorn, P.S., Donigian Jr, A.S., Voos, K.A., Schanz, R.W., Meeks, Y.J., and
Carsel, R.F. 1989. Risk of UnsaturatedlSaturated Transport and Transformation of Chemical
L-3
-------
Concentrations.(RUSTIC). Volume II. Users Manual. EPA-600/3-89/048b. U.S. EPA,
Athens, GA.
DeCola, L. (1989) Fractal analysis of a classified Landsat Scene. Photo grammetric Engineering
and Remote Sensing 55: 601-610.
Domsch, K.H. (1985) Influence of management on microbial communities in soil. In Microbial
Communities in Soil, ed. V. Jensen, A. Kjoller and L.H. Sorensen. Elsevier Applied Science
Publishers, London, pp. 355-367.
Domsch, K.H., Jagnow, G., and Anderson, T.-H. (1983) An ecological concept for the
assessment of side-effects of agrochemicals on soil microorganisms. Residue Reviews 86: 65-
105.
Duvick, D.N. (1984) Genetic diversity in major farm crops on the farm and in reserve.
Economic Botany 38: 161-178.
Ebing, W., Pflugmacher, J. and Hague, A. (1984) The earthworm as a key organism for the
measurement of soil contamination with foreign chemicals. Ber. Landwirtschaftwiss. 62:222-
255.
Edwards, C.A. and Lofty, J.R. (1977) Biology of Earthworms. 2nd Ed. John Wiley and Sons,
Inc., New York.
Elliott, E.T. and Cole, C.V. (1989) A perspective on agroecosystem science. Ecology 70:1597-
1602.
Ellis, E.G., Arickson, A.R. Wolcott, Zabik, M. and Leavitt, R. (1977). Pesticide Runoff Losses
From Small Watersheds in Great Lakes Basin. EPA-600/3-77-112. U.S. EPA, Athens, GA.
Evans, L.T. 1980. The Natural History of Crop Yield. American Scientist 68:388-397.
Flora, C.B. (1990a) Sustainablity of agriculture and rural communities. In: Francis, C.A.,
Flora, C.B., and L.D. King (Eds). Sustainable Agriculture in Temperate Zones. John Wiley and
Sons, N.Y. pp. 343-359.
Flora, C.B. (1990b) Policy issues and agricultural sustainability. In: Francis, C.A., Flora,
C.B., and L.D. King (Eds). Sustainable Agriculture in Temperate Zones. John Wiley and Sons,
N.Y. pp. 361-379.
Forman, R.T.T. and Baudry, J. (1984) Hedgerows and hedgerow networks in landscape
ecology. Environmental Management 8: 495-510.
Forman, R.T.T. and Godron, M. (1986) Landscape Ecology. John Wiley and Sons, New York.
L-4
-------
Franson, S.E. (1990). Data Confidentiality in the Environmental Monitoring and Assessment
Program (EMAP): Issues and Recommendations. US EPA Report 600/X-90/219. U.S. EPA,
Las Vegas, NV.
Freemark, K.E. and Merriam, H.G. (1986) Importance of area and habitat heterogeneity to bird
assemblages in temperate forest fragments. Biological Conservation 31: 115-42.
George, T.A. and Choate, J. (1989) A first look at the 1987 National Resources Inventory. /.
Soil and Water Conservation (Nov-Dec.): 555-556.
Gianessi, L.P. (1985) The Resources for the Future data base for nonpoint source policy
assessments.
EPA 440/5-85-001. U.S.EPA Office of Water Regulations and Standards, Washington, D.C.
Glooschenko, W.A. (1989) Sphagnum fiscum moss as an indicator of atmospheric cadmium
deposition across Canada. Environmental Pollution 57: 27-33.
Godron, M. and Forman, R.T.T. (1983) Landscape modification and changing ecological
characteristics. In Disturbance and Ecosystems, ed. H.A. Mooney and M. Godron, Springer-
Verlag, Berlin, pp. 12-28.
Gottfried, B.M. (1979) Small mammal populations in woodlot islands. American Midland
Naturalist 102: 105-12.
Gutierrez, RJ. and Carey, A.B. (Eds) (1985) Ecology and Management of the Spotted Owl in
the Pacific Northwest. USDA Forest Service GTR PNW-185.
Harris, L.D. (1988) The nature of cumulative impacts on biotic diversity of wetland vertebrates.
Environmental Management 12: 675-93.
Hawksworth, D.L. (1973) Mapping Studies. In: Air Pollution and Lichens, ed. B.W. Ferry,
M.S. Baddeley and D.L. Hawksworth. The Athlone Press.
Heck, W.W., Campbell, C.L., Moser, T.J., Rawlings, J.O., Hess, G.R., Breckenridge, R..B.,
Smith, C.N., Byers, G.E. and Finkner, A.L. (1989). EMAP-Agroecosystem Research Plan.
Informal Report
Heck, W.W. (1990) Impacts of air pollution on agriculture in North America. In: Ecological
Risks: Perspectives from Poland and the United States, ed. Grodzinshi, W., Cowling, E., and
Breymeyer, A., National Academy Press, Washington, D.C., pp. 171-195.
Helliwell, D.R. (1976) The effects of size and isolation on the conservation value of wooded
sites in Britain. /. Biogeography 3: 409-16.
L-5
-------
Henderson, M.T., Merriam, G., and Wegner, J. (1985) Patchy environments and species
survival: chipmunks in an agricultural mosaic. Biological Conservation. 31: 95-105.
Hileman, B. (1990) Alternative Agriculture. Chemical and Engineering News 68: 26-40.
Huff, F.A. and J.C. Neill. (1982) Effects of natural climatic fluctuations on the temporal and
spatial variation in crop yields. /. Applied Meteorology 21: 540-550.
Humanik, F.J., Smolen, M.D., Spooner, J. Dressing, S.A. and Maas, R.P. (1984) Best
Management Practices for Agricultural Nonpoint Source Control, Voil-TV. North Carolina
Agricultural Extension Service, Raleigh, NC.
Hunsaker, C.T. and Carpenter, D.E., Eds. Environmental Monitoring and Assessment Program:
Ecological Indicators. EPA/600/3-90/060. U.S. EPA, Office of Reserach and Development,
Washington, D.C.
Ireland, M.P. (1979) Metal accumulation by the earthworms Lumbricus rubellus, Dendrobaena
veneta and Eiseniella tetraedra living in heavy metal polluted sites. Environmental Pollution
19: 201-206.
Jahn, L.R. (1988) The potential for wildlife habitat improvements. /. Soil Water Conservation
43: 67-69.
Jenkinson, D.S. and Powelson, D.S. (1976) The effects of biocidal treatments on metabolism in
soil I. Fumigation with chloroform. Soil Biology and Biochemistry 8: 167-177.
Johnson, H.P. and Baker, J.L. (1982) Field-to-Stream Transport of Agricultural Chemicals
and Sediment in an Iowa Watershed: Part I. Data Base for Model Testing (1976-1978). EPA-
600/3-82-032. U.S. EPA, Athen, GA.
Jones, K.C., Stratford, J.A., Tidridge, P., Waterhouse, K.S. and Johnston, A.E. (1989a)
Polynuclear aromatic hydrocarbons in an agricultural soil: long-term changes in profile
distribution. Environmental Pollution 56: 337-351.
Jones, K.C., Stratford, J.A., Waterhouse, K.S., Furlong, E.T., Giger, W., Kites, R.A., Schaffner,
C. and Johnston, A.E. (1989b) Increases in the polynuclear aromatic hydrocarbon content of an
agricultural soil over the last century. Environ. Sci. Technol. 23.
Karr, J.R. and Schlosser, I.J. (1978) Water resources and the land-water interface. Science 20:
229-34.
Kelly, J.R. and M.A. Harwell. (1988) Indicators of ecosystem response and recovery. In
Ecotoxicology: Problems and Approaches, ed. S.A. Levin, M.A. Harwell, J.R. Kelly and K.D.
Kimball. Springer-Verlag, New York. pp. 9-35.
L-6
-------
Klepper, R., W. Lockeretz, B. Commoner, M. Gertler, S.Fast, D. O'Leary and R. Blobaum.
1977. Economic performance and energy intensiveness on organic and conventional farms
in the Corn Belt: A preliminary comparison. Amer. J. Agr. Econ. X: 1-12.
Knapp, C.M., Marmorek, D.R., Baker, J.P., Thornton, K.W., Klopatek, J.M., and Charles, D.P.
(1990). The Indicator Development Strategy for the Environmental Monitoring and Assessment
Program. Draft Report. Environmental Research Laboratory, U.S. EPA, Corvallis, OR.
Krummel, J.R., Gardner, R.H., Sugihara, G., O'Neill, R.V., and Coleman, P.R. (1987)
Landscape patterns in a disturbed environment. Oikos 48: 321-324.
Kutz, F.W. and Linthurst, R.A. (1990). A systems-level approach to environmental assessment.
Toxicological and Environmental Chemistry 28: 105-114.
Laflamme, R.E. and Kites, R.A. (1978) The global distribution of polycyclic aromatic
hydrocarbons in recent sediments. Geochim. Cosmochim Acta 42: 289-303.
Lam, N. Siu-N. (1990) Description and measurement of Landsat TM images using fractals.
Photogrammetric Engineering and Remote Sensing 56: 187-195.
Langdale, G.W. and R. Lowrance. (1984) Effects of soil erosion on agroecosystems of the
humid United States. In Agricultural Ecosystems: Unifying Concepts, ed. R. Lowrance, B.R.
Stinner and G.J. House. John Wiley and Sons, New York, pp. 133-144.
Laurance, J.A. and Grieteur. (1984) Development of a Biological Air Quality Indexing System.
Minnesota Environmental Quality Board, St. Paul, MN.
Leonard, R.A., Knisel, W.G., and Still, D.A. 1987. GLEAMS: Groundwater loading effects of
agricultural management systems. Transactions of the American Society of Agricultural
Engineers 30: 1403-1418.
Logan, T.J. (1990) Sustainable agriculture and water quality. In: Sustainable Agricultural
Systems,.ed. Edwards, C.A., R. Lai, P. Madden, R.H. Miller and G. House. Soil and Water
Conservation Society, Ankeny, IA. pp. 582-613.
Lowrance, R., Hendrix, P.P., and Odum, E.P. (1986). A hierarchal approach to sustainable
agriculture. Am. J. Alternative Agriculture I: 169-173.
Lynch, J.F. and Whigham, D.F. (1984) Effects of forest fragmentation in Maryland, USA.
Biological Conservation 28: 287-324.
MacArthur, R.H. and Wilson, E.G. (1967) The Theory of Island Biogeography. Princeton Univ.
Press, Princeton.
L-7
-------
MacClintock, L., Whitcomb, R.F., and Whitcomb, B.L. (1977) Evidence for the value of
corridors and minimization of isolation in preservation of biotic diversity. American Birds 31:6-
16.
Madden, J.P. and Dobbs, T.L. (1990) The role of economics in achieving low-input farming
systems. In: Edwards, C.A., Lai, R., Madden, P., Miller, R.H., and House, G. (Eds) Sustainable
Agricultural Systems. John Wiley and Sons, N.Y.
Magurran, A. (1988). Ecological Diversity and its Measurement. Princeton U. Press, Princeton,
N.J.
Malik, K.A. and Azam, F. (1980) Effect of salinity of ^C-labeled microbial biomass and its
contribution to soil organic matter. Pak. J. Bot. 12: 117-127.
Mandelbrot, B.B. (1977) Fractals: Form, Chance and Dimension. Freeman, San Francisco.
Manning, W.J., and Feder,.W.A. (1908). Biomonitoring Air Pollutants with Plants. Applied
Science, Ltd., London.
Maranto, G. (1985). A once and future desert Discover, June 1985: 32-39.
Mayrose, V.B., Bache, D.H., Libal, G. (1985). Pork Industry Handbook. NC Ag. Extension
Service, Raleigh, NC.
McKee, I.E. and Wolf, H.W. (1963) Water Quality Criteria. California State Water Resources
Control Board. Reprinted in 1971.
Meyer, J.R., Peck, S., Rawlings, J.O., Campbell, C.L. (1990). Environmental Monitoring and
Assessment Program. Annual Statistical Summary Report on Agroecosy stems: An Example.
EPA 600/X-90/208. U.S. EPA, Environmental Monitoring Systems Laboratory, Las Vegas,
NV.
Meyer, J.R., Campbell, C.L., Moser, T.J., Hess, G.R., Rawlings, J.O., Peck, S. and Heck, W.W.
(1991) Indicators of the ecological status ofagroecosystems. In: Proceedings of the
International Symposium on Ecological Indicators. October 1990, Ft. Lauderdale, FL (in press)
Milne, B.T. (1988) Measuring the fractal geometry of landscapes. Applied Mathematics and
Computation 27: 67-79.
Monk, C. (1966) Ecological importance of root/shoot ratios. Bulletin of the Torry Botanical
Club 93: 402-406.
Moreau, D.H. (1990) Draft Final Report of Recommendations of the Governor's Blue Ribbon
Panel on Environmental Indicators. Office of the Governor, Raleigh, N.C.
L-8
-------
Morse, R.A., Culliney, T.W., Gutenmann, W.H., Littman, C.B., and Lisk, D.J. (1987)
Polychlorinated biphenyls in honey bees. Bulletin of Environmental Contamination and
Toxicology 38: 271-276.
National Research Council (1989). Alternative Agriculture. National Academy Press.
Washington, D.C.
National Research Council (1990) Managing Troubled Waters: The Role of Marine
Environmental Monitoring. National Academy Press, Washington, D.C.
Naveh, Z. and Lieberman, A.S. (1984). Landscape Ecology: Theory and Application. Springer
Verlag, N.Y.
Nealon, J.P. (1984). Review of Multiple and Area Frame Estimators. SF and SRB Report No.
80. USDA/NASS.
Neuhauser, E.F., Malecki, M.R, and Loehr, R.C. (1984). Growth and Reproduction of the
Earthworm, Eisenia foetida after Exposure to Sublethal Concentrations of Metals.
Pediobiologia. 27:89-97.
Neuhauser, E.F., and Callahan, C.A. (1990). Growth and Reproduction of the Earthworm
Eisenia fetida Exposed to Sublethal Concentrations of Organic Chemicals. SoilBiol. Biochem.
22:175- 179.
Noss, R.F. (1983) A regional landscape approach to maintain diversity. BioScience 33: 700-6.
Noss, R.F. (1987) Corridors in real landscapes: A reply to Simerloff and Cox. Conservation
Biology 1: 159-164.
Noss, R.F. and Harris, L.D. (1986) Nodes, Networks and MUMs: Preserving Diversity at All
Scales. Environmental Management 10: 299-309.
Odum, E.P. (1989) Input management of production systems. Science 243: 177-182.
Office of Technology and Assessment (OTA) (1982) Impacts of Technology on U.S. Cropland
and Rangeland Productivity. OTA, U.S. Govt. Printing Office, Washington, D.C.
Office of Technology and Assessment (1983) Water-related Technologies for Sustainable
Agriculture in U.S. Arid/Semiarid Lands. OTA, U.S. Govt. Printing Office, Washington, D.C.
Office of Technology and Assessment (1986) Technology, Public Policy and the Changing
Structure of American Agriculture. OTA, U.S. Govt. Printing Office, Washington, D.C.
Olson, G.L. and Breckenridge, R.P. (1990) Assessing Agroecosystem Sustainability- An
Integrated Approach. Proceedings of the International Symposium on Ecological Indicators,
October 16-19,1990, Fort Lauderdale, FL, USA (in press).
L-9
-------
Olson, G.L., Breckenridge, R.P. and Wiersma, G.B. (1990) Assessment of federal databases to
evaluate agroecosystem productivity. Idaho National Engineering Laboratory. Informal Report
EGG-CEMA-8924. February 1990.
Omernik, J.M., Abernathy, A.R., and Male" L.M. (1981) Stream nutrient levels and proximity
of agricultural and forest land to streams: some relationships. Journal of Soil and Water
Conservation 36: 227-31.
Omernick, J.M. 1987. Ecoregions of the conterminous United States. Annals of the Association
of American Geographers 77: 118-125.
O'Neill, R.V., Ausmus, B.S., Jackson, D.R., Van Hook, R.I., Van Voris, P., Washburne, C., and
Watson, A.P. (1977) Monitoring terrestrial ecosystems by analysis of nutrient export. Water,
Air and Soil Pollution 8: 271-277.
O'Neill, R.V., DeAngelis, D.L., Waide, J.B., and Allen, T.F.H. (1986) A Hierarchical Concept
of Ecosystems. Monographs in Population Biology 23, Princeton University Press, Princeton.
O'Neill, R.V., Krummel, J.R., Gardner, R.H., Sugihara, G., Jackson, B., DeAngelis, D.L., Milne,
B.T., Turner, M.G., Zygmunt, B., Christensen, S., Dale, V.H., and Graham, R.L. (1988)
Indices of landscape pattern. Landscape Ecology 1: 153-162.
Paarlberg, D. (1980). Farm and Food Policy. University of Nebraska Press, Lincoln, NE.
Patton, D.R. (1975) A Diversity Index for Quantifying Habitat "Edge". Wildlife Society Bulletin
3: 171-173.
Paul, E.A. and Robertson, G.P. (1989) Ecology and the agricultural sciences: A false
dichotomy? Ecology 70: 1594-1597.
Peterjohn, W.T. and Correll, D.L. (1984) Nutrient dynamics in an agricultural watershed:
Observations on the role of a riparian forest. Ecology 65: 1466-1475.
Pilegaard, K.L. (1979) Heavy metals in bulk precipitation and transplanted Hypogymnia
physodes and Dicranoweisia cirrata in the vicinity of a Danish steelworks. Water, Air, and Soil
Pollution 11:77-91
Pimentel, D. (1980) Handbook of Energy Utilization in Agriculture. CRC Press, Inc: Boco
Raton, Florida.
Pimentel, D. and Edwards, C.A. (1982) Pesticides and ecosystems. BioScience 32: 595-600.
Pimentel, D. and Levitan, L. (1986) Pesticides: Amounts applied and amounts reaching pests.
BioScience 36: 86-91.
L-10
-------
Pimentel, D., Culliney, T.W., Butler, I.W., Reinemann DJ. and Beckman, K.B. (1989) Low-
input sustainable agriculture using ecological management practices. Agriculture, Ecosystems
and Environment 27: 3-24.
Postel, S. (1985) Managing Fresh Water Supplies. In: The State of the World. The
Worldwatch Institute. Washington, D.C.
Posthumus, A.C. (1976). The Use of Higher Plants as Indicators of Air Pollution in the
Netherlands. In: Proceedings of the Kuopio Meeting on Plant Damages Caused by Air Pollution.
L. Karenlampi (ed.), Kuopio, pp. 115-120.
Rabatin S.C. and Stinner, B.R. (1989) The significance of vesicular-arbuscular mycorrhizal
fungal-soil macroinvertebrate Alterations in agroecosystems. Agriculture, Ecosystems and
Environment 27: 195-204.
Rapport, DJ. (1989) What constitute ecosystem health? Perspectives in Biology and Medicine.
33: 120-132.
Raw, F. (1959). Estimating Earthworm Populations by Using Formalin. Nature. 184:1661-1662.
Reichelderfer, K. and Phipps, T.T. (1988) Agricultural Policy and Environmental Quality
National Center for Food and Agricultural Policy, Resources for the Future. Washington,
D.C.
Renard, K.G., Asce, F. and Simanton, J.R. (1990) Application of RUSLE to Rangelands. In:
Watershed Planning and Analysis, eds. R.B. Riggins, E. Bruce Jones and P. A. Richard.
American Society of Civil Engineers, N.Y., N.Y., pp. 164-173.
Risser, P.O. (1985) Toward a holistic management perspective. BioScience 35: 414-418.
Risser, P.O., Karr, J.R., and Forman, R.T.T. (1983) Landscape Ecology: Directions and
Approaches. Illinois Natural History Survey Special Pub. No. 2, Champaign, IL.
Ritchie, J.C. and McHenry, J.R. (1990) Application of radioactive fallout Cesium-137 for
measuring soil erosion and sediment accumulation rates and patterns. /. Environmental Quality
19: 215-233.
Ross, H.B. (1990) On the use of mosses (Hylocomium splendens and Pleurozium schreberi) for
estimating atmospheric trace metal deposition. Water, Air and Soil Pollution 50: 63-76.
Roth, R.R. (1976) Spatial heterogeneity and bird species diversity. Ecology 57: 773-82.
Ruppel, R.F., and C.W. Laughlin. (1977). Toxicity of Some Soil Pesticides to Earthworms. /.
Kansas Entomol. Soc. 50: 113-118.
L-ll
-------
Schaffer, D.J., Herricks, E.E. and Kerster, H.W. (1988) Ecosystem health: I. Measuring
ecosystem health. Environmental Management 12: 445-455.
Schlosser, I.J. and Karr, J.R. (1981) Water quality in agricultural watersheds: impact of riparian
vegetation during base flow. Water Resources Bulletin 17: 233-240.
Schroeder, R.L. (1986) Habitat Suitability Index Models: Wildlife Species Richness in
Shelterbelts. U.S. FWS Biological Report 82 (10.128), U.S. Fish and Wildlife Service,
Washington, D.C.
Shannon, C.E. and Weaver, W. (1949) The Mathematical Theory of Communication. Univ.
Illinois Press, Urbana.
Sharpe, D.D. Lieth, H. and Whigham, D. (1975) Assessment of regional productivity in North
Carolina. In: Primary Productivity of the Biosphere, ed. H. Lieth and R.H. Whittaker.
Springer-Verlag, New York, pp. 131-146.
Sharpe, D.D., H. Leith, G. R. Noogle, H.D. Gross. (976). Agricultural and Forest Primary
Productivity in North Carolina 1972-1973. North Carolina Agricultural Experiment Station
Tech. Bull. 241.
Short, H.L. (1984) Habitat Suitability Index Models: The Arizona Guild and Layers of Habitat
Model. FWS/OBS-82/10.70, U.S. Fish and Wildlife Service, Washington, D.C.
Short, H.L. (1985) Management Goals and Habitat Structure. In Riparian Ecosystems and Their
Management: Reconciling Conflicting Uses, USD A Forest Service GTR RM-120, ed. R.R.
Johnson, C.D. Ziebell, D.R. Patton, P.P. Folliott and R.H. Hamre, USDA Forest Service,
Washington, DC, pp. 257-262.
Short, H.L. (1987) A habitat structure model for natural resource management. /.
Environmental Management. 27: 289-305.
Short, H.L. (1989) A wildlife habitat model for predicting effects of human activities on nesting
birds. In Freshwater Wetlands and Wildlife, CONF-8603101, DOE Symposium Series No. 61,
ed. R.R. Sharitz and J.W. Gibbons, U.S. Dept. of Energy, Office of Scientific and Technical
Information, Oak Ridge, TN, pp. 957-975.
Short, H.L. (1990) The use of the Habitat Linear Classification System (HLCS) to inventory and
monitor wildlife habitat. Unpublished manuscript. U.S. Fish and Wildlife Service, Arlington,
VA.
Short, H.L. and Williamson, S.C. (1986) Evaluating the structure of habitat for wildlife. In
Wildlife 2000: Modeling Habitat Relationships ofTerretrial Vertebrates, ed. J. Verner, M.L.
Morrison and C.J. Ralph, Univ. Wisconsin Press, Madison, pp. 97-104.
L-12
-------
Skidmore, E.L. (1986) Wind Erosion. In: Soil Erosion Research Methods, ed. R. Lai. Soil and
Water Conservation Society, Ankeny, IA, pp. 203-233.
Simpson, E.H. (1949) Measurement of diversity. Nature 163: 688.
Smirle, M.J., Winston, M.L. and Woodward, K.L. (1984) Development of a sensitive bioassay
for evaluating sublethal pesticide effects on the honey bee (Hymenoptera: Apidae) /. Economic
Entomology 77: 63-67.
Smith, C.N., Parrish, R.S., and Brown, D.S. (1990) Conducting field studies for testing
pesticide leaching models. Intern. J. Environ. Anal. Chem. 39: 3-21.
Smith, C.N., Leonard, R.A., Langdale, G.W. and Bailey, G.W. (1978). Transport of
Agricultural Chemicals from Small Upland Piedmont Watersheds. EPA-600/3-78-056. U.S.
EPA,. Athens, GA.
Smith, J.L. and Paul, E.A. (1990) The significance of soil microbial biomass estimations. In:
Soil Biochemistry, Vol. 6. ed. J.-M. Bollag and G. Stotsky. Barcel Dekker, Inc. New York, pp.
357-396.
Soule, J., Carre, D., and Jackson, W. (1990). Ecological impact of modern agriculture. In:
Agroecology. ed. C.R. Carroll, J.H. Vandermeer and P. Rosset. McGraw-Hill Publ. Co., New
York, pp. 165-188.
Stanley, T.W. and Verner, S.S. (1985). The U.S. Environmental Protection Agency's Quality
Assurance Program. In: Quality Assurance for Environmental Measurements. ASTM STP 867:
12-19. American Society for Testing and Materials, Philadelphia, PA.
Stone, J.J., Lopes, V.L., Asce, M. and Lane, L.J. (1990) Water erosion prediction project
(WEPP) watershed model: Hydrologic and erosion calculations. In: Watershed Planning and
Analysis, eds. R.B. Riggins, E. Bruce Jones and P. A. Richard. American Society of Civil
Engineers, N.Y., N.Y., pp. 184-190.
Stokes, B. 1983. Water shortages: The next energy crisis. The Futurist (April): 38-47.
Strachan, W.M.J. and Glooschenko, W.A. (1988) Moss bags as monitors of organic
contamination in the atmosphere. Bulletin Environ. Contam. Toxicol. 40: 447-450.
Strong, D., Lawton, J. and Southwood, T.R.E. (1984) Insects on Plants: Community Patterns
and Mechanisms. Blackwell, Oxford, England.
Suter, G.W. H. (1990). Endpoints for regional ecological risk assessments. Environmental
Management 14: 9-23.
Tangley, L. (1986) Crop productivity revisited. BioScience 35: 142-147.
L-13
-------
Taylor, J.K. 1987. Quality Assurance of Chemical Measurements. Lewis Publishers, Chelsea,
MI.
Temple, S.A. (1986) Predicting Impacts of Habitat Fragmentation on Forest Birds: A
Comparison of Two Models. In Wildlife 2000: Modeling Habitat Relationships of Terrestrial
Vertebrates, ed. J. Verner, M.L. Morrison and C.J. Ralph, U. Wisconsin Press, Madison, pp.
301-304.
The Conservation Foundation. (1986) Agriculture and the Environment in a Changing World
Environment. The Conservation Foundation, Washington, D.C.
Thomas, W. (1979) Monitoring organic and inorganic trace substances by epiphytic mosses: A
regional pattern of air pollution. Trace Subst. Envir. Health 12: 285-289.
Thomas, W. (1986) Representativity of mosses as biomonitor organisms for the accumulation
of environmental chemicals in plants and soils. Ecotoxicology and Environmental Safety 11:
339-346.
Tiner, R.W. (1984) Wetlands of the U.S., Current Status and Recent Trends. U.S. Fish and
Wildlife Service. National Wetlands Inventory. U.S. Government Printing Office.
Washington, D.C.
Tingey, D.T. (1989) Bioindicators in air pollution research: Applications and constraints. In:
Biologic Markers of Air Pollution Stress and Damage in Forests. National Academy Press,
Washington, D.C.
Treshow, M. (1984) Air Pollution and Plant Life. John Wiley and Sons, Ltd., London.
Turner, M.G. (1987) Land use changes and net primary production in the Georgia, USA,
landscape: 1935-1982. Environmental Management 11: 237-247.
Turner, M.G. (1989) Landscape ecology: The effect of pattern on process. Annual Review of
Ecological Systematics 20: 171-197.
U.S. Department of Agriculture (1985) 1986 Fact Book on Agriculture. USDA Misc. Publ.
1063. Washington, D.C.
U.S. Department of Agriculture (1988). Agricultural Statistics 1988. U.S. Government
Printing Office, Washington, DC.
U.S. Department of Agriculture (1989) The Second RCA Appraisal. Soil Water and Related
Resources on Nonfederal Land in the United States. Analysis of Conditions and Trends. USDA
Misc. Publ. 1482. Washington, D.C.
U.S. Department of Agriculture, Economics Research Service (1984) U.S. Cropland,
Urbanization and Land Ownership Patterns. USDA/ERS, Washington, D.C.
L-14
-------
U.S. Department of Agriculture, Economic Research Service (1989) Production and Efficiency
Statistics. USDA/ERS, Washington D.C.
U.S. Department of Agriculture, Economic Research Service (1990) Agricultural Resources:
Inputs and Outlook Report. AR-17, USDA/ERS, Washington, D.C.
U.S Department of Agriculture, National Agricultural Statistics Service (1990) Agricultural
Statistics. USDA/NASS, Washington, D.C.
U.S. Department of Agriculture, National Soil Erosion-Soil Productivity Research Planning
Committee (J.R. Williams, Chair) (1981) Soil erosion effects on soil productivity: A Research
Perspective. /. of Soil and Water Conservation 36: 82-90.
U.S. Department of Agriculture, Soil Conservation Service. (USDA/SCS) (1981). Land
Resource Regions and Major Land Resource Areas. USDA/SCS Agricultural Handbook 296.
U.S. Government Printing Office, Washington, D.C.
U.S. Department of Agriculture, Soil Conservation Service. (USDA/SCS) (1987) 1987National
Resources Inventory, Summary Report. Statistical Bulletin No. 790. Iowa State University
Statistical Laboratory, Ames, IA.
U.S. Department of Commerce (DOC), Bureau of the Census (1990) 1987 Census of
Agriculture^ U.S. Govermentt Printing Office, Washington, D.C.
U.S. Environmental Protection Agency (1976) Quality Criteria for Water U.S. EPA,
Washington, D.C.
U.S. Environmental Protection Agency (1978) Methodologies for valuation of agricultural crop
yield changes: A review. EPA 600/5-78/018. U.S. EPA., Washington, D.C.
U.S. Environmental Protection Agency (1990) Environmental Monitoring and Assessment
Program: Overview. EPA/600/9-90/001. U.S. El x, Office of Modeling, Monitoring Systems
and Quality Assurance. Washington, D.C.
U.S. Geological Service (1977) National Handbook of Recommended Methods for Water Data
Acquisition. Office of Water Data Coordination. USGS, Washington, D.C.
U.S. Geological Service (1986-1987) Water Resources Investigations Reports. Reconnaissance
Investigations of Water Quality, Bottom Sediment and Biota Associated with Irrigation
Drainage. Lower Rio Grande Valley and Laguna Atascosa National Wildlife Refuge, Texas.
Report #87-4277. USGS, Washington, D.C. (see also reports #87-4243; #88-4001; #888-4002:
#87-4244)
U.S. Geological Service (1987) Water Resources Investigations Report #88-4011. USGS,
Washington, D.C
L 15
-------
VanHook, R.I. (1974). Cadmium, Lead and Zinc Distributions between Earthworms and Soils:
Potentials for Biological Accumulation. Bulletin of Environmental Contamination & Toxicology.
12:509-512.
Villeneuve, J. and Holm, E. (1984) Atmospheric background of chlorinated hydrocarbons
studied in Swedish lichens. Chemosphere 13: 1133-1138.
Wallwork-Barber, M.K., Ferenbaugh, R.W. and Gladney, E.S. (1982) The use of honey bees as
monitors of environmental pollution. American Bee Journal 122: 770-772.
Weaver, M. and Kellman, M. (1981) The effects of forest fragmentation on woodlot tree biotas
in southern Ontario. /. Biogeography. 8: 199-210.
Weinberg, A.C. (1990) Low-input agriculture reduces nonpoint source pollution. /. Soil Water
Conservation 45: 48-49.
Weller, M.W. (1988) Issues and approaches in assessing impacts on waterbird habitat in
wetlands. Environmental Management 12: 695-701.
Werner, M.R. and Dindal, D.L. (1990) Effects of conversion to organic agricultural practices on
soil biota. American J. of Alternative Agriculture 5: 24-32.
Williams, J.R. (1990) Social traps and incentives: Implications for low-input, sustainable
agriculture. /. Soil Water Conservation 45: 28-30.
Williams, J.R., Renard, K.G. and Dyke, P.T. (1983) EPIC: A new method for assessing
erosion's effect on soil productivity. /. Soil and Water Conservation (Sept-Oct): 381-383.
Wilson, LJ. (1987) Iowa Groundwater Protection Strategy and the Iowa Groundwater
Protection Act 1987. Environmental Protection Commission, Iowa Department of Natural
Resources.
Wischmeier, W.H. and Smith, D.D. (1978) Predicting Rainfall Erosion Losses-A Guide to
Conservation Planning. Agriculture Handbook 537, U.S. Department of Agriculture,
Washington, D.C.
Wright, M.A., and A. Stringer. (1980). Lead, Zinc and Cadmium Content of Earthworms from
Pasture in the Vicinity of an Industrial Smelting Complex. Environmental Pollution. 23:313-
321.
Yahner, R.H. (1988) Changes in wildlife communities near edges. Conservation
Biology 2: 333-339.
L-16
-------
Yeates, G.W. and Coleman, D.C. 1982. Role of nematodes in decomposition. In: D.W.
Freckman, ed. Nematodes in Soil Ecosystems. University of Texas Press, Austin. Pp.
55-80.
Young, R.A., Onstad, C.A., Bosch, D.D. and Anderson, W.P. (1989) AGNPS: A nonpoint-
source pollution model for evaluating agricultural watersheds. J. Soil Water Conservation
(March-April): 168-173.
L-17
-------
APPENDIX 1
AGROECOSYSTEM RESOURCE GROUP MEMBERS
Steering Committee
Walter W. Heck, Chairman
USDA/ARS Air Quality Program
1509 Varsity Drive
Raleigh, NC 27606
(919)737-3311
FTS 672-4069
E-Mail EPA8250
Fax (919) 737-3593
Robert Bass
USDA-NASS
Room 4151 South Bldg.
14th Independence, SW
Washington, DC 20250-2000
(202) 447-6170
FTS 447-6170
E-Mail
Fax (202) 382-0507
Robert P. Breckenridge
Idaho National Engineering Lab
P.O. Box 16258
Idaho Falls, ID 83415
(208) 526-0757
FTS 583-0757
E-Mail EPA8206
Fax (208) 526-0603
Roy E. Cameron
Lockheed Engineering
& Science Co.
Environmental Programs
1050 E. Flamingo Road
Las Vegas, NV 89119
(702) 734-3318
E-Mail EPA8271
Fax (702) 796-1084
C. Lee Campbell, Assoc. Chairman
Department of Plant Pathology
NC State University
Box 7616
Raleigh, NC 27695-7616
(919) 737-2751
E-Mail EPA8250 (Heck)
Fax (919) 737-7716
Tom Moser
NSI Technology Services Corp.
200 Southwest 35th Street
Corvallis, OR 97333
(503) 757-4666
FTS 420-4666
E-Mail EPA8414
Fax (503) 757-4335
John O. Rawlings
Department of Statistics
NC State University
Box 8203
Raleigh, NC 27695-8203
(919) 737-2535
E-Mail EPA8250 (Heck)
Fax (919) 737-3593 (Heck)
Charles N. Smith
U.S. EPA-ERL
College Station Road
Athens, GA 30613
(404) 546-3175
FTS 250-3175
E-Mail EPA8431
Fax (404) 546-3340
Al 1
-------
Exofficio (Administration)
Bruce Jones
U.S. EPA
EMSL-VL, BAD
P. O. Box 93478-3478
Las Vegas, NV 89193-3478
(702) 798-2671
FTS 545-2671
E-Mail EPA8272
Fax (702) 798-2637,2638 or 2654
Ann Pitchford
U.S. EPA
EMSL-LV, BAD
P. O. Box 93478
Las Vegas, NV 89193-3478
(702) 798-2219
FTS 545-7438
E-Mail
Fax (702) 798-2454 or 2221
Members (Current)
Jerry Byers
Lockheed Engineering
& Sciences Co.
Environmental Programs
1050 E. Flamingo Road, Suite 200
Las Vegas, NV 89119
(702) 734-3337
E-Mail EPA8231
Fax (702) 796-1084
George Hess
Air Quality Program
NC State University
1509 Varsity Drive
Raleigh, NC 27606
(919)737-3311
FTS 672-4069
E-Mail EPA8161
Fax (919) 737-3593
Jay Giles
NSI Technology Services Corp.
200 Southwest 35th Street
Corvallis, OR 97333
(503) 757-4719
FTS 420-4719
E-Mail EPA8410
Fax (503) 757-4335/FTS 420-4335
Virginia Lesser
906 Park Ridge Drive
Durham, NC 27713
(919) 962-7016 (office)
(919) 493-8371 (home)
Fax (919) 737-3593
Craig Hayes
Department of Agric. Statistics
NC Department of Agriculture
1 West Edenton Street
P. O. Box 27761
Raleigh, NC 27601
(919) 733-6333
FTS 672-4394
Fax (919) 856-4139
Julie Meyer
NSI Technology Services Corp.
NC State University
1509 Varsity Drive
Raleigh, NC 27606
(919)737-3311
FTS 672-4069
E-Mail EPA8192
Fax (919) 737-3593
Al 2
-------
Members Continued
Exoffico
Deborah Neher
Air Quality Program
NC State University
1509 Varsity Drive
Raleigh, NC 27606
(919)737-3311
FTS 672-4069
E-Mail EPA8192 (Meyer)
Fax (919) 737-3593
Karl Hermann
NSI Technology Services Corp.
P. O. Box 12313
2 Triangle Drive
RTP, NC 27709
(919) 541-4119
FTS
E-Mail EPA8255
Fax
Gail Olson
Idaho National Engineering Lab
P.O. Box 16258
Idaho Falls, ID 83415-2213
(208) 526-1870
FTS 583-1870
E-Mail EPA8206 (Breckenridge)
Fax (208) 526-0603
Douglas G. Lewis
DEHNR
P.O. Box 27687
Raleigh, NC 27611-7687
(919) 733-6376
FTS
E-Mail
Fax (919) 733-2622
Steve Peck
Department of Statistics
NC State University
1509 Varsity Drive
Raleigh, NC 27606
(919)737-3311
FTS 672-4069
E-Mail EPA8250 (Heck)
Fax (919) 737-3593
Susan Spruill
Department of Statistics
NC State University
1509 Varsity Drive
Raleigh, NC 27606
(919)737-3311
FTS 672-4069
E-Mail EPA8192 (Meyer)
Fax (919) 737-3593
Consultant
Alva L. Finkner
5425 Ironwood Lane
Raleigh, NC 27613
(919) 787-1483
Fax (919) 737-3593 (Heck)
Al 3
-------
AGROECOSYSTEM
Organizational Structure
February 1991
EMAP Management
Team
Bruce Jones (EPA)
(Program Director)
EMAP
Technical Directors
EMAP
Tech. Coordinators
T
Ann Pitchford (EPA)
(Adm. Director)
Walter W. Heck
(USDA/ARS)
Technical Director
Julie Meyer (METI)
RTF-EPA
(Indicator Development)
Lee Campbell (NCSU)
Associate Director
(Indicators/Logistics)
George Hess (NCSU)
(Indicator Development)
Vacant (NCSU)
Information Manager
Deborah Neher (NCSU)
(Indicator Development)
Beth Sherill (NCSU)
(Indicator Testing)
John O. Rawlings (NCSU)
Statistical Design
(Sampling/Design, Stat.)
Susan Spruill (NCSU)
Data Management
(General Consultant)
Steve Peck (NCSU)
Statistical Design
(Data Analysis)
Virginia Lesser (UNC-CH)
Statistical Design
(Special Stat. Prob.)
Alva Finkner
Statistical Design
(Consultant)
Robert Bass (USDA/NASS)
National NASS Liaison
(Design/Implementation)
Craig Hayes (USDA/NASS)
N.C. NASS Liaison
(Design/Implementation)
Vacant (USDA/SCS)
SCS Liaison
Douglas Lewis (NC DEHNR)
N.C. Blue Ribbon Panel
Exofficio
Al 4
Clara B. Edwards (NCSU)
Adm. Assistant
Adm. Support (NCSU)
(WP/Clerical/Library)
Tom Moser (METI)
Corvallis
(Indicators/Gen.)
Bob Breckenridge (INEL)
Idaho Falls
(Indicator Development)
_L
Gail Olson (INEL)
Idaho Falls
(Indicator Development)
Jerry Byers (LESC)
Las Vegas
(QA/Logistics)
Roy Cameron (LESC)
Las Vegas
(Agricultural Consult.)
Charlie Smith (EPA)
Athens
(Indicator Development)
Karl Hermann (METI)
Raleigh
(CIS Specialist)
ARS Research Staff
(Indicator Testing)
(Consulting)
2/1/91
-------
APPENDIX 2
AGROECOSYSTEM PEER REVIEW PANEL
James M. Davidson, Chair
Dean for Research
Institute of Food and Agric. Sci.
1022 McCarty Hall
University of Florida
Gainesville, FL 32611
(904) 392-1784
Fax: (904) 392-4965
(soils, and nonpoint
source pollution)
David C. Coleman
Dept. of Entomology
University of Georgia
Athens, GA 30602
(404) 542-2309
(agroecosystems, general)
John A. Lawrence
Boyce Thompson Institute
Ithaca, NY 14853
(607) 254-1215
(environmental impacts)
Donald T. Searls
Dept. of Math & Applied Stat.
University of No. Colorado
Greeley, CO 80635
(303) 351-2055
(Home: 303/353-5359)
(design and statistics)
Jan L. Flora
Dept. of Agricultural Economics
213 Hutcheson Hall
Virginia Polytechnic Institute
Blacksburg, VA 24061-0401
(703) 231-9441
Fax: (703) 231-4163
(socioeconomics)
A2 1
-------
APPENDIX 3
SUMMARY OF SELECTED EXISTING DATABASES
The table on the following pages lists a selection of existing databases which have
potential application in the development of indicators. Databases will be evaluated in detail by
the Agroecosystem Resource Group as specific data needs are identified.
Key to References:
(1) Abramovitz, J.N., D.S. Baker and D.B. Tunstall (1990). Guide to Key Environmental Statistics
in the U.S. Government. World Resources Institute (April 1990), Washington, DC.
(2) Olsen, G.L. and R.P. Breckenridge (1989). Summary of Federal Contaminant Monitoring
Programs and Related Data Bases: A Fish and Wildlife Perspective (Draft May 29,1989). Center
for Environmental Monitoring and Assessment, Idaho National Engineering Laboratory, Idaho
Falls, ID.
(3) Olsen, G.L., R.P. Breckenridge and G.B. Wiersma (1990). Assessment of Federal Databases
to Evaluate Agroecosystem Productivity. Report Number EGG-CEMA-8924, Idaho National
Engineering Laboratory, Idaho Falls, ID.
(4) USGS (1983). Scientific and Technical, Spatial, and Bibliographical Data Bases and Systems
of the USGS, 1983. Geological Survey Circular 817, USGS, Alexandria, VA.
(5) USDA (1983). Scope and Methods of the Statistical Reporting Service. USDA Statistical
Reporting Service Misc. Pub. No. 1308 (September 1983). US Govt. Printing Office, Washington,
DC.
(6) USDC (1990). 1987 Census of Agriculture, Volume 2, Subject Series, Part 2, Coverage
Evaluation (AC87-S-2). US Department of Commerce, Bureau of the Census, Agricultural
Division. US Govt. Printing Office, Washington, DC.
A3 1
-------
Table A3.1: Selected databases which have potential application in the development of
indicators.
Program
Agency
Description
Coverage
Frequency
Refs
General Resource Inventories
National Resources
Inventory (NRI)
Major Uses of Land in the
U.S.
National Land Use and Land
Cover Mapping Program
Water, Land and Related
Data
National Wetlands Inventory
USDA / SCS
USDA / ERS
USGS
BOR
USF&W
Data on the status, condition, and trends of soil, water, and
related natural resources on non-federal land.
Inventory of major land uses in the U.S., including
cropland, idle cropland, pasture and range, farmsteads,
forest, transportation, recreation, and defense.
Land use and land cover data in digitized format (vector
polygon or grid cell).
These data have been published by the Bureau of
Reclamation on BOR projects in the western US annually
since 1920 on acreage, yield, production, gross crop value,
and multi-purpose functions such as municipal and industrial
water service, power, recreation and flood control.
Comprehensive information on the extent and characteristics
of wetland resources. Includes estimates of wetland loss.
National
National
National
Western
states
Lower 48
states
Every 5
years
Every 5
years
Once,
approx.
1975-85
Annual
1
1
1
3
1
Agricultural Data
Agricultural Statistics
USDA / NASS
Crops: estimates of acreage farmers intend to plant, planted
acreage, acreage intended for harvest, probable yields, actual
harvest, actual yield; utilization, disposition, and value of
commodities produced.
Livestock: production, breeding intention, breeding success;
estimates of manufactured dairy products.
Other: Fertilizer, number and size of farms, farm labor and
wages, prices received for produce, grain stocks.
National
Annual
5
~
-------
Table A3.1: Selected databases which have potential application in the development of
indicators.
Program
Census of Agriculture
RFF Pesticide Data
Commercial Fertilizers
Agency
USDC / Ag.
Census Division
Resources for
the Future
TVA
Description
Comprehensive information covering crop farming,
ranching, livestock, irrigation, agri-chemicals and
productivity.
Estimates of pesticide use by county.
Statistics on commercial fertilizer consumption. Data were
collected by USDA / NASS before 1985; TVA since 1985.
Coverage
National
National
National
Frequency
Every 5
years
-
Annual
Refs
1,6
3
3
Water Resources
National Hydrological
Benchmark Program
National Water Quality
Assessment Program
(NAWQA)
National Water Data
Exchange (NAWDEX)
National Stream Quality
Accounting Network
(NASQAN)
USGS
USGS
USGS
USGS
Initiated in 1964 to provide a nationally uniform basis for
assessing long-term trends in the physical and chemical
characteristics of surface waters which should be largely
unaffected by land use.
A pilot program started in 1986 to provide nationally
consistent description of water quality and an understanding
of the factors which influence water quality. Pilot scheduled
to be completed in 1990.
A computerized index of water data from over 375,000 sites
and 410 organizations. Information includes site location,
collecting organization, types of data available, spatial and
temporal coverage. Access to databases which contain the
data (eg, WATSTORE, STORET)
Since 1972, provides a nationally uniform basis for assessing
large-scale, long-term trends in the physical, chemical and
biological characteristics of the nations 's surface waters.
37 states
National
Varies
National
Varies
from
monthly to
quarterly
Pilot
Varies
Varies
from bi-
monthly to
quarterly
1,4
2,4
2
1,4
-------
Table A3.1: Selected databases which have potential application in the development of
indicators.
Program
WATSTORE
Acid Rain Effects on
Watersheds
STORE!
National Stream Survey
National Pesticide Survey
Water Data Management
System
Snow Survey and Water
Supply Forecasting
Agency
USGS
USGS
EPA
EPA
EPA
BLM
USDA / SCS
Description
The Water Resources Division of the USGS investigates the
occurrence, quantity, quality, distribution and movement of
US surface and underground waters.
Wet deposition; pH; dissolved oxygen; metals; NO3; NH4;
SO4; and others.
A computerized database system containing data pertinent to
water quality. The data are collected, entered, and
maintained by various federal, state and local government
agencies and their contractors.
A probability sample of stream reaches within areas if the
US expected to contain waters of low acid neutralizing
capacity. Chemical variables measured included pH,
extractable aluminum, fluoride, dissolved organic carbon,
acid neutralizing capacity, phosphorous, ammonium and
turbidity.
A one-time survey of drinking water wells to determine the
national occurrence of 126 pesticides and nitrates.
Water use on public lands.
Snowpack and other hydrometeorological data are collected
to produce seasonal water supply forecasts.
Coverage
National
National
National
National, in
areas know
to have
acidic
deposition
problems
National
Western
US
AK, AZ,
CA, CO,
ID.MT,
NV, NW,
OR, UT,
WA, WY
Frequency
Varies
.
Varies
-
Varies
.
Monthly
during
snow
season
Refs
4
.
2
-
2
.
1
Air Resources
>
J^
-------
Table A3.1: Selected databases which have potential application in the development of
indicators.
Program
National Acidic Precipitation
Assessment Program
(NAPAP)
Aeromatic Information
Research System (AIRS)
Storage and Retrieval of
Aeromatic Data (SAROAD)
Agency
Inter-agency
EPA
EPA
Description
Wet and dry atmospheric deposition on soil, water, flora.
Ambient air quality emissions and compliance data.
A system for editing, storing, summarizing and reporting
ambient air quality data.
Coverage
National
National
National
Frequency
Varies
Varies
Varies
Refs
2
2
2
Wildlife
National Contaminant
Biomonitoring Program
(NCBP)
North American Breeding
Bird Survey
Christmas Bird Count
Threatened & Endangered
Species Data (TEDS)
Ecological Site Inventory
Integrated Habitat Inventory
Classification System
(IfflCS)
USF&W
US F&W
Audubon
Society
BLM
BLM
BLM
The NCBP's objective is to determine geographic and
temporal trends of persistent organic and inorganic
contaminants in fish and wildlife tissue.
Provides a uniform basis for assessing long-term trends in
avian populations throughout North America since 1966.
Members of the National Audubon Society indentify and
count all of the birds they see in a geographical area. The
counts have been taken every December since 1900.
Location of threatened and endangered plant and animal
species.
Plant species composition.
Wildlife habitat sites.
National
Lower 48,
Alaska,
Canada
Global
Western
US
Western
US
Western
US
Under
revision
Annual
Annual
-
-
-
1
1
-
2
2
2
-------
Table A3.1: Selected databases which have potential application in the development of
indicators.
Program
Wildlife Observation Report
Data System (WORDS)
Biological Effects Program
Agency
BLM
National Park
Service
Description
Wildlife counts.
Effects on flora of ozone, SOX, NOX, HF and trace elements.
Coverage
Western
US
National
Frequency
-
-
Refs
2
-
-------
oo
i
-J
Table A3.1: Selected databases which have potential application in the development of
indicators.
Program
Agency
Description
Coverage
Frequency
Refs
Miscellaneous
Wind Erosion Conditions
National Climate Data
Center
USDA / SCS
NOAA
Wind erosion reports in the Great Plains states provide
information on the status and condition of affected soil and
vegetation.
Meteorological data.
CO, KS,
MT, NB,
MM, ND,
OK, SD,
TX, WY
National
Three
times a
season
1
-------
APPENDIX 4
LANDSCAPE CHARACTERIZATION
Analysis of agroecosystem indicators in a landscape context requires a well designed
geographic information system (GIS) database. Such a database would include detailed, spatially
accurate data on land use, land cover, soils, topography, and other landscape attributes over broad
areas. The EMAP Landscape Characterization Database (LCD) is intended to provide detailed
characterization data for large areas of the nation and will be available in a GIS format
(ARC/INFO). During the early design phases of the EMAP LCD each EMAP Ecosystem
Resource Group submitted characterization requirements. EMAP-Agroecosystem requirements
are shown in Table A4.1. The initial intent of the EMAP LC program was to characterize the
12,600 EMAP hexagons according to the criteria submitted by the Resource Groups. However,
the EMAP Landscape Characterization program is presently undergoing extensive revision as a
result of a June, 1990 peer review of the characterization process. At this time the ARG does not
know what data the EMAP LC Group will ultimately provide to the Ecosystem Resource Groups.
A general framework for the analysis of landscape data for agroecosystems is proposed
in Appendix 8.3, Land Use and Landscape Descriptors (Figure AS.3.1). Landscape
characterization, at least to the level of detail described in Table A8.3.2 of that same section, is
required for the analysis and interpretation of the proposed land use and landscape descriptors
and for the analysis of indicators in a landscape context. As the EMAP LCD will not be available
during the Agroecosystem pilot program, an alternative approach is being developed. Land use
data from the June Enumerative Survey will be collected by NASS for all sampled segments.
Aerial photographs of each segment will also be available. Initially, the aerial photographs will
not be digitized for use in a (GIS), precluding calculation of the full suite of land use and
landscape descriptors. During the pilot program, members of the Agroecosystem Resource Group
and NASS will explore various options for digitizing the land use and cover data using the NASS
photography. Soils information for the segments may be obtained from Soil Conservation Service
soil survey maps. General soil maps for North Carolina, the proposed site of the initial ARG pilot
project, have been digitized by the North Carolina Center for Geographic Information and
Analysis (CGIA) (T.Tribble, personal communication); digitization of detailed soil maps is in
progress. A partial list of data available from the North Carolina CGIA is shown in Table A4.2.
By utilizing data from these and other sources, the ARG will develop procedures for the analysis
of indicator data in a landscape context.
A4 1
-------
Table A4.1: Proposed landscape classification categories.
General1
Urban or built-up
Industrial
Utilities
Transportation
Agriculture
Wetlands
Forested
Non-forested
Non-forested, floating
Scrub-shrub
Forest
Deciduous
Coniferous
Mixed
Rangeland
Herbaceous
Scrub-shrub
Arid lands
Barren lands
Surface water
Other
Agricultural
Use
Row crops, by species
Field crops, by species
Pasture
Idle
Fallow
Orchards, to species
Nurseries, tree / shrub
Horticultural and
specialty crops,
by species
Managed animal
production, to cattle,
swine, fowl, etc.
Aquaculture
Farm ponds
Farmsteads, buildings
Management
Irrigated
Till / no-till
Grassed waterways
Terraced
Soils
Soil associations
Physiography
Organic deposits
Flood plain moraine
Aeolian deposits
Glacial lake bed
Drumlin
Useful information
Erosion potential
Percolation rate
Slope
Depth to water table
Crop potential
Land modification
Fill/soil deposition
Waste deposition
Grading
Tilling
Ditched
Irrigated
Impounded
1. Identify dominant species in appropriate categories.
A4 2
-------
Table A4.2: Partial list of landscape characterization data available from the North Carolina Center for Geographic
Information and Analysis as of July, 1990.
Data1 (source2)
General soils (SCS)
Detailed soils (SCS)
Geology (NC Geologic Survey)
Hydrography (USGS maps)
Land use (NHAP photos)
Natural Heritage Inventory (DPR)
Coastal reserves (DCM)
Game lands (WRC)
County boundaries (USGS maps)
Municipal boundaries (DOT)
1980 census boundaries
1990 census boundaries
Coverage
Statewide
12 counties
Statewide
Statewide
12 quads
Statewide
Statewide
Statewide
Statewide
Statewide
Statewide
Statewide
Scale
1:250,000
1:24,000
1:250,000
1:100,000
1:24,000
N/A
1:24,000
1:126,720
1:100,000
1:126,720
1:63,360
1:63,360
Status
Available
Available
Available
In progress
Available
In progress
Available
Available
Available
Available
Available
Available
1. All data are stored in ARC/INFO.
2. DCM NC Division of Coastal Management
DPR - NC Division of Parks and Recreation
DOT NC Department of Transportation
NHAP - National High Altitude Photography
SCS USDA / Soil Conservation Service
USGS - US Geological Survey
WRC NC Wildlife Resources Commission
A4 3
-------
APPENDIX 5
INDICATOR DATA NEEDS AND DATA SOURCES
Indicator and Data Needs
Crop productivity
Outputs:
crop yield of each crop
acreage per crop
crop use (forage, grain, etc)
net primary productivity
harvest index for each crop
moisture content of harvested crop
root:shoot ratio for each crop
Inputs:
type, rate and frequency of fertilizer use
type, rate and frequency of pesticide use
type, rate and frequency of herbicide use
lime added
sewage sludge added
irrigation water
fuel use: # tillage passes and type of machinery
Ancillary data:
temperature (min, max, degree days)
total rainfall
rainfall during growing season
solar radiation (growing season and total)
catastrophic events (hail, hurricanes, flooding)
severe pest or disease outbreak
costs of inputs
plant variety?
Data Source
NASS question
NASS question
NASS question
EMAP calculation
literature
NASS question
literature
NASS
NASS
NASS
NASS
NASS
question
question
question
question
question
NASS question
Nat. Weather Service?
NWS
NWS
NWS?
NWS?
EMAP sample?
literature
Soil productivity
(1) nutrient-holding capacity
exchangeable acidity
total C
total N
totals
specific conductance
cation exchange capacity (CEC)
exchangeable cations
NASS sample
A5-1
-------
extractable P
bulk density
coarse fragment
% sand, silt, clay (texture)
soil permeability
infiltration rate
hydraulic conductivity
organic matter (ignition method)
Ancillary data:
use of manure/kind/amount
tillage practices
use of fertilizers and micronutrients
(2) contaminants
As, Cd, Cr, Hg, Ni, Pb, V
Ancillary data:
use of sewage sludge and amount applied
use of manure/kind/amount
(3) erosion
rill/sheet erosion (USLE):
SCS soil type classification
T value
length of field
slope of field
rainfall factor (R)
soil erodability factor (K)
crop/vegetation factor (C)
management/mechanical factor (P)
Ancillary data:
tillage practices
NASS question
NASS question
NASS question
NASS sample
NASS question
NASS question
SCS office
NRI SOILS-5 dataset
EMAP measure?
i
EMAP measure?
SCS office?
NRI SOILS-5 dataset
NASS question
NASS question
NASS question
To develop: wind erosion and gully erosion. Consider the Wind Erosion Equation used by SCS.
Consider estimating the extent of land area affected by gully erosion fron aerial photographs.
(4) microbial biomass
soil organic matter as surrogate until further developed
NASS sample
Agricultural chemical use
type, rate and frequency of fertilizer use
type, rate and frequency of pesticide use
type, rate and frequency of herbicide use
NASS question
NASS question
NASS question
A5-2
-------
Ancillary data:
costs of chemical inputs
literature
Nonpoint source loadings
(1) initial surrogate: nitrates and atrazine
in adjacent streams, groundwater, farm ponds
and irrigation drainwater
(2) initial surrogate: soil displaced/field
EMAP sample
see soil erosion
Irrigation water quality and quantity
electrical conductivity (EC)
pH, Cl, NO3, Se, B, Na, Ca, Mg
sulfates?
total dissolved solids (TDS)
agricultural chemical residues, e.g. atrazine
acres irrigated
type of irrigation system (flood, sprinkler)
volume of water applied
source of water
availability of water
Ancillary data:
water temperature
cost of irrigation water
flow rate?
NASS or EMAP measure
NASS or EMAP sample
NASS or EMAP sample
NASS or EMAP sample
NASS question
NASS question
NASS question
NASS question
NASS question?
NASS or EMAP measure
Pest density: nematodes and weeds
abundance and diversity of plant parasitic nematode genera
abundance and diversity of weed genera
Ancillary data:
type, rate, and frequency of herbicide use
type, rate, and frequency of nematicide use
use of soil fumigants
crop variety
crop rotation history
temp and rainfall during growing season
stage of crop maturity at sampling
soil physical and chemical parameters
NASS sample?
EMAP sample?
NASS question
NASS question
NASS question
NASS question
A5-3
-------
Density of beneficial insects
number of pests and number of beneficials/ft row
number of parasitized pests/ft row
Ancillary data:
type, rate and frequency of insecticide use
participation in CRP or other set-aside programs
crop variety
crop rotation history
temp and rainfall during growing season
EMAP or NASS sample
EMAP or NASS sample
NASS question
NASS question
NASS question?
NASS question
NWS?
Landscape descriptors
woodlot area and perimeter
hedgerow/ shelterbelt area and perimeter
stream corridor area, perimeter and width
vertical profile of noncrop vegetation
predominant vegetation
percent closure of overstory
EMAP measure
EMAP measure
EMAP measure
EMAP measure
EMAP measure
EMAP measure
Land use
land area in cultivation
land area in specific crops and cultivars
land area left in woodlots
land area in farm ponds
land area in grass waterways
land area in buildings and paved areas
previous use of the land
irrigated land area
NASS
NASS
NASS
NASS
NASS
NASS
NASS
NASS
question
question
question?
question
question
question
question
question
Status of biomonitor species:
(1) honey bees in stationary hives
contaminants in foragers
contaminants in honey
number of colonies
honey production
bee mortality
bee disease
(2) earthworms
number of earthworms
contaminants in earthworms
Ancillary data: soil type
NASS or ABA
NASS or ABA
NASS or ABA
NASS or ABA
NASS or ABA
NASS or ABA
NASS sample
NASS sample
SCS office
A5-4
-------
(3) white clover
need more information
(4) lichens and mosses
number of species EMAP measure
frequency of occurrence EMAP measure
% cover EMAP measure
contaminants EMAP sample
Foliar symptoms
(to be developed)
Socio-economic factors
(tentative)
farm size NASS question
number of farms calculation
rural population Bureau of Census?
social institutions Bureau of Census?
sodbuster/CPR acreage NASS question
ownership NASS question
federal commodity programs NASS question
indebtedness NASS/ERS question?
perceptions of land manager regarding the environment NASS question?
indebtedness
consumption of fossil fuels
farm program financing criteria
A5-5
-------
APPENDIX 6
NASS QUESTIONNAIRE FOR PILOT SURVEY
A6.1. ADDITIONS TO THE JUNE ENUMERATIVE SURVEY
Insert new land use categories into Section D:
(a) woods
(b) wetlands
(c) farm ponds
(d) Conservation Reserve Program
(e) other set-aside programs
(f) irrigated (yes/no)
A6.2. QUESTIONS FOR THE NASS ENUMERATORS IN A DECEMBER SURVEY
(on the sample fields only)
1. Insert the questions from the December survey (Sections 2 & 3) on yield,
acreage, and use by crop.
2. If appropriate, ask if the yield in item 1 is the wet or dry weight. (Obtain
moisture content for corn and soybeans).
3. Ask the questions from the Form H: Cropping Practices Survey, Parts II
and III (Fertilizers and Pesticides)
Incorporate the following questions into the form H format:
4. If lime was added, what form of lime?
5. If manure was added, how much?
(List any pertinant data for office computation)
6. Has sewage sludge been applied to this field in the past five years?
If yes, how much was applied?
When was the last application (year)?
7. Do you currently use any erosion control measures?
(from Form W)
a. terracing?
b. contour stripping?
c. strip cropping?
A6 1
-------
d. grassed waterways?
e. other (list)
8. Which system best describes the current tillage practice used on this sample
field? (from Form W)
a. No-Till
b. Ridge-Till
c. Mulch-Till or other conservation tillage
d. Conventional without moldboard plow
e. Conventional with moldboard plow
f. other (list)
9. What crops were previously planted in this field during the past four growing
seasons (include cover crops)?
If not cropped, list field use (pasture, idle, forest, wetland, other?)
Fall 1990
Spring 1990
Fall 1989
Spring 1989
Fall 1988
Spring 1988
Fall 1987
Spring 1987
10. Has this field been irrigated in the past 5 years? (Form W)
11. Did you irrigate the sample field in 1991? (from Form W)
12. If yes, what was the quantity of water applied per acre to the sample field in
1991? (from Form W)
(If not in acre/feet, list other pertinent data for office computation)
13. What was the source of the irrigation water?
(modified from Form W)
a. surface water
b. farm ponds
c. wells
d. purchased water
e. other (list)
A6 2
-------
14. What type of irrigation system was used?
(modified from Form W)
a. sprinkler irrigation system
b. gravity (furrow or flood)
c. drip or trickle
d. subirrigation (water applied beneath the ground or the maintenance of the
water table at a predetermined depth)
e. other (list)
15. Was the application of irrigation water managed by use of soil moisture
sensing devices? (from Form W)
16. Is there any limitations in the water supply to irrigate this field? (If yes, list)
A6.3. ADDITIONAL DATA NEEDED FOR PILOT STUDY FROM OTHER
SOURCES THAN THE NASS SURVEY QUESTIONNAIRE
soil type (and soil record number)
LS value of soil map unit (nonterraced field)
LS value of soil map unit (terraced field)
R factor
C factor
K factor
P factor
T-value for the soil type (from SCS)
soil analyses
water analyses
temperature (min, max, degree days)
total rainfall
rainfall during growing season
harvest index for each crop
root:shoot ratio for each crop
specific land uses (woodlots, hedgerows, ponds, etc.) -- aerial photography
A6 3
-------
-3-
SECTION B - RESIDENCE AND SCREENING
1. [Does the operator of this tract live inside or outside the segment?!
Q INSIDE - [Enter 1J
(~1 OUTSIDE - [Enter 2, then go to SECTION C.I
168
2. Were there any other persons living in this household on June 1 who operated a farm or ranch?
n YES-[Entername(s).]
[Assign separate tract letter(s) on Screening Questionnaire, then go to Item 3.]
D NO
3. On June 1, did you operate land under any other name or land arrangement other than
[name listed on Face Page]!
Q YES - [Assign separate tract letter(s) on Screening Questionnaire for other arrangements),
then go to SECTION C]
n NO
********************************
SECTION C - SECTIONS TO BE COMPLETED
1. [Are all three boxes at bottom of page checked?]
Q YES - [Go to SECTION D.]
n NO
2. [Is this a new segment, or a new operator in an old segment?]
Q3 YES -[Check all boxes at bottom of page, then go to SECTION D.]
n NO
3. [Was this tract nonagricultural last year?]
Q] YES -[Check all boxes at bottom otpage, then go to SECTION D.]
D N0
4. [Have you changed or corrected the operation name, combination of individual names,
or operator's name on Face Page?]
Q] YES - [Check all boxes at bottom of page, then go to SECTION D.]
n NO
5. [Have you changed, corrected, or deleted any partner's names in SECTION A, page 2 ?]
Q YES - [Check all boxes at bottom of page, then go to SECTION D]
Q NO -[Go to SECTION D.]
***********
HOGS -CROPS
CATTLE -SHEEP
AG LABOR
A6-4
-------
-4-
SECTION D - CROPS AND LAND USE ON TRACT
How many acres are inside this blue tract boundary drawn on the photo (or map)!
Now I would like to ask about each field inside this blue tract boundary and its use during 1989.
FIELD NUMBER
1 Total acres in field
2 Crop or land use [specify!
3 Occupied farmstead or dwelling
4 Woods, waste, roads, ditches, etc
Permanent - not in crop rotation
5 Pasture
Cropland - used only for pasture
7 Idle cropland -Idle all during 1989
8 Two crops planted in this field for harvest
this year or two uses of the same crop?
[Specify second crop or use 1
Acres
9 Acres left to be planted?
15. Planted
16. Forgram
17 Planted and to be planted
18. Forgram
19 Planted and to be planted
20. For grain
21. Planted and to be planted
22 Forgram
23 Planted and to be planted
24 Forgram
26 Sorghum Planted and to be planted
[Exclude crosses
27 with Sudani for grain
28. Other uses of grains planted Use
silage, etcj Acres
,„ Alfalfa ana alfalfa mixtures
Cut and to be cut for hay
30 Hay Grain -Cut and to be cut for hay
32 Other hay - Cut and to be cut for hay
33 Planted and to be planted
34 Following another crop
35c. Burley Acres
35d Flue-cured Acres
36 Peanuts Planted and to be planted
38 Upland Cotton [net acres if skip rowedl
Planted and to be planted
46 Irish Potatoes Planted and to be planted
47 Sweetpotatoes Planted and to be planted
48 Other crops Acres planted or in use
01
328
843
•
841
842
•
856
857
•
a Yes a MO
844
*
610
540
*
541
•
547
•
548
•
533
•
534
535
536
530
»
531
•
570
571
•
.
653
*
656
•
654
•
bOO
602
732
•
315
690
524
»
884
558
*
02
828
^§$MIM§^
841
842
856
857
D Yes D No
844
•
610
*
540
•
541
547
•
548
533
534
535
536
•
530
•
531
570
•
571
•
653
•
656
654
600
602
•
732
•
315
690
524
884
558
•
•
03
828
•
841
*
842
•
856
857
•
Q Yes D No
844
610
540
•
541
•
547
•
548
533
•
534
•
535
•
536
•
530
531
570
•
571
•
653
656
654
*
600
*
602
732
315
690
524
•
884
•
558
•
_
04
828
•
841
842
856
•
857
D Yes D No
844
•
610
540
•
541
•
547
548
533
•
534
•
535
536
•
530
531
•
570
571
.
653
656
654
600
602
732
•
31 5
•
690
•
524
•
•
558
•
•
05
828
*
841
•
842
856
•
857
•
Q Yes a No
844
610
•
540
•
541
547
•
548
•
533
•
534
•
535
•
536
•
530
531
570
571
•
.
653
656
•
654
600
602
732
315
•
690
524
•
•
558
•
•
A6 5
-------
December 1989 Agricultural Survey
ACJUS OPERATED
1. How many total Kiti of land w«f« m ihn operation or> Dtttmbtr 1?
Include farmstead. all cropland, woodland, pasturtlarxJ. wasteland,
government program larid. »fl land owned, rented or managed
land rented lo others Kid all graung Ijrx) uv«d on in AUM (ff»
2. 0' the total acres m this operation, how many acres would b« considered
croplind. including Und in hjy »nd crooi»nd m government pcogrimj'
How to complete this section
• • Report for ill ihe Urtd yog operjle. including i*nd renied (rom othefi.
• • If njry«tunol complete, mike your r>nintim«teof »crn and total production
• - Production i » eoual to Kf « harvested and lo be harvested times averag* yield p«f a grain and seed
Total grjm and seed produdion
Acreicut (o< silage
Total silage production
Acre) (or all other purposes, including Abandonment
SORGHUM Imilo):
(trxJudt other gran and tongt sorgnum. f ic'i/de sorgnu/n *
Acres planted for all purposes
Acrei harveited and to be hirvMted for 9rjin
Total grain production
Axrei cut for silage
Toul sJlage produ
-------
SECTION 3 - CROPS (Continued)
Acres harvested
Total production .
Include hay nay/age tnd jreen cftop^ ...
Total production ol dry hay m tom
OR(Vo o'twfes
Acres cut for dry hay (fcludt ilfjw. nay/age. tr>d grtenchop)
Total production of dry hay m torn
ALL OTHER HAY (include c/over. dmo(ny. (lover andgrau murluf«.
brome. Judan. suOan crosses, millet, other lame and »vi/o nay )
Acres cut 'or dry h*y (eic'uoe fiiyltge ina green ihop)
Total production of dry hay m torn
OR (Wo oltalet
$WEETPOTATOES:
Acres harvested
Total production .
OTHER CROPS (Specify):
Acres Of ALL PASTURE LAND (mc/uo'e only /ind anted in/s yeac
tnd not hji\ested lor gran or ftay eic/ude gn
ANY OTHER LAND not reported, including woods, waste, ponds, orchards.
farms'.fads idle land, farm lots, eu (t'flude grmng i/lotmentl)
WINTER WHEAT AND RYE SEEDINGS FOR THE 1990 CROP YEAR
2. Please repo'V WINTER WHEAT and RYE seed.ngs lor the 1990 CROP YEAR.
WINTER WHEAT acres seeded and 10 be seeded for all purposei
RYE acres seeded and to be seeded for all purposes
Continue On Neit Page
1 • incomplete. Has Crops
2 • Incomplete, Crop*
Presence Unknown
138
3 valid Zero
A6 7
-------
UNITED STATES DEPARTMENT OF AGRICULTURE
NATIONAL AGRICULTURAL STATISTICS SERVICE
Parti. Seeding
FORM H: SOYBEAN YIELD SURVEY -1989
CROPPING PRACTICES INTERVIEW
YEAR.CROP.FORM, MONTH
(1-5)
928
Starting Time (Military Time)
Form Approved
O.M.B Number0535-0088
Expiration Date 7/31/92
CE.-120034H-3
101
1. Copy acres of soybeans planted in Sample Field No.
from Form A, Table A, Col. 3 .
If Item 1 has
A Zero entry- Enter ending time and go to Form A.
An Acreage entry - Continue.
2. Has this field been (or will it be) irrigated? YES = 1, NO 2
3. What was the seeding rate per acre? .
4. What was the source of soybean seed used?
1 = Homegrown or traded
2 = Purchased
5. If purchased, what was the cost of the seed? .
6. What crop(s) were grown(harvested) on this field in 1988?
103
. Enter code
. Bushels
OR
Pounds
.Enter code
105
111
110
. . Dollars per Bushel
OR
Dollars per 50 # bag
. . . Enter Code
If double cropped, enter 1st crop in Item Code 125 and 2nd crop in Item Code 126.
7. What crop(s) were grown(harvested) on this field in 1987? . ... . . .Enter Code
If double cropped, enter 1st crop in Item Code 127 and 2nd crop in Item Code 128.
115
120
$
121
$
125
126
127
128
For crops not listed on card^
90 Other, specify
91 Other, specify
199
Completion Code
A6 8
-------
Part II. Fertilizer
1.. Was livestock or poultry manure applied to this field?
2. Did you apply ANY chemical fertilizer on this soybean field?
YES * 1. Continue NO = 2, Go to Item 3
YES = 1. NO = 2. . Enter Code
201
. Enter Code
203
Include all amounts applied specifically for this soybean crop since preparation of field began.
WHEN
APPLIED
1 - Before seeding
(fall)
2 -Before seeding
(spring)
3- At seeding
4- After seeding
Col. 1
211
217
223
229
235
241
MATERIAL USED
(Enter percent analysis or actual
pounds of plant nutrients
applied per acre).
N
NITROGEN
Col. 2
212
218
224
230
236
242
P205
PHOSPHORUS
Col. 3
213
219
225
231
237
213
K2O
POTASH
Col. 4
214
220
226
232
238
244
QUANTITY
Used Per
Acre!/
Col. 5
215
221
227
233
239
245
UNIT
Pounds - 1
Gallons - 2
Actual plant
nutrients = 3
Col. 6
216
222
228
234
240
216
11 If actual pounds of plant nutrients are reported in columns 2, 3, or 4, leave
column 5 blank and enter a code 3 in column 6.
Kind
EXAMPLES OF ANALYSIS BY KIND
Percent Analysis Kind
Percent Analysis
Diammonium Phosphate 18 46 0
Triple Superphosphate 0 46 0
Potassium Chloride 0 0 60
Mixed Fertilizers 6 24 24
Mixed Fertilizers 9 23 30
Mixed Fertilizers 3 9 18
3. Was LIME applied to this field? YES = 1, NO 2
If YES. how many pounds were applied per acre?
4. Was SULPHUR applied to this field? YES 1. NO = 2.
If YES, how many pounds were applied per acre? .
5. Were any MICRO NUTRIENTS applied to this field? YES = 1, NO
6. Was a soil test performed on this field in 1988 or 1989? YES = 1, NO
If yes, did it include a Nitrogen test? YES = 1, NO
7. Was a soil test performed on this field in 1987? YES 1, NO
If yes, did it include a Nitrogen test? YES - 1, NO
Enter Code
.Pounds
. Enter Code
.Pounds
2 . . Enter Code
2 . . Enter Code
. 2 . . Enter Code
i 2 . . Enter Code
i 2 . . Enter Code
Completion Code
380
381
383
384
385
386
387
388
399
A6-9
-------
Part III. Pesticides
1. How many times during the growing season was this field cultivated for weed control?
(Zero is a valid answer. Enter a positive number or a "0")
Enter Code
Enter Code
401
402
403
404
2. Were any herbicides used on this soybean field? YES = 1, NO = 2 . . .
3 Were (or will) any insecticides (exclude seed treatment) be used on
this soybean field? YES = 1. NO = 2, DON'T KNOW = 3
4. Were (or will) any fungicides (exclude seed treatment) be used on
this soybean field? YES = 1. NO = 2. DON'T KNOW = 3 . Enter Code
If Question 2, 3 or 4 is YES, complete the table below. Otherwise, go to Part IV.
A. If two or more products were mixed together in the spray tank before application to this field,
REPORT them on the SAME LINE using the applicable product codes listed on the Pesticide Card.
EXAMPLE: For Treflan applied alone-- record code 124 under the first product used, col. 2.
Fora tank-mix of Dual + Sencor-on the same line record code 109 under
the first product and code 120 under the second product, col. 2.
B. If the same product was used more than once- record EACH application on a separate line.
ASK for the products used on this field starting in the Fall of 1988.
Start with before planting applications.
s
b
Q
U
E
N
c
E
1
2
3
4
5
6
7
8
9
10
When Applied:
1 - Before planting
2 = At planting
3 = After planting
Col. 1
411
417
423
429
435
441
447
453
459
465
Product Code(s):
(Listed on card)
1st
2nd
3rd
Col 2
412
418
424
430
436
442
448
454
460
466
413
419
425
431
437
443
449
455
461
467
414
420
426
432
438
444
450
456
462
468
How was it Applied:
1 = Broadcast, Ground
2 = Broadcast, Air
3 = In furrow
4 = Irrigation water
5 = Banded in/over row
6 = Directed spray
9 = Spot treatment
Col. 3
415
421
427
433
439
445
451
457
463
469
Who Applied:
1 = Farmer
2 = Custom
Col. 4
416
422
428
434
440
446
452
458
464
470
For pesticides not listed on card:
Herbicide?: 190 Other, specify
Insecticides: 290 Other, specify
Fungicides: 390 Other, specify
_, 191 Other, specify
_, 291 Other, specify
, 391 Other, specify
Completion Code
699
A6- 10
-------
Part IV. Tillage and Planting Operations
1. Was a cover crop seeded in this field for soil erosion protection purposes during the winter of 1988 -1989?
YES = 1. NO = 2 EnterCode
2. Has the Soil Conservation Service (SCS) notified you that this field is "Highly Erodible" (HEL)?
YES = 1. NOT"HEL"-2, FIELD NOT EVALUATED = 3 ...EnterCode
3. Now would you list the tillage and planting implements used on this field beginning with the first trip over the
field after harvest of the preceding crop and continuing through seeding of this crop? Do not include fertilizer
or pesticide implements. Use the implement codes listed on the Implement Card.
If two or more implements were used at the same time. REPORT them on the SAME LINE.
Trip
over
the
field
1st
2nd
3rd
4th
5th
6th
7th
8th
9th
10th
Implements
Used
Col. 1
Implement
Code(s)
1st
Col. 2
711
717
723
729
735
741
747
753
759
765
2nd
Col. 3
712
718
724
730
736
742
748
754
760
766
When
used
fall = 1
spring = 2
Col. 4
713
719
725
731
737
743
749
755
761
767
Width in
FeetV
Col. 5
714
720
726
732
738
744
750
756
762
768
Acres
Covered
Per Hour
Col. 6
715
721
727
733
739
745
751
757
763
769
TRACTOR
PTO
HP2/
Col. 7
716
722
728
734
740
746
752
758
764
770
// Moldboard plow, record the number of bottoms: row planter and bedder, record number of rows
21 If unknown, record model number and manufacturer. For tractors without Power Takeorfs, record
Engine Horsepower
ENUMERATOR CHECK: We have listed a total of
Is that correct?
trips over the field for tillage and planting purposes.
For implements not listed on card:
90 Other, specify
91 Other, specify
For Columns 5. 6 or 7:
888 Custom operation, respondent does not know
Ending Time (Military Time)
Enumerator
Return to Form A - - Complete the Interview
Completion Code
STATUS
CODE
A6- 11
-------
APPENDIX 7
EXAMPLES OF DATA SUMMARY TABLES
The following hypothetical data tables were prepared to show the kind of information
EMAP-Agroecosystem will collect in the pilot project and how the information might be
presented. Only basic data summary tables are included here. Many other graphing,
mapping and analytical tables are planned for future data presentation.
Table 1. Crop productivity
Table 2. Agricultural chemical use
Table 3. Quality of irrigation water
Table 4. Irrigation water use in U.S. agroecosystems
Table 5. Sources of irrigation water in U.S. agroecosystems
Table 6. The nutrient-holding capacity of agricultural soils
Table 7. Salinization of irrigated and nonirrigated soils
Table 8. Acidification of agricultural soils
Table 9. Presence of heavy metals in agricultural soils
Table 10. Soil erosion by tillage practice
A7-1
-------
Table 1. Crop Productivity
mean (s.d.)a
minimum
maximum
75 %ile
median
25 %ile
total hectares
no. of samples'5
Nppc
Agricultural
chemical input(
Fuel usee
Irrigation watei
use
Production input
index
Crop
productivity
indexf
a
b
c
d
e
f
Mean and standard deviation of (n) fields segment and (n) segments/region
(n) fields segment and (n) segments/region
Net primary productivity (NPP). The total biomass of crop plants in kg/ha calculated for each crop
from the crop yield, moisture content, harvest index and root:shoot ratio.
Sum of inputs
Calculated from the number of tillage passes and type of machinery
Net primary productivity/total inputs calculated on a per field basis. Regional mean calculated
from (n) fields/segment and segments/ region.
Table 2. Agricultural chemical use
meana
standard
deviation
minimum
maximum
75 %ile
median
25 %ile
total hectares
no. of samples'3
Insecticides0
Herbicides0
Fungicides0
Other
pesticidesc
Fertilizers0
N
P
K
lime
manure
a
b
c
Mean of (n) fields segment and (n) segments/region
(n) fields segment and (n) segments/region
Total amount applied and total active ingredient will be presented in appropriate units.
A7-2
-------
Table 3. Quality of irrigation water
a
b
c
meana
standard deviation
minimum
maximum
75 %ile
median
total hectares
number of samples"
% samples exceeding
quality criteria0
Total dissolved solid
(TDS)
Electrical
conductivity (EC)
Mean and range of (n) fields segment and (n) segments/region
(n) fields segment and (n) segments/region
Recommended United States Geological Service (USGS) quality criteria for TDS
in irrigation water is 1200 mg/1.
Table 4. Irrigation water use in U.S. agroecosystems.
Region
1
2
3
4
5
6
7
8
9
10
Total
Amount of irrigation water used (l/ha)a
drip
sprinkler
furrow
other
total
Mean, standard deviation, maximum and minimum values, 75 and 25 %ile and
median values will all be presented. Calculated from (n) samples/fields, (n)
fields/segment and (n) segments/region.
A7-3
-------
Table 5. Sources of irrigation water.
Region
1
2
3
4
5
6
7
8
9
10
Total
Proportion of total irrigated fields (%)a
groundwater
surface water
other
Calculated from (n) samples/fields, (n) fields/segment and (n) segments/region.
Table 6. The nutrient-holding capacity of agricultural soils
mean
standard
deviation
minimum
maximum
75 %ile
25 %ile
median
total hectares
number of
samples a
organic mattei
(OM) (%)
OM/total N
CEC
(meq)
base satn(%)
texture
nutrient holding capacity
index value
Calculated from (n) samples/field, (n) fields/segment and (n) segments/region.
A7-4
-------
Table 7. Salinization of irrigated and nonirrigated soils
mean
standard deviation
minimum
maximum
75 %ile
median
25 %ile
total hectares
number of samples0
Irrigated
Naa
ECb
Not irrigated
Na (meq)
EC
a Na=sodium content (meq/100 g soil)
b EC= electrical conductivity
c Calculated from (n) samples/field, (n) fields/segment and (n) segments/region.
Table 8. Acidification of Agricultural Soils
mean
standard deviation
minimum
maximum
75 %ile
median
25 %ile
total hectares
no. of samples3
exchangeable
acidity
(meq)
PH
CEC
(meq)
exchangeable
aluminum
(meq)
base satn
(%)
acidification
index value
a Calculated from (n) samples/field, (n) fields/segment and (n) segments/region.
A7-5
-------
Table 9. Presence of heavy metals in agricultural soils
mean
standard deviation
minimum
maximum
75 %ile
median
25 %ile
total hectares
no. of samples3
Contaminant (uj
As
Cd
Cr
Hg
?/g soil)
Ni
Pb
V
a (n) samples/field, (n) fields/segment and (n) segments/region.
Table 10. Soil erosion by tillage practice
mean
standard deviation
minimum
maximum
75 % ile
25 %ile
total hectares
no. of samplesa
Amount of soil displaced
(tons/ha)
Total
Conservation
tillage
practiced
No conservation
tillage
practiced
From land in the
Conservation
Reserve
Program
Estimate of sheet and rill erosion using the Universal Soil Loss Equation (USLE)
on a per field basis (n) fields/segment and (n) segments/region.
A7-6
-------
APPENDIX 8
INDICATOR FACT SHEETS
A8.1 Crop Productivity
1 Endpoint
2 Indicator Category
3 Status
4 Description and Rationale
5 Data Needs
6 Index Period
7 Data Collection
8 Estimated Costs
9 Data Variability
10 Future Work
A8.2 Soil Productivity
A8.3 Land Use and Landscape Descriptors
A8.4 Habitat Linear Classification System and Habitat Layer Index
A8.5 Irrigation Water Quality
A8.6 Irrigation Water Quantity
A8.7 Agricultural Chemical Use
A8.8 Nonpoint Source Loading
A8.9 Biomonitors
A8.10 Pest Density
A8.11 Density and Diversity of Beneficial Insects
A8.12 Socioeconomics
A8.13 Livestock Production
A8.14 Genetic Diversity
A8.1-1
-------
A8.1. Crop Productivity
A8.1.1 Endpoint: Sustainability
A8.1.2 Indicator category: response
A8.1.3 Status: research, active
A8 1.4 Description and Rationale:
Crop productivity is an estimate of the efficiency of agricultural inputs in achieving
the commercial yield. The approach to using crop productivity as an indicator includes
assessing crop yields in relation to inputs and determining if the quantities of specific inputs
needed to maintain yield are changing
Crop yields alone express only the commercially usable portion of the crop and
therefore address only the agronomic perspective of production. An estimate of crop
productivity which has both an agronomic and ecological interpretation is net primary
productivity (NPP). NPP is measured of the organic matter produced by the plant less the
amount respired and thus can be considered a measurement of what the land produces. NPP
is commonly used to assess the productivity of natural ecosystems. The advantage of using
NPP as an indicator is that all crops could be incorporated into a single index for a region.
NPP, expressed in kg or tons/ha, is calculated by multiplying the yield of each crop by an
appropriate conversion factor (harvest index constant = the ratio of total above-ground
production to the commercial yield), and correcting for moisture content (Sharpe et al 1975;
Turner 1987). Estimates of below-ground production (roots), estimated at approximately
10% for annual plants (Monk 1966), can be added to obtain total NPP estimates.
There is no single conventional way that crop productivity is estimated by
agricultural economists, or a consensus on which inputs should be included in the
calculation. The USDA/ERS aggregates seven major categories of inputs into an index of
total farm inputs in their multifactor productivity index that is used to measure annual
changes in the volume of resources used in farm production. These categories are labor; real
estate; machinery; feed, seed and livestock purchases; agricultural chemicals; taxes; and
interest, and miscellaneous. A second computation used by the ERS is the total factor
productivity index, in which three main categories, labor, capital and management inputs
(e.g. feed, seed, agricultural chemicals), are used. The ARG has initially chosen agricultural
chemicals, fuel use and irrigation as the three main input categories for the crop productivity
indicator because these are the main inputs into crop production and because of their
potential environmental impacts.
A crop productivity index could form the basis of a sustainability index. Olson and
Breckenridge (1990) proposed a comprehensive productivity function which includes
agricultural chemicals, fuel use and irrigation as well as quantitative estimates of
environmental inpact, such as run-off of fertilizer or soil into farm ponds (considered as
waste disposal "costs"); pollutants from external sources (e.g. atmospheric ozone), and
A8.1-2
-------
natural calamities affecting yield (e.g. hailstorms) (Figure A8.1.1). Outputs include crop
yield and the residual effects of management practices such as the status of critical biological
and soil factors, which could be estimated by other agroecosystem indicators. The concept
of a "sustainability function" will be further explored in 1991.
A8.1.5 Data Needs
Outputs:
crop yield of each crop
acreage per crop
crop use (forage, grain, etc)
net primary productivity
harvest index for each crop
moisture content of harvested crop
rootrshoot ratio for each crop
Inputs:
type, rate and frequency of fertilizer use
type, rate and frequency of insecticide use
type, rate and frequency of herbicide use
lime added
sewage sludge added
other amendments
irrigation water
fuel use: # tillage passes and type of machinery
environmental impact cost
Ancillary data:
temperature (min, max, degree days)
total rainfall
rainfall during growing season
solar radiation (growing season and total)
catastrophic events (hail, hurricanes, flooding)
severe pest or disease outbreak
costs of chemical, fuel and water inputs
plant variety?
Source of data
NASS question
NASS question
NASS question
EMAP calculation
literature
NASS question
literature
NASS question
NASS question
NASS question
NASS question
NASS question
NASS question
EMAP calculation
Nat. Weather Service?
NWS
NWS
NWS?
NWS?
EMAP sample?
literature
A8.1.6. Index Period
The best time for the collection of the data on yield and inputs is after the crops are
harvested, e.g. early winter . One collection per season is adequate.
A8.1-3
-------
Figure A8.1.1. Comparison of a current productivity function and a modified production function for use
in EMAP-Agroecosytems.
Current crop productivity function used by the U.S Department of Agriculture, Economic
Research Service (Ball 1985):
Total Productivity Function = OUTPUTS INPUTS
animal products labor
feed and food grains capital
other field crops energy
vegetables and fruits fertilizer and lime
fruits and nuts feed and seed
The economic production function can signal changes in productivity and tracks several
anthropogenic inputs. However, it does not track pollution impacts or residual ecological
benefits.
Modified production function for EMAP: "sustainability function"
Sustainability function = OUTPUTS INPUTS
yield management inputs
residual effects / fertilizer and pesticide
/ soil status / energy (tillage)
/ biological status / irrigation
/ microbial status pollutants from external sources
waste disposal "costs"
climatic disasters, e.g. hailstorms
A sustainability production function would account for ecologic benefits that carry over to
following years and the unrecognized "costs" to farmers for waste disposal.
A8.1-4
-------
A8.1.7 Data Collection
NASS enumerators currently ask farm operators for crop yield information at the end
of the growing season. NASS would be willing to add questions on crop yield to a
December survey for EMAP. Data on the moisture content of soybeans are currently
available from NASS. Estimates of the moisture content of crops could also be obtained
from the literature. Data on fertilizer and pesticide inputs and cropping practices could be
obtained from a December enumeration survey conducted by NASS for EMAP, using
questionnaire formats from previous NASS surveys (Example questionnaire for pilot in
Appendix 6). Harvest indices to calculate NPP could be obtained from the literature.
A8.1.8 Estimated Costs
To be arranged with NASS. Would include the cost of a December survey plus costs
in retrieving, manipulating and analyzing the data.
A8.1.9 Data Variability
Variability (error) in yield estimates within a sample unit based on growers estimates
will be minimal. Biological variability in yield among sample units within a region can
range from 10-30% of the mean yield (Huff and Niel, 1982). The differences in crop yields
observed within a region are due to (in order of importance): inherent soil fertility, plant
variety used, temperature and rainfall of that season, N added, other fertilizers added, and
management (e.g., timing) (R.E. Jarrett, NCSU Small Grains Extension Specialist, personal
communication). Variability in crop yields across the U.S. may be up to 30% of the mean
yield due to regional differences in the variables listed above; however, data will be
interpreted mainly within, and not among, regions. Year-to-year variability can be large,
mainly due to seasonal rainfall and temperature. Variability in estimates of moisture
contents would be insignificant.
Estimates of conversion factors (e.g., harvest index) used to convert commercial yield
to NPP must be selected on a regional basis. This is mainly because of the differences in
crop varieties and type of varieties that are planted (e.g., soft vs. hard wheat, determinate vs.
indeterminate soybeans). Estimates of the conversion factors within a region were not found
to significantly affect overall NPP (Turner, 1987). She found that a 10% increase in harvest
indices brought about increases in NPP that ranged from 0-5%, but since NPP is calculated
from many different crops, the index of a particular crop does not significantly affect the
overall NPP. Initially, conversion factors in the literature can be used and updated over time
as better estimates are published.
A8.1.10 Interpretation
Net primary productivity values might be compared to several of the available
A8.1-5
-------
ecological models which predict NPP for a given physiographic region. NPP estimates are
generally based on elevation, temperature, rainfall, soil type, etc. and are more developed for
forest ecosystems than for annual plants. Nevertheless, they would provide a baseline for the
"potential" productivity of a region. This approach has been used by Turner (1987), who
compared her estimates of agricultural and forest productivity in Georgia with estimates of
NPP from various ecological models of productivity. An integrated EMAP regional
assessment of productivity could be attempted using data from EMAP Agrpecosystems,
Forests, Arid Lands and Wetlands.
The USDA/ERS and the Council for Environmental Quality evaluates productivity
by comparing inputs and output indices to the value of a base year (usually 1977), chosen
somewhat arbitrarily as a "normal" production year (e.g., no drought, or change in
agricultural policy). The baseline year is changed about every 10 years (Canter 1986, CEQ
1981, USDA/ERS 1988). Another approach is to express inputs and outputs in units of
energy (e.g. kcal) used to produce the inputs related to the energy value of the crop (Evans
1980, Klepper et al 1977, Pimentel 1980).
Interpretation of crop productivity will have to take into account several climatic and
socio-economic forces, including variability due to the plant variety (genotype) planted, the
direct and indirect of weather (e.g., soil moisture levels, planting data, pests and diseases)
and federal policies which result in land taken out of production. Usually the marginal land
is taken out first, resulting in greater productivity per unit land area.
Examples of basic questions that could be addressed with crop productivity data
include:
1. What is the spatial pattern of crop productivity across a region?
2. How does this pattern compare with patterns in soil types and prime farmland?
3. Are quantities of pesticides changing significantly to sustain yield?
4. Are quantities of irrigation water changing significantly to sustain yield?
A8.1.11 Future Work
We need to test out the statistics and mathematics of a crop productivity indicator
using existing data. It is not known yet if the data on outputs and inputs can be meaningfully
combined into a single index, or if they should be analyzed separately. It is not known yet
what units will be most appropriate (monetary, weight, volume, energy, e.g. kcal).
It's not known yet how to quantify environmental impacts or beneficial effects of
good management for computation of a sustainability index. Different ideas need to be
tested with existing data.
A8.1-6
-------
A8.2 Soil Productivity
A8.2.1 Endpoint: Sustainability, Contamination of natural resources
A8.2.2 Indicator category: response, exposure
A8.2.3 Status: research, active
A8.2.4 Description and rationale
The concept of soil productivity as an indicator is to find quantitative ways of
determining the status of soil health in terms of the physical, chemical and microbial
components of soil. It is anticipated that ranges for specific physical, chemical and
biological components can be worked out for soil types in different climates which
could be used to assess the general status of agricultural soils in the U.S. on a
regional basis.
The four components of the soil productivity indicator are nutrient-holding
capacity, erosion, contamination, and a microbial component. Nutrient-holding
capacity is intended to be a measure of the ability of a soil to hold nutrients, which is
determined by physical and chemical factors such as the organic matter content,
cation exchange capacity, pH, texture, etc. Soil erosion is a measure of the amount
of soil displaced from the field. Contamination refers to the amounts of industrial
or agricultural contaminants such as trace metals or persistent pesticide residues. The
microbial component is intended to be a measure of a microbial process or
population that indicates a healthy and functioning microflora.
It is not yet known if data from the four components could or should be
combined into a single index of soil productivity. Separate analysis of the data on all
four components of soil productivity would also provide meaningful and important
information. This issue will be explored further in 1991 and 1992.
A8.2.5 Data Needs
(1) nutrient-holding capacity NASS sample and
pH contract laboratory
analysis
exchangeable acidity for all data needs
total C
total N
totals
specific conductance
cation exchange capacity (CEC)
exchangeable cations
extractable P
A8.2-1
-------
bulk density
coarse fragment
% sand, silt, clay (texture)
soil permeability
infiltration rate
hydraulic conductivity
organic matter (ignition method)
Ancillary data:
use of manure/kind/amount
tillage practices
use of fertilizers and micronutrients
(2) contaminants
As, Cd, Cr, Hg, Ni, Pb, V
Ancillary data:
use of sewage sludge and amount applied
use of manure/kind/amount
(3) erosion
rill/sheet erosion (USLE):
SCS soil type classification
T value
length of field
slope of field
rainfall factor (R)
soil credibility factor (K)
crop/vegetation factor (C)
management/mechanical factor (P)
Ancillary data:
tillage practices
NASS question
NASS question
NASS question
NASS sample
NASS question
NASS question
SCS office
NRI SOILS-5 dataset
EMAP measure?
EMAP measure?
SCS office?
NRI SOELS-5 dataset
NASS question
NASS question
NASS question
(4) microbial component
soil organic matter as surrogate until further developed
NASS sample
A8.2.6 Index period
Soil samples for physical and chemical analyses will be taken in December by
NASS enumerators. Field data needed to calculate soil erosion could be taken by the
enumerators at that time. It is not known yet when soil samples for microbial
analyses will be taken.
A8.2-2
-------
A8.2.7 Data Collection
NASS enumerators could collect soil samples for chemical, physical and
microbial analysis according to a protocol developed by EMAP.
Several techniques of estimating and interpreting rates of soil erosion are
being evaluated for their suitability for a regional monitoring program such as the
agroecosystem component of EMAP. These include the Universal Soil Loss
Equation (USLE; Wischmeier and Smith 1978) and sediment delivery ratios, the
revised USLE (RUSLE) (Renard et al. 1990), 137Cesium distribution in soils
(Ritchie and McHenry 1990), the Erosion Productivity Impact Calculator (EPIC)
(Williams et al. 1983), the Water Erosion Prediction Project (WEPP) (Stone et al.
1990), wind erosion equations (Skidmore 1988), and the interpretation of aerial
photography and other remote sensing products to estimate gully erosion.
Initially, the data needed to calculate soil erosion with the USLE will be
collected in order to standardize with the SCS who currently uses USLE for their
NRI survey. The SCS is actively evaluating the RUSLE and WEPP models and is
likely to implement WEPP in the mid-1990's.
A8.2.8 Estimated Cost
Cost for the enumerators to take the sample will be negotiated with NASS as
part of the December survey conducted by NASS for EMAP. The main expense will
be in the chemical and physical laboratory soil analyses. Cost for the analysis at
different laboratories vary greatly and are being investigated.
A8.2.9 Variability
The coefficients of variation (CV) for bulk density are 2-17%; organic matter
42-125%; porosity, 4-18 %; 15 bar water content, 18-87%; saturated hydraulic
conductivity, 48-320%; and infiltration rate, 40-97% (Smith et al. 1987). Further
analysis of variability will be conducted on soil chemical variables in 1991 on
existing data and on data from soil samples collected in North Carolina in December
1990.
The variability associated with USLE estimates has not yet been determined
by the ARG. The spatial and temporal variability associated with soil microbial
measurements is likely to be high, and is anticipated to be a limiting factor as to the
choice of measurement.
A8.2.10 Interpretation
A8.2-3
-------
Soils could be grouped by multivariate techniques to assess the regional
distribution of soils based on their structure, nutrient-holding capacity, degree of
fertility, acidity, salinization, contamination, etc. It is anticipated that this
information could be used as an essential data layer in the assessment of multiple
indicators, such as crop productivity, land use, and agricultural chemical use.
The ARG, or the cross-cutting EMAP soils group, needs to determine ranges for
good, marginal and poor values of the different physical, chemical and biological
variables. It is likely that no single range can be applied to all soils, so we need to
determine the criteria for different soil groups (probably based on some level of
taxonomy) for which ranges will be established.
A8.2.11 Future Work
We need to develop appropriate measurements of wind erosion and gully erosion;
the Wind Erosion Equation used by SCS is being considered. The extent of land area
affected by gully erosion could possible be estimated from aerial photographs.
We need to evaluate to what extent EMAP could and should use soil erosion data
from SCS/National Resource Inventory.
It is not known yet what an appropriate measure of microbial status would be.
A8.2-4
-------
A8.3. Name: Land Use and Landscape Descriptors
A8.3.1. Endpoint: Quality of Agricultural Landscapes
A8.3.2. Category: Response / Stressor
A8.3.3. Status: Research
A8.3.4. Description, Rationale and Interpretation
As described in the indicator section regarding the Quality of Agricultural Landscapes,
the spatial structure of a landscape affects the flow of energy, material, and organisms across its
components (Forman & Godron 1986; Turner 1989). The land use and landscape descriptors are
a set of numerical values which provide a quantitative description of a landscape. (Table A8.3.1)
Table A8.3.1: Description and rationale for landscape descriptors.
Measure
Amount of land in a given use or cover
Proportion of land in a given use or cover
Diversity indices
Distribution of patch sizes
Distribution of patch dissection index
Landscape dissection index
Nearest neighbor analysis
Contagion index
Fractal dimension
Habitat Linear Classification System
Habitat Layer Index
(Appendix 8.4)
Description and Rationale
Overall land use and cover data.
May be correlated with various ecological and socio-
economic processes and other indicators.
Overall diversity (heterogeneity) of landscape.
May be correlated with ecological processes and other
indicators.
Abundance and size of patches.
Some flora and fauna may have minimum patch size
requirements.
Shape of patches.
Some species have preferences for edge habitat; others
require interior habitat.
Fragmentation and dumpiness of landscape.
May be correlated with the ability of energy , materials,
organism and disturbances to move through a landscape.
Broad-scale measure of pattern.
May be correlated with intensity of agriculture, as well as
other ecological processes.
Vertical structure of habitat.
May be correlated with species habitat require.nents.
A8.3 - 1
-------
summarizes the landscape descriptors currently being explored by the Agroecosystem Resource
Group.
General Framework for Analysis and Interpretation
Using NASS segments as a representative sample of a region, the value of the various
land use and landscape descriptors can be calculated for each region sampled. The relationship
between these values and the values of other EMAP indicators may be explored, and hypotheses
generated concerning the relationships between spatial pattern and ecological and socio-economic
processes. A framework for these analyses, based on an agricultural landscape hierarchy
(Lowrance et. al. 1986) is proposed (Figure A8.3.1). Field-sampled indicators are correlated with
landscape data to develop hypotheses regarding the interaction of landscape pattern and
ecological process.
Agricultural Landscape
Hierarchy
Data Analysis
Farm Field
Field-sampled Indicators
J
NASS Segment
Land use and landscape descriptors
based on NASS segments
Watershed / Region
J
Broad-scale landscape characterization from
EMAP Landscape Characterization Database
(as available)
Watershed / Regional analysis of
correlations between landscape patterns
and ecological processes
Continent / Nation
J
Continental / National Analysis
Figure A8.3.1: General framework for the analysis of land use and landscape descriptor data.
A8.3 2
-------
Amount and Proportion of Land in a Given Use or Cover
A very simple way in which to describe a landscape is to provide the area of land devoted
to each of a set of possible land uses or covers. In order to more easily compare landscapes, the
proportion, p,, of land use (or cover) i may be obtained by dividing the area in use / by the total
area of the landscape in question.
Land use information will be used to analyze the reallocation of land between agricultural
and other land uses as well as among agricultural uses. Changes in land use are likely to reflect
trends in crop preferences, socio-economic conditions, and grower perceptions concerning market
opportunities and profitability. Land use patterns may also be correlated with other indicators and
ecological and socio-economic processes. For example, Turner (1987) associates an increase in
net primary productivity in Georgia landscapes with changing land uses.
Diversity Indices
A variety of diversity indices may be found in the literature (Magurran 1988). All are
fairly easy to calculate given the proportions of all types of land use (or cover) for a particular
landscape. Two common diversity indices are shown below.
Simpson's Index
Simpson's Index (Simpson 1949; Magurran 1988), developed to address the question of
species diversity, gives the probability of any two individuals drawn at random from an infinitely
large community belonging to different species as:
Pi = proportion of individuals in the z'th species.
The value of D ranges between 0 and 1 and decreases with increasing diversity (it is often
expressed as l-D or 1/D), and is highly weighted towards the most abundant species.
The index may be used to described landscapes by defining p- as the proportion of land
in use or cover category /. A completely homogenous landscape has an index of 1; values less
than 1 would reflect the degree of heterogeneity in the landscape, with maximum heterogeneity
as D -> 0.
A8.3 3
-------
Shannon-Wiener Index
The Shannon-Wiener Index is given by:
H = -Ep,log(p,)
p; = proportion of landscape in use or cover /
The Shannon-Wiener index has been used to describe the heterogeneity of landscapes (e.g.,
Barrett et. al. 1990) and accounts for both the evenness and the abundance of land use categories.
The value of the Shannon-Wiener index is 0 (no diversity) if a landscape is completely
homogenous. The value is maximal if all possible uses (or covers) are present in exactly the same
amount. In such a case, the value would be -log(l/L), where L is the number of land use (or
cover) categories possible.
An additional measure of diversity, called the Shannon-Wiener evenness measure, is
defined as the ratio of observed diversity to maximum diversity:
LI TJ
E = = "
E ranges between 0 and 1 where 1 represents the situation in which all land use categories are
equally abundant, and 0 represents a completely homogenous landscape.
Both the Simpson and Shannon-Wiener indexes are based on the following assumptions:
o All of the species (land use categories) of interest are represented in the sample.
o The data are from a random sample of an effectively infinite population.
Further, if index values for several landscapes are to be compared they must all be measured at
the same spatial resolution. Because NASS segments are designed to be a random sample of a
county's agricultural lands, and there is a complete accounting of all land within each segment,
the indices may be calculated and compared on a county-by-county basis. Another approach
would involve the calculation and comparison of diversity indices on a regional basis. These two
approaches would reflect changes in diversity at different scales. The validity and robustness of
the assumptions, as well as the various ways in which to calculate and interpret these indices for
landscapes, will be tested using simulations and historical data during the Pilot Program.
A8.3 4
-------
Distribution of Patch Sizes
The size of a woodlot has often been correlated to the species richness found within it
(e.g., Freemark and Merriam 1986, Lynch and Whigham 1984, Gottfried 1978). In these studies
a correlation was established between the abundance of species sensitive to fragmentation and
the size of remnant patches. They are based on the theory of island biogeography (MacArthur
and Wilson 1967) which states that, given two environmentally similar islands, the larger will
contain more species than the smaller. Although originally developed for oceanic islands, the
theory has been applied to remnant island patches in a "sea" of dissimilar habitat. These concepts
may also be applied to wetland, grassland or other patches found in agroecosystems.
Given the areas of a set of patches, a distribution of patch area by frequency may be
plotted and statistical measures generated. The mean and variance of patch area will provide a
measure of the distribution of patch sizes within the landscape. These values may also be
compared from landscape to landscape and correlated with other indicators.
Dissection Index
The simple area measures described above do not address the question of patch shape.
Some species, such as white-tailed deer, are edge adapted species; others, such as the spotted owl
(Gutierrez and Carey 1985), prefer deep forest interior habitat. A circular patch with a 1 km.
radius provides about the same area as a rectangular patch 200 m. wide by 15 km. long (about
315 ha.). However, the narrow rectangular patch is presents a far larger edge (30.4 km.) than the
circular patch (6.28 km.) and, while useful to an edge-adapted species, may be of no use to
species requiring interior habitat. The movement of energy and material across boundaries, and
the ability of a patch to serve as a corridor for species movement may also be sensitive to the
shape of a patch (Forman & Godron 1986).
Patton's (1975) Dissection Index is designed to provide a measure of the area of a patch
relative to its perimeter. Because a circle has the smallest possible perimeter for a given area, the
index compares the perimeter of a random polygon to that of a circle of equal area. The
Dissection Index, DI, for a patch is calculated by:
A = area of patch
P = perimeter of patch
The minimum value of DI is 1 (for a circle) and there is no upper limit. Patches with larger
amounts of edge relative to area have larger DI values.
A8.3 5
-------
A Dissection Index can also be calculated for an entire landscape by (Bowen & Burgess
1981):
landscape ~ ,
AT = total area of all patches of a given type
PT = total perimeter of all patches of a given type
Nearest Neighbor Analysis
Calculation of nearest neighbor probabilities requires that the landscape be overlaid by
a grid and each grid cell assigned a land use or cover category (as in a raster-based geographic
information system). Nearest neighbor probabilities, qtj, represent the probability that cells of land
use category j are adjacent to cells of land use type j (Turner et. al. 1989). They are calculated
by dividing the number of cells of type / that are adjacent to cells of type j by the total number
of cells of type i. A landscape with very large patches of type / will have a relatively high
-------
Calculation of Fractal Dimension for a Landscape
The fractal dimension (Mandelbrot 1977), D, of the boundaries of a set of patches in a
landscape is obtained by using the perimeter and area of the patches to approximate D in the
equation (Krurnrnel et. al. 1987; Milne 1988; Turner et. al. 1989):
P = cADf2
P = perimeter
c = constant
A = area
D = fractal dimension
Taking the log of the above equation gives
log(P) =
Regression of logC/5) against log(A) for the data from a set of patches in the landscape results in
a line with a slope which is one-half of the fractal dimension, D, of their boundaries (Krummel
et. al. 1987).
The value of the fractal dimension, ranges from 1 for simple geometric shapes to 2 for
plane-filling shapes. The higher the dimension, the more complex the boundaries of the shapes
are. DeCola (1989) found that the fractal dimension of agricultural regions tends to be inversely
related to the intensity of cultivation. Krummel et. al. (1987) used the fractal dimension of forest
patches to formulate hypotheses concerning the spatial scale of pattern-process interactions.
A8.3.5. Data Needs
Land use and cover categories of interest to the Agroecosystem Resource Group are
shown in Table A8.3.2. Table A8.3.3 shows the minimum data required to calculate each of the
landscape descriptors. Detailed land use and land cover data in a geographic information system
format are required to calculate the full suite of landscape descriptors. A more detailed
description of data the ARG has requested of the EMAP Landscape Characterization Group may
be found in Appendix 4, Landscape Characterization.
A8.3 - 7
-------
Table A8.3.2: Land use and cover categories of interest to the Agroecosystem Resource Group.
Agricultural Land Uses and Cover
Cultivated land, by crop
Permanent pasture
Land in set-aside programs
Fallow land
Managed animal production
Farm ponds
Border areas (hedgerows, shelterbelts, etc)
Grassed waterways
Buildings and paved areas
General Land Uses and
Urban or built-up
Transportation
Range
Agriculture
Forest (identify dominant
Surface water
Cover
species)
Wetland (identify dominant species)
Arid lands (identify dominant species)
Barren
Table A8.3.3: Landscape descriptors that can be calculated as more detailed data become available.
Available data
Area of each land use and cover type
+ Area on a patch-by-patch basis
+ Perimeter of each patch
+ Data in a geographic information system
Descriptors which can be calculated
Amount and proportion in use / cover
Diversity indices
Distribution of patch sizes
Distribution of patch dissection index
Landscape dissection index
Fractal dimension of patches
Nearest neighbor analysis
Contagion index
A?.3 - 8
-------
A83.6. Index Period
NASS collects land use data annually during the June Enumerative Survey.
A8.3.7. Data Collection
The area of each land use and cover type in all sampled segments will be routinely
available from the NASS June Enumerative Survey through an inter-agency agreement with
NASS. These data are collected by enumerators who visit each sampled segment annually and
account for all land within the segment. NASS also has aerial photographs (1:12,000 scale) of
all sampled segments. Although the details have yet to be completed, the area and perimeter of
all patches may be obtained through analysis of these photographs. Digitizing these images for
a geographic information system would provide the data required to calculate the complete suite
of landscape descriptors.
The EMAP Landscape Characterization Database is intended to provide detailed
characterization data for large areas of the nation and will be available in a GIS format
(ARC/INFO). This would allow the calculation of landscape descriptors at the watershed or
regional level. However, the EMAP Landscape Characterization program is undergoing extensive
revision as a result of a June, 1990 peer review of the characterization process. At this time we
do not know what data the EMAP Landscape Characterization Group will ultimately provide to
the other Resource Groups. (Appendix 4)
The ARG is also exploring the use of GIS data developed by the North Carolina Center
for Geographic Information and Analysis. A partial list of data available from that source may
be found in Appendix 4 (Table A4.2).
A83.8. Estimated Costs
The additional cost of analyzing and digitizing the aerial photographs of each NASS
segment has not been determined.
A8.3.9. Data Variability
A precise statement of variability cannot be made at present. Known sources of variability
are:
o field collection of agricultural land use data
o interpretation of land use or cover from aerial photography
o digitizing imagery for use in a GIS
A8.3 - 9
-------
A8.3.10. Interpretation (See A8.3.4.)
A8.3.11. Future Work
The concepts described above will be further developed and tested during the
Agroecosystem Pilot Program. Research will continue on the relationships between landscape
pattern and ecological process. Specific activities planned include:
o continuing to explore the literature for more and / or more appropriate land use and
landscape descriptors
o work with NASS to develop procedures to obtain the full suite of land use and landscape
descriptors from NASS images of the sampled segments
o work with the EMAP Landscape Characterization Group to ensure that ARG requirements
are satisfied and to develop procedures for landscape analyses
o work with the North Carolina Center for Geographic Information and Analysis to
determine how the ARG may utilize GIS data they have developed
o develop procedures for the calculation of land use and landscape descriptors using the
ARC/INFO geographic information system (ARC/INFO is the EPA standard)
o identify and analyze historical data for correlations between the land use and landscape
descriptors and wildlife populations
o explore the relationship between field-sampled indicator data collected during the Pilot
Program and land use and landscape descriptors
A8.3 - 10
-------
A8.4. Name: Habitat Linear Classification System / Habitat Layer Index
A8.4.1. Endpoint: Quality of Agricultural Landscapes
A8.4.2. Category: Response / Stressor
A8.4.3. Status: Research
A8.4.4. Description, Rationale and Interpretation
The Habitat Linear Classification System (HLCS) is being developed by the US Fish and
Wildlife Service (Short 1990) to assess the suitability of habitat for particular wildlife species.
It has been successfully applied to bird species and guilds in areas of the western United States
(Short 1990). The HLCS provides detailed information about the vertical structure of vegetation
for all terrestrial ecosystems and for wetland ecosystems with emergent vegetation. It has been
proposed for application in all EMAP ecosystems.
As developed by Short (1990) an HLCS sampling unit is a 45 x 45 meter quadrat
containing 25 9 x 9 meter sub-quadrats. Each sub-quadrat is sampled at 25 points, for a total of
Main Quadrat
HLCS Sampling Grid |
Sub-Quadrat
45 meters
* Sample all layers
+ Sample first vertical meter only
9 meter square
Figure A8.4.1: The HLCS sampling grid, with main quadrat, sub-quadrai and sampling
points,
A8.4 1
-------
725 sample points (Figure A8.4.1). The HLCS is based on vegetation intercept data collected by
placing an 8 m. long pole in a vertical orientation at each sampling point and recording 1) the
type of substrate upon which the pole rests, 2) the type of vegetation that the pole intercepts in
1 m. intervals, and 3) the type of vegetation that occurs directly above the pole. A variety of
upland, wetland and vegetation traits have been identified as inventory data and a classification
process has been developed (Short 1990) (Table A8.4.1). Data are analyzed for density and
cl'impiness and may be compared from site to site, or at the same site ovei time, and related to
known habitat requirements for individual species.
Table A8.4.1: Substrate and vegetation types used in the HLCS.
Code
Class
WATF
WATB
WATS
Water, Fresh
Water, Brackish
Water, SaJty
WEAR
WROC
WLIW
WLIL
WHBG
WEBF
WHBO
Wetland Bare (mud, sand, soil, gravel - no vegetation)
Wetland Rock (>65mm)
Wetland Woody Litter (on surface > 10cm; stumps < 1m)
Wetland Other Organic Litter (on surface < 10cm; includes leaves)
Wetland Herbaceous (grass, grasslike)
Wetland Herbaceous (forbs)
Wetland Other Organic (algal, aquatic moss, vascular, etc.)
MANO
UBAR
UROC
ULIW
UHBO
UHBG
UHBF
UHBO
CACT
Manmade Object
Upland Bare (sand, soil, gravel - no vegetation)
Upland Rock (> 65mm)
Upland Woody Litter (on surface > 10cm; stumps < 1m)
Upland Other Organic Litter (on surface < 10cm; includes leaves)
Upland Herbaceous (grass, grasslike)
Upland Herbaceous (forbs)
Upland Other Organic (ferns, mosses, lichens)
Cactus Stems and Pads
BLOW
BLNW
NLDW
NLNW
LBLD
LEND
Woody Broadleaf Deciduous
Woody Broadleaf Non-deciduous
Woody Needleleaf Deciduous
Woody Needleleaf Non-deciduous
Woody Liana: Broadleaf Deciduous
Woody Liana: BroadJeaf Non-deciduous
DESW
SNAG
BOLE
Dead Standing Woody
Number of snags (> 25cm dbh) in subplot
Number of living tree boles (> 25cm dbh) in subplot
Source: Short 1990
A8.4 - 2
-------
The HLCS is an extension of earlier work at Fish and Wildlife on a Habitat Layer Index
(HLI) (Short and Williamson 1986). The HLI is a related measure that can also be used as an
indicator to evaluate habitat structure in order to monitor wildlife diversity. The Index was used
in the Second Resource Act Conservation Act (RCA) Appraisal (USDA 1989) as part of an
evaluation of the condition of the nation's wildlife habitat. The method, based on ground-truthed
interpretation of aerial photography, maps the vertical habitat within a sample site into the
following layers:
tree canopy
stem bole
midstory
terrestrial surface
terrestrial subsurface
water surface
The extent of each layer within the sample site is determined. In forested regions, the lower
layers are inferred (with appropriate ground-truthing) from the dominant tree species, the degree
of canopy closure, and the height class of the trees.
The vertical complexity of the habitat at a sample site is described by a numerical index,
the Habitat Layer Index, calculated from information about the extent of each habitat layer. The
Index compares the layers of habitat actually measured at a site to a theoretical maximum that
would occur if all layers were present uniformly. The value of the Index ranges from 0 (no
structural habitat) to 1 (all habitats uniformly available throughout the site). Short and Williamson
(1986) give the equation for six habitat layers as:
//!/=—^-
/ = the number of layers of habitat present at the site
n = the number of different cover types at the site
At = the area of layer / at the site
B = the surface area of cover type j at the site
The numerator represents the total area of all layers present at the site. The denominator
represents the theoretical maximum area of all layers at the site. It is the product of the maximum
number of habitat layers (6) that could occur at the site and the maximum area of those habitat
layers (5 times the area). Although originally developed for the six habitat layers shown above,
the equation can be modified to accommodate any number of layers. In fact, the Second RCA
Appraisal (USDA 1989) considered seven layers in its evaluation, adding water column to the
categories. The HLI may also be calculated using data collected for the HLCS.
A8.4 3
-------
A8.4.5. Data Needs
Data required for both the HLCS and the HLI are described in the Description, Rationale
and Interpretation section above. Data for the HLCS would be collected from hedgerows,
shelterbelts and woodlots of appropriate fields on each segment sampled by the ARG. The
intensity of sampling required is not known.
New aerial imagery of each segment would be required at regular intervals; probably
every 3-5 years.
A8.4.6. Index Period: Not yet determined
A 8.4.7. Data Collection
Data for the HLCS might be collected by NASS enumerators. Members of the ARG
would develop a detailed procedural manual and participate in the training of NASS enumerators.
The aerial photographs of each NASS segment could be interpreted for calculation of the
HLI. NASS currently does not update this imagery on a regular basis. New photographs are
acquired when 1) the enumerator reports major changes in the segment or 2) the sampling frame
is updated.
A8.4.8. Estimated Costs
The time required to collect HLCS data for a quadrat depends upon the vegetative cover
on the quadrat. It ranges from a minimum of four hours for a sparsely vegetated quadrat, up to
two days for a densely forested quadrat.
The cost of updating the aerial photographs on a regular basis is unknown.
A8.4.9. Data Variability: Not yet determined
A8.4.10. Intepretation (See section A8.4.4.)
A8.4.1L Future Work
An HLCS workshop was hosted by the Agroecosystem Resource Group in July, 1990. The
workshop was given by Hank Short, and attended by representatives of the EMAP
A8.4 - 4
-------
Agroecosystem and Arid Lands Resource Groups, USDA/NASS, and the North Carolina Wildlife
Resources Commission. All present agreed that the methodology, properly refined and uniformly
applied, would be a valuable tool for assessing and monitoring the condition of the nation's
wildlife habitat. Many issues were raised by the participants, among them:
o statistical questions about the HLCS sampling technique
o the high cost of collecting the data, especially in densely wooded areas
o the incorporation of plant species data
o simplification and standardization of the sampling procedure
o uniform application and coordination of the HLCS across EMAP Resource Groups
o appropriateness of HLCS for landscapes that have fairly uniform vertical structure, such
as some desert and agroecosystem landscapes
o the potential for developing remote sensing techniques to collect HLCS data
o data interpretation and presentation
The EMAP Forests Resource Group attempted to apply a variant of this technique during
their 1990 field pilot program. The results of this study are not yet available.
If resources are available, the EMAP Agroecosystem Resource Group intends to test
alternative ground sampling procedures during 1991. Statistical sampling questions will also be
addressed.
A8.4 - 5
-------
A8.5 Irrigation Water Quantity
A8.5.1 Endpoint: Sustainability of commodity production
A8.5.2 Indicator category: exposure
A8.5.3 Status: research, active
.\8.5.4 Description and Rationale
Irrigation water quantity will be a measure of agricultural irrigation extent and
usage. This indicator also includes availability of irrigation water and impacts to
groundwater levels, either directly as a result of pumping for irrigation, or as a result
of depletion of the recharge capacity of the system. The result of expanding
irrigation is often the depletion of the groundwater. Such effects have had serious
implications for water management in many areas of the country.
A8.5.5 Data Needs
Source of data
acres irrigated NASS question
type of irrigation system (flood, sprinkler) NASS question
volume of water applied NASS question
source of water NASS question
availability of water NASS question?
Ancillary data:
cost of irrigation water NASS question
if the water is drained, type of drainage system
flow rate?
Many of these data are already being collected by the Geological Survey (Regional
Aquifer System Analysis); the Department of Commerce (Farm and Ranch Irrigation
Survey) and by the Bureau of Reclamation in the western U.S. (Summary Statistics:
Water, Land and Related Data). Many states report on irrigation water use.
A8,5.6 Index Period
Data supplied by farm operators could be collected during the December
survey conducted by NASS for EMAP.
A8.5.7 Data Collection
A8.5-1
-------
NASS enumerators can obtain many of the listed data needs in an
enumerative survey with the farm operators. Some questions, such as the availability
of irrigation water,
will be difficult to formulate in a way that consistent, meaningful data will be
obtained.
Irrigation water is sometimes supplied and managed by a contractor. It is not
known how the ARG will be able to obtain the data in these cases.
A8.5.8 Estimated Costs
Would be part of the NASS survey done for EMAP in December. Cost is to
be negotiated with NASS.
A8.5.9 Variability
Estimates by farm operators are likely to be quite accurate. Variability among
years will depend on the climate and may be very low in dry areas where crops are
always irrigated.
A8.5.10 Interpretation
Analyses could include spatial patterns of irrigated acreage in a region and on
specific crops. The amount of water used, the availability of water and the type of
irrigation methods would be an indicator of the efficiency of water use and the
sustainability of the system. Correlations among spatial patterns of groundwater
levels and irrigation water use could be attempted. Crop productivity, land use, soil
productivity and irrigation water use can be compared in data layers.
Many of these data are being collected by the Departments of Interior and
Commerce. Some effort needs to go into identifying the compatibility of the
databases; for example, the Bureau of Reclamation keeps track of all water diverted
for irrigation.
A8.5.11 Future Work
Existing datasets need to be more closely examined for use in data analysis
approaches. The ARG needs to work out a meaningful way to collect data on the
availability of irrigation water.
A8.5-2
-------
A8.6 Irrigation Water Quality
A86.1 Endpoint: Sustainability; Contamination of natural resources
A8.6.2 Indicator category: exposure
A8.6.3 Status: research, active
A8.6.4 Description and rationale
It is known that water retained in agricultural soil becomes progressively
more saline over time, especially in poorly drained soils. This process is believed to
be responsible for the failure of many irrigation projects through history (Carr 1966,
OTA 1983). Irrigation water of poor quality can affect the permeability and aeration
of the soil, and affect plants through the presence of phytotoxic substances or through
the modification of processes that limit the water uptake by the plant (OTA 1983).
Surveys by the USDA before 1985 showed that 2.9 million of California's 10.1
million irrigated acres show signs of salt damage (Maranto, 1985).
In both arid and non-arid irrigated ecosystems, the drainwater from
irrigated soils often contains a higher salt burden, and may contain toxic
concentrations of trace metals (e.g. cadmium, selenium) and persistent organics,
which are recognized threats to wildlife (OTA 1983, USGS 1986). Data will be
collected on the type and concentrations of selected chemicals in the irrigation
water and in the drainwater.
A8.6.5 Data Needs
Source of data
electrical conductivity (EC) NASS or EMAP measure
pH, Cl, NO3, Se, B NASS or EMAP sample
Ca, Mg, sulfates NASS or EMAP sample
total dissolved solids (TDS) NASS or EMAP sample
agricultural chemical residues, e.g. atrazine NASS or EMAP sample
(may include heavy metals)
Ancillary data:
water temperature NASS or EMAP measure
cost of irrigation water NASS question?
type of irrigation system (flood, sprinkler) NASS question
volume of water applied NASS question
source of water NASS question
flow rate?
A8.6.6 Index Period
A8.6-1
-------
Must be sampled while irrigation water is available at the field. In most
regions, this is likely to be mid-summer. Worst-case situations for irrigation water
quality would exist in the fall, when water is at low flow, and the dilution factor is
low. In drainage ponds, water should be sampled later in the summer (late July or
August), after constituents of concern have leached from the soil.
A8.6.7 Data Collection
Water samples from flowing water before and after diversion from the
drainwater (or holding ponds) would be taken.
Electrical conductivity (EC) is an easy measurement taken in the field. NASS
enumerators could take this measurement with little training. Since EC is
temperature dependent, it would be necessary to take water temperatures at the same
time.
A knowledge of water quality sampling procedures would be required to
collect samples. A sampling protocol will be developed in 1991. Water sampling is
generally done by skilled technicians. The ARG will have to carefully develop a
protocol for water sampling based on the experience of water quality experts. It is
anticipated that NASS enumerators will take irrigation water samples in the pilot tests
after a thorough training session on the protocol.
Sample analyses must be done at a laboratory that has demonstrated
adherence to quality assurance/quality control practices (e.g., has participated in
interlaboratory comparison programs). Both EC and TDS measurements have been
taken for many years by the USGS, and the techniques and quality of data obtained
are good.
If drainage water is to be sampled, water samples should be taken upstream
and downstream of an irrigated agroecosystem of interest, and drainwater should be
collected from the appropriate pond(s). Drainwater from the drains may also be
collected as necessary to identify specifically the source of contamination. To be
consistent with the current monitoring efforts of the Department of Interior Task
Force on Irrigation Water Quality, methods from the USGS (1977) should be used.
A8.6.8 Estimated Costs
The cost of collecting the sample will be negotiated with NASS. Because the
sampling must be done at the time of irrigation, it is likely that data collection will
not be a part of the December survey or an established NASS survey. This would
mean a special visit to farm operations by the enumerators to collect the water
sample.
A8.6-2
-------
Laboratory chemical analyses of water samples will be at least $40 per
sample.
A8.6.9 Data Variability
Variability for TDS and EC will be low enough that results would be
meaningful from year to year. The most important factors that would influence
variability would be flow and temperature. Samples should be taken at a specified
time each year to minimize effects from flow differences.
The variability in trace elements, pesticides, nutrients, and other elements
would likely be quite high seasonally and spatially. The data should be adjusted for
flows. For example, although the concentration of selenium may be high from a
particular drain, the actual volume or amount of selenium draining into a drainfield
may be only moderate, i.e., the selenium may be concentrated in a small amount of
water
A8.6.10 Interpretation
The characteristics of water most often considered when determining
irrigation water quality include (1) TDS and EC; (2) the proportion of sodium to
calcium and magnesium; and (3) boron, chloride and sulfate content (McKee and
Wolf 1963; U.S. EPA 1976). Major contaminants in irrigation drainwater of
ecological concern are increased salt burdens, fertilizer and pesticide residues,
synthetic industrial organics, selenium, and trace metals. Whether water quality is
good or poor may be initially determined by comparing the levels of the analyte
found in the irrigation water with the allowable levels established by the EPA or
states. For example, in Utah, 50 micrograms of zinc per liter is the aquatic wildlife
standard (USGS Report 88-4011). A site exceeding this standard could be considered
poor quality and the proportion of irrigated land within a region receiving (or
draining) good or poor quality irrigation water could be calculated.
Initially, baseline values for salts, trace elements, pH, dissolved solids and
agricultural chemicals should be established for each region. After these have been
established, trends of ecosystem condition with respect to irrigation water quality can
be analyzed.
Irrigation water quality should be integrated with several of the other
indicators under development, such as crop productivity, agricultural chemical use,
soil productivity, including soil sodium and soil salinity (soil would be a better
indicator for infrequent monitoring), and land use (e.g., tillage). Comparison of
irrigation drainwater concentrations versus the concentrations in natural streams of
the same region could be made.
A8.6-3
-------
Indices could be established to project the expected levels of elements in
water based on such information as the local geology, soils, natural water
concentrations, etc. Observed versus expected levels of any of the elements could be
determined. Indices from the Soil Conservation Service have already been proposed
to look at retention of pesticides in soils (reference here?). These could be tested as
well.
Useful, but not critical, ancillary data are hydrology, soil type, soil
permeability, depth to groundwater, groundwater quality, and leaching fraction data
(Corwin et. al, 1990). These data would help ascertain the susceptibility of the soil to
salinization and to determine what constituents are present that could be leached. For
example, it may be important to know that the geological parent material of a region
is shale, since some shales are high in selenium. Impermeable layers also need to be
identified, and whether or not the site is drained. The hydrological setting of the site
should be clearly identified (i.e., identification of water supply, agricultural
diversions and return flows, and the impact of groundwater on the water supply).
Flow information will be needed in order to calculate volumes of salts and elemental
constituents. Precipitation and temperature could be used to identify the potential for
and impact of evapotranspiration.
A8.6.11 Future Work
Numerous studies have examined salinity and toxicity in irrigation water and
drainwater that could provide more information for the development of this indicator.
The USDA has operated the Salinity Laboratory in California for decades, and there
are specialists there that can provide information on sample variability and sample
timing. It is not yet known if NASS or EMAP will take the samples, and how
difficult it will be to work out an appropriate protocol. This will receive high priority
in 1991.
Individuals involved in the Irrigation Drainwater Program should be
consulted regarding their findings, which include: (1) if there is a closed system,
there will most likely be an elemental problem, (2) the constituent of most concern is
selenium, (3) areas where the source rocks are seleniferous (e.g., shales) will
probably have selenium problems, (4) water quality is greater in areas of high
precipitation, (5) the elemental problems are generally site-specific. Findings from
monitoring programs by the USGS, such as National Stream Quality Accounting
Network, should also be integrated into this effort.
A8.6-4
-------
A8.7 Agricultural Chemical Use
A8.7.1 Endpoint: Sustainability; Contamination of natural resources
A8.7.2 Indicator category: stressor
A8.7.3 Status: research, active
A8.7.4 Description and Rationale
The agricultural chemical use indicator is a quantitative measure of rates and
spatial and temporal distributions of applied chemicals. It is a major variable in
determining potential ecological impacts of agrichemical use.
Complete data on actual chemical use in U.S. agroecosystems do not yet exist.
The data that are available are estimated values and do not include all chemical types
and crops. The most widely-used data base is the National Pesticide Usage Database
developed by Resources for the Future, Inc. (RFF) with support from EPA, USGS
and the USDA (Gianessi, 1985). The database contains estimates of the average
annual use of selected pesticide active ingredients by crop and county. Usage data
are developed from expert opinions from various states, including Extension Service
personnel, and USDA usage surveys. Estimates are calculated using two coefficients:
the percentage of acres that are treated and the average annual application rate per
treated acre, and are keyed to planted crop acreage estimates reported in the Census
of Agriculture (U.S. DOC 1990).
EMAP data on agricultural chemical use is intended to supplement these
existing data sets with data on actual usage, frequency, rate, etc. Data on agricultural
chemical use will be used by the ARG to calculate crop productivity indices
(input/output ratios) and to estimate the potential nontarget environmental impacts of
fertilizer and pesticide use.
A8.7.5 Data Needs
Source of data
type, rate and frequency of fertilizer use NASS question
type, rate and frequency of pesticide use NASS question
type, rate and frequency of herbicide use NASS question
(Questions on agricultural chemical use in the NASS questionnaire for the EMAP
pilot include time of application, material used, quantity applied, units, and how
applied. See Appendix 6 for example questionnaire).
Ancillary data:
costs of chemical inputs literature
total acreage where chemical was applied NASS question
A8.7-1
-------
crop NASS question
row spacing, planting date? NASS question
A8.7.6 Index period
Data could be collected from the farm operators by NASS enumerators during
a December survey conducted by NASS for EMAP.
A8.7.7 Data Collection
NASS enumerators
A8.7.8 Estimated Cost
Would be part of the December survey conducted by NASS for EMAP. Cost
for the survey will be negotiated with NASS.
A8.7.9 Expected Variability
Accuracy of collected data will depend on farmer's willingness to cooperate,
share data, and participate in record-keeping. Year-to-year variability will depend on
climate factors (temperature and rainfall), type of crops planted, cost of inputs, and
pest populations.
A8.7.10 Data Interpretation
A data base on actual pesticide and fertilizer use would be a valuable database
to establish. Spatial relation to other indicators, such as nonpoint source pollution
loading or wildlife populations is known to be difficult. However, usage is the
primary variable driving the ecological effects of agrichemicals.
Agricultural chemical use can be a major determinate of crop yield. The
amount of fertilizers and other agrichemicals added can mask deterioration of soils
due to erosion and intensive monoculture. The amount of inputs needed to maintain
yield will be explored as an indicator of sustainability.
A8.7.11 Future work
Approaches to data analysis and interpretation need to be explored with
existing data.
A8.7-2
-------
A8.8. Nonpoint source loading
A8.8.1 Endpoint: Contamination of natural resources
A8.8.2 Indicator category: stressor
A8.8.3 Status: candidate, active
A8.8.4 Description and Rationale
The nonpoint source loading indicator is a quantitative measure of multi-
media chemical transport losses within and from the agroecosystem to surrounding
ecological resources and is indicative of the efficient use of inputs into
agroecosystem. Nonpoint source loadings include agricultural chemicals (fertilizers,
pesticides and their breakdown products), animal wastes, eroded soils and perhaps
genetically engineered organisms.
Nonpoint source pollution is characterized by highly variable loadings with
rainfall and other environmental characteristics dominating the timing and magnitude
of transport. Chemicals are exported from their site of application to nearby streams
and lakes by runoff and subsurface flow, leaching to groundwater, drift from aerial
and ground application equipment, chemical dust transport and
volatilization/deposition to and from the atmosphere. Agricultural land management
practices such as tillage, tile drainage and channelization impact the export of
chemicals, sediments and other nonpoint source loadings via soil erosion, runoff and
leaching. The use of reduced tillage alters export losses and causes shifts in the
receiving media. Additional concerns are being raised relative to chemical transport
of dust from agricultural tillage operations.
Nonpoint source loading as a quantitative indicator presents problems in
measurement and interpretation. Sampling approaches within the EMAP design are
good for sampling resources that do not change rapidly over time and that are
relatively homogeneous within some defined spatial scope. Nonpoint source
loadings, however, either change rapidly in value as a function of time or are highly
variable spatially. Others are event driven.
Because of the time and space dependencies, it is unlikely that actual
leaching, runoff, and volatilization can be accurately monitored in a single index
period. However, the ARG will attempt to use existing modeling efforts by U.S.
EPA and other groups to identify the soil, climate, chemical use and land
management variables that drive pollution loading phenomena, and measure these
variables as an estimate of the potential for nonpoint source loadings on a regional
basis (Table 6.6)
A8.8-1
-------
A8.8.5 Data Needs
(1) initial surrogate: nitrates and atrazine EMAP sample
in adjacent streams, groundwater, farm ponds
and irrigation drainwater
(2) initial surrogate: soil displaced/field see soil erosion
(3) initial surrogate: Key variables driving nonpoint source loading as identified by
current modelling efforts will be evaluated for their ability to predict potential
nonpoint source loading (Smith et al. 1990, Table 6.6).
Additional variables will to be considered include, pesticide degradation rates,
sorption partition coefficients, soil characteristics (such as soil series, horizon depth,
bulk density), depth to water table, and daily weather records of precipitation,
evaporation, max/min air temperature.
A8.8.6 Index Period
Not yet determined because it depends on which variables are chosen. For
leaching, groundwater samples might be considered rather than soil cores.
A8.7.8 Estimated Cost
Multiple pesticide (or degradation products) analysis may cost up to $200 per
sample. Analytical costs associated with fertilizer elements (nitrogen, phosphorus,)
would be considerably less.
A8.8.9. Variability
Not known yet and will depend on the variables chosen to monitor. No
information on the variability associated with runoff or large-scale groundwater
systems has been found yet by the ARG.
A8.8.10 Data Interpretation
Mathematical models have become useful tools for predicting movement of
chemicals from agricultural sites to surrounding environmental media. Several
models for transport and transformation have been developed during the last decade.
These include an PRZM (Carsel et al. 1984), RUSTIC (Dean et al. 1989), GLEAMS
(Leonard et al. 1987) and TEEAM (Dean et al. 1989). The models could be used to
evaluate monitoring requirements and to estimate potential nonpoint source loads to
A8.8-2
-------
other ecosystems. Ranges for high, medium and low potential for nonpoint source
loadings could be worked out. The proportion of agricultural land in a region with
high, medium and low potential for nonpoint source loading could be computed.
Spatial patterns of nonpoint source loading potential could be developed by
GIS overlay techniques with climate, soil, agricultural chemical use, and land
management data.
A8.8.11 Future Work
Determine key parameters from transport models. Determine best index
periods. Determine data analysis and interpretation approaches.
A8.8-3
-------
A8.9. Biomonitors
A8.9.1 Endpoint: Contamination of natural resources
A8.9.2 Indicator Category: exposure, response
A8.9.3 Status: Honey Bees research, active
Earthworms candidate, active
Ladino clover research, active
Lichens & mosses candidate, active
A8.9.4 Description and Rationale
The long-term monitoring of chemical contaminants in the air, water and soil, as
well as exposure and response measurements of selected biomonitoring organisms can be
viewed as a connecting link between the abiotic and biotic components of the
environment, providing information on contaminant exposure, fate, bioavailability and
effects; and therefore leading to a more meaningful assessment of contamination of
natural resources.
Biomonitoring refers to the use of biologically-based measurements of one or
more variables of a biological unit to determine spatial and/or temporal trends in the
condition of the variable(s) in order to make assessments of environmental quality.
Advantages in using biomonitors include: 1) integration of physical and chemical factors
(stressors) to provide a more realistic assessment of stressor exposure, 2) relatively
inexpensive compared to sophisticated instrumentation required to monitor
environmental media, 3) some species occur over broad areas allowing for the
development of regional (and possibly national) networks, and 4) tissues can be archived
(dried or frozen) for future analyses.
Although there are several organisms that could be used as biomonitors, currently
the following four biomonitors are being evaluated for use in EMAP-Agroecosystems: 1)
honey bees, 2) earthworms, 3) ladino clover, and 4) lichens or mosses.
An assessment endpoint of concern for agroecosystems is the contamination of
both cultivated land and the immediate surrounding natural resources. The four
biomonitors recommended can be used to directly address contamination from
agricultural management practices (e.g., pesticides, trace metals) and general urban-
industrial pollution (synthetic organics and trace metals) transported via the atmosphere
or surface waters.
Questions that can be asked from biomonitoring data include:
o What proportion of agroecosystems contain biomonitors with enhanced levels of
trace metals and persistent synthetic organics?
A8.9-1
-------
o What regions of the country contain biomonitors with enhance levels of trace metals
and persistent synthetic organics?
o In what proportion of agroecosystems are biomonitor population size, mortality, and
productivity adversely impacted?
o What regions of the country are biomonitor population size, mortality and
productivity adversely impacted?
o What are the magnitudes of spatial and/or temporal changes in the level of
bioavailable contaminants in agroecosystems?
(1) Honey Bees
Honey bees (Apis mellifera L.) have been used effectively as biomonitors of
agroecosystems, where they have been shown to be effective bioaccumulators of trace
metals, radionuclides, pesticides and industrial organics. Standardized protocols exist for
the use of bees as exposure monitors and for laboratory bioassays of toxicity. Well
developed guidelines have been provided for in situ assessments of a wide array of
population responses, and a standardized protocol for response assessments is being
developed under an EPA cooperative agreement (Bromenshenk, J.J., personal
communication).
Forager bees, in search of food and water, often accumulate various contaminants
such as fluoride, trace metals, radionuclides, and synthetic organic chemicals, including
persistent agricultural chemicals as well as industrial contaminants such as
polychlorinated biphenyls (Anderson and Wojtas 1986, Bromenshenk and Preston 1986,
Bromenshenk et al. 1985, Morse et al. 1987, Smirle et al. 1984, Wallwork-Barber et al.
1982). Exposures to many of these chemicals can result in dysfunction or death of
individual bees, but only in severe cases does the bee colony perish. Thus, the colony is
relatively rugged and long-lived, yet it responds to pollution stress by alterations in
reproduction, population size, shortened life span, mortality, behavior, and productivity
(e.g., stores of pollen, nectar, and honey).
Honey bees are being considered as both exposure and response indicators. As an
exposure indicator, forager bees will be used to determine the bioavailability of
contaminants present in various environmental media (air, water, and soil). As response
indicators, honey bee population parameters (mortality, reproduction, incidence of
disease) and honey production can be used to determine overall quality of their
environment. Since there a thousands of potentially contaminants, it is essential to
identify specific contaminants (e.g., persistent organics used by farmer in vicinity of
sampling area) or categories of contaminants that are of interest.
(2) Earthworms
A8.9-2
-------
Because of their common occurrence in the majority of soils and their intimate
contact with the soil, earthworms have been shown to be ideal biomonitors for assessing
soil contamination of trace metals and persistent organics (Neuhauser and Callahan 1990,
Callahan 1988, Ebing et al. 1984, Neuhauser et al. 1984, Beyer and Gish 1980, Wright
and Stringer 1980, Ireland 1979, Ruppel and Laughlin 1977, VanHook 1974).
Earthworms have been selected as a key indicator organism for ecotoxicological testing
of the toxicity of industrial chemicals, not only by the European Economic Community,
but also by the Organization for Economic Cooperation and Development, the Food and
Agriculture Organization of the United Nations, and the Food and Drug Administration.
The U.S. EPA is currently finalizing both a laboratory and field protocol for using
earthworms for assessing the impact of toxic chemicals to the soil ecosystem (C.A.
Callahan, personal communication).
Earthworms will be sampled from cultivated fields and pasture acreage and
possibly adjacent areas to determine tissue contamination (trace metals and persistent
toxic organics), species composition and abundance.
(3) Ladino Clover
A number of higher plants (e.g., tobacco, potato, alfalfa, cotton, ragweed,
soybean) have been successfully used as biomonitors of air pollution (Manning and
Feder 1980, Posthumus 1976). Recently, a monitoring system using ozone (03) resistant
(R) and sensitive (S) clones of ladino clover (Trifoliwn repens L.) has shown promise as
a useful indicator of ambient 03 concentrations. In studies conducted to date, the R and
S clones have been grown under several 03 regimes where response parameters
(biomass, foliar injury and leaf chlorophyll) between the clones were compared. Growth
of the R clone is not affected significantly by seasonal ambient 03 levels (May to
August; 12 h day mean of 50 to 90 ppb), while growth of the S clone is decreased 20% to
50%. The S/R ratio of these three measures was related to the 03 concentration and
could be calibrated to show their relationship to yield of important crop species as well as
ambient 03 concentrations (A. L. Heagle, personal communication).
The rationale for recommending the R and S clones of ladino clover includes: 1)
lack of genetic variability due to the use of clonal material, 2) easily propagated, 3)
grows under all conditions suitable for agriculture, 4) a ratio of S/R can be generated for
the response parameters (biomass, foliar injury and leaf chlorophyll) to indicate doses of
ozone during the growth period.
The ladino clover S/R method could serve as a monitor of the effects of ambient
03 levels on the yield of important crop species. It may also be useful to develop a
production index that would include short-term climatic variables. This system could
A8.9-3
-------
provide information on the temporal and spatial distribution of the occurrence of elevated
levels of 03 and possibly other atmospheric pollutants.
(4) Lichens and Mosses
Lichens and mosses are ideal organisms to assess exposure to airborne chemicals
since they absorb most of their nutrients (and contaminants) directly from the atmosphere
(wet/dry deposition). To determine if airborne contaminants have an influence on
epiphytic lichens or mosses, it is necessary to decide which parameters of the lichens are
to be investigated. In areas where air contaminants are subnominal, 1) fewer epiphytic
lichen species will be present (due to the disappearance of sensitive species), 2) the
species that are present will be poorly developed, and 3) the percentage cover and the
frequency of the sensitive species present will be less in contaminated areas.
Because epiphytic lichens or mosses are to be used to determine the influence of
airborne contaminants, the influence of all other environmental factors need to be
excluded to the degree possible. This is a difficult problem since not all of these factors
are known and consequently they cannot all be taken into account. Important factors
affecting lichen distribution that need to be considered are tree species (or genus) and
microclimate
A8.9.5 Data Needs
(1) Honey Bees
contaminants in forager bees
contaminants in honey
number of colonies
honey production
bee mortality
bee disease
Source of data
NASS sample or ABF
NASS sample or ABF
NASS question or ABF
NASS question or ABF
NASS question or ABF
NASS question or ABF
(2) Earthworms
contaminants in earthworms
number of species
number of earthworms/species
condition of integument
NASS sample
NASS sample
NASS sample
NASS sample
(3) Ladino Clover
biomass
foliar injury
chlorophyll concentration
NASS sample
NASS measure
NASS sample
A8.9-4
-------
(4) Lichens and Mosses
contaminants in tissues EMAP sample
number of species EMAP measure
frequency of occurrence EMAP measure
percent cover EMAP measure
Ancillary data:
agricultural insecticide use NASS question
crop type, yield and acreage NASS question
urban-industrial toxic emissions TRI
air quality data Various sources
meteorological data NWS
urban-industrial sludge use NASS question
tillage practices NASS question
irrigation water quality EMAP indicator
soil type classification SCS
agricultural pesticide use NASS question
landscape descriptors EMAP measure
A8.9.6 Index Period
(1) Honey Bees
Annual samples of forager bees and honey made in late summer would suffice for
monitoring contaminants, such as trace metals and persistent agricultural and industrial
organics. For response indicator measurements, routine beekeeper or apiculture inspector
reports (mortality, poor performance, etc.)could be used. On-site monthly measurements
of reproduction, forager longevity, population size and productivity is advisable but is
likely to be beyond the resources available to EMAP.
(2) Earthworms
Earthworms would be sampled once per year. An early fall sampling period
would be adequate for exposure and response measurements.
(3) Ladino Clover
The starting date for placing ladino clover in the field will be standardized for a
given agricultural area, but will probably be at least partly dependent on latitude.
Biomass (leaf chlorophyll and foliar injury estimates optional) will be made at or
A8.9-5
-------
immediately following the final crop harvest at each sampling site. These parameters
could be taken more often (e.g., monthly intervals), but at an additional cost.
(4) Lichens and Mosses
For use as exposure indicators, resident lichen or moss species should be sampled
once annually in the early fall. For response measurements, a 4 year measurement cycle
should be adequate.
A8.9.7Data Collection
(1) Honey Bees
Information about bees (number of colonies, yields, problems, etc.) can be
obtained from various bee organizations such as the American Beekeeping Federation
and the American Honey Producers. Other sources of information are state regulatory
agencies. Some states maintain data on colony strength and condition (for beekeepers
who intend to rent bees as pollinators). Many states require beekeepers to register the
number of colonies that they own or control and the sites where they are deployed. A
survey of beekeepers could possibly be performed by NASS enumerators.
(2) Earthworms
Samples of 0.1 m^ will be taken for measurement of the abundance and number
of earthworm species, as well as observations of the condition of the integument
(swellings and lesions). Soil will be wet with 0.22% formalin (Raw 1959) and
earthworms will be collected over a 30 minute period as they emerge from the soil.
After 30 minutes, the 0.1 m^ soil core will be removed from the soil and sorted by hand
for remaining earthworms. Whole body burden measurements of contaminants (suite of
trace elements and synthetic organic residues) will be measured from a subsample of the
collected earthworms from each sample core. The ARG is exploring the possibility of
NASS enumerators collecting the samples.
If problem areas are determined with the above sampling, on-site placement of
earthworms could be used to monitor acute effects such as mortality, condition of the
integument, and the uptake and retention of the contaminant(s) in earthworm tissue over
a specified sampling period. Enclosed, permeable field containers (e.g., nylon) with a
common species and number of worms could be monitored in both field crop and
surrounding locations. Other observations could include growth, enzyme induction (e.g.,
metallothionen) and enzyme depression (e.g., cholinesterases).
A8.9-6
-------
(3) Ladino clover
At each sampling period, biomass (above ground only) measurements will be
made on each 15 liter pot for each clone. For each 15 liter pot, clover will be cut at a
height of 5 cm above the soil level. The stems and leaves will be placed in pre-labeled
bags. The bags (one per pot) will be oven-dried and weighed. Leaf injury ratings
(chlorosis and necrosis in 5 % increments) and leaf chlorophyll (a and b) contents are
optional. Foliar injury estimates require trained persons (0.5 hr per site). It is not known
who would collect this data.
(4) Lichens or mosses
At each sampling period, data for the following parameters would be collected: 1)
number of epiphytic lichen species, 2) frequency of occurrence and 3) percent cover.
Lichen or moss samples would be collected at the same time for trace metal (ICP) and
persistent organic contaminants (GC/MS) analyses. It is not known who would collect
the data.
A8.9.8 Estimated Cost
(1) Honey Bees
Collecting response data and samples for exposure analyses for honey bees has
the potential to be the least expensive of the biomonitors recommended, as there exists
the possibility of enlisting the assistance of bee organizations.
Bee hive/colony costs for the initial screening could be passed onto the
beekeepers (just use their existing colonies). If bees are not available at a sample
location, either renting them ($10- $35 each) or the use mini-colonies ($15-$30 each)
would be an appropriate solution.
The sampler/observer needs to have coveralls, gloves, a bee veil, and a smoker.
Also needed are a vacuum ($0-$30/site), a nozzle ($l-$2/site), sample bags ($1.00/site),
and preferably dry ice (10 pounds per day ($0.25 1.00/site).
Inorganic chemical screening may run from $15-150 per sample, depending on
number of chemicals, detection limits, etc. Organics start at $60 and can exceed $1,000.
(2) Earthworms
The major on-site efforts would be in treating each sampling area with formalin
and in the subsequent digging and sorting through soil for earthworms. If proper care
A8.9-7
-------
were taken for storing earthworms (cool, moist) during collection, response
measurements could be taken under laboratory conditions by trained personnel.
Earthworm samples collected for tissue analyses would be relatively straight
forward (again avoiding inadvertent contamination) - the major cost would be the actual
preparation and analyses.
(3) Ladino Clover
Biomass measurements involve collection, drying and weighing plant tissue. The
costs are not yet known.
(4) Lichens and Mosses
Collecting data on lichens as response indicators would be very labor intensive
and would require individuals trained in lichen identification and sampling techniques.
Possibly a list of easily identified lichen species (sensitive to tolerant) could be prepared
with color photographs to make identification of species easy for non-taxonomists.
Lichen samples collected for tissue analyses would be relatively straight forward
(protocol for avoiding inadvertent contamination of the sample would need to be
developed) — the major cost would be the actual preparation and analyses. For all the
biomonitors used as exposure indicators, a necessary first step (at least for synthetic
organic contaminants) would be to limit the number of organic residues that would be
analyzed.
A8.9.9 Variability
(1) Honey Bees
Variability for honey bees as exposure and response indicators is not known at this time
scientific literature needs to be assessed for determining expected variability.
Some contaminant values (ppm/dry weight) measured in forager bees in the Pacific
Northwest, New York and Florida are (Bromenshenk, J.J., personal communication):
Element
Arsenic
Cadmium
Fluoride
Lead
Zinc
Low
0.1
0.1
4.3
0.2
66.0
High
18.5
6.9
405.0
298.7
222.8
(2) Earthworms
A8.9-8
-------
Variability for earthworms as exposure and response indicators is not known at
this time.
(3) Ladino Clover
It is assumed that there will be less variability for clones than for cultiv.ars. From recent
research data, the coefficient of variation for the ladino clover S/R ratio for biomass was
10% over three harvest periods.
(4) Lichens and Mosses
Variability for lichens and mosses as exposure and response indicators is not
known at this time.
Some contaminant values (micrograms/gram dry weight) measured in lichens are (T.H.
Nash III, personal communication):
Element Low Hi eh
Arsenic
Cadmium
Mercury
Lead
Zinc
Nickel
0.26
0.05
0.00
3.60
9.10
0.00
8
330
29
12,000
25,000
300
A8.9.10 Interpretation
(1) Honey Bees
The purpose of using honey bees as exposure/response indicators is to determine
if there are elevated concentrations of specific contaminants or adverse responses to
chemical stresses from their activities within a multi-media environmental context (air,
water, soil, vegetation). For example, an increase in bee mortality or a decrease in honey
production for a given agricultural region may be indicative of increased exposure to
contaminants, especially if tissue concentrations were elevated and there was evidence of
a contaminant emission source impacting the area (e.g . TRI data, air quality data).
However, other factors, such as adverse weather (e.g., drought), disease, or the
availability food (crop type and acreage), need to be considered for interpretation. The
BEEPOP and BEEKILL models, as well as the BEETOX data base may be very useful
in assisting in data interpretation.
A8.9-9
-------
The home-range of forager bees are greater than a single field. Therefore, issues
on the different spatial relations between bee data and data collected on a single field will
have to be resolved.
(2) Earthworms
The purpose of using earthworms as exposure/response indicators is to determine
if there are elevated concentrations of specific contaminants or adverse responses to
chemical stresses from their close association with soil. Interpretation of the data will
require correlative analysis of earthworm and ancillary data listed in Section 8.9.5. For
example, a decrease in the abundance of earthworms through time in a given agricultural
region may be indicative of increased exposure to trace metals in soil amendments such
as sewage sludge, especially if soil and tissue concentrations were elevated in trace
metals (e.g., Cd) and there was a history of sewage sludge use for this area. However,
other factors, such as adverse weather (e.g.j drought), tillage practices, or insecticide use
need to be considered for interpretation.
(3) Ladino Clover
Ladino clover is intended as an active, diagnostic indicator of ambient 63
concentrations in agricultural areas, as well as a surrogate for estimating crop yield losses
due to 63. Research to date as demonstrated the S/R ratios decrease as the 03 dose
increases. Once the ladino clover system has been adequately tested, confidence limits
will be available for estimates of ambient 63 concentrations to S/R ratios, as well as
estimates of yield loss of crops for which 03 dose-yield loss information exists.
Interpretation of the data will require correlative analysis of the S/R ratios with other
potential stressors or modifiers, such as CC>2 levels, other air pollutants, or disease.
(4) Lichens and Mosses
The purpose of using epiphytic lichens or mosses as exposure and response
indicators is to determine if there are elevated concentrations of specific airborne
contaminants or adverse population responses to these airborne contaminants
A decrease in the frequency of occurrence of certain lichen species through time
for a given agricultural region may be indicative of increased exposure to airborne
contaminants, especially if tissue concentrations were elevated and there was evidence of
a contaminant emission source impacting the area (e.g., TRI data, air quality data).
However, other factors, such as gradual climate change, succession, or physical alteration
of the natural habitat, resulting in modification of the microclimate (e.g., a few edge trees
cut resulting in increased light penetration, increased daytime temperatures and decreased
relative humidity for the more interior habitat).
A8.9.11 Future Work
A8.9-10
-------
(1) Honey Bees
For exposure monitoring, the main problem is twofold: 1) many contaminants
may have subtle affects but the contaminant in question may be transitory (will not
persist in the bee tissue), and 2) scaling-up to monitor at a national scale. More
information about the physical/chemical properties of synthetic organic contaminants is
required to focus on chemicals that will indeed persist in tissues. Calibration of bees to
more conventional ambient monitors could be improved. This might be accomplished by
co-location of high volume air samplers and bee colonies.
Additional effort are also needed to: 1) determine the extent and quality of
historical exposure and response data collected by state and agricultural agencies, and 2)
determine the possibilities of enlisting cooperation from the American Beekeeper
Federation for a national monitoring network.
(2) Earthworms
As with honey bees, the main problem with using earthworms is: 1) many contaminants
may have subtle affects but the contaminant in question may be transitory, and 2)
scaling-up to monitor at a national scale. More information about the physical/chemical
properties of synthetic organic contaminants is required to focus on chemicals that will
indeed persist in tissues. Additional attention needs to be given to the evaluation of
biomarkers that will reveal subtle effects from contaminant exposure and that could be
cost effective at a regional/national level.
Departures from nominal trends in response indicators, (e.g., decline in
earthworm species and numbers) will help pinpoint problem areas, but determining the
potential causes for the departures will need more intensive research. Even if the
observed responses can be correlated with increased contaminant tissue concentrations,
one will still have to account for some of the other possible causes. More information is
required on the influence of soil type, tillage, weather, predators and disease on
earthworm population dynamics.
(3) Ladino Clover
Current and future research under laboratory and field conditions will address the
following issues: 1) refine ozone exposure-response relationships, 2) determine the
effects of climate and edaphic factors on (^-induced S/R ratios, 3) test the relative
sensitivity of S and R clones to SC»2 and other air pollutants, 4) test the sensitivity of the
S and R clones to diseases and biological pests, 5) calibrate S/R ratios with ozone-
induced yield losses of major agricultural crops, 6) test whether different growth media
used for pot culture affect the S/R ratios, and 7) determine if conducting measurements at
final harvest (rather than monthly) will result in an appreciable loss of information.
A8.9-11
-------
Using the ladino clover system as an indicator of 63 effects on important
agricultural crops will require research data to allow calibration of clover clones to crops
response to 63 under a wide range of environmental conditions.
(4) Lichens and Mosses
The main problem associated with the use of lichens and mosses are similar to
those discussed for honey bees and earthworms.
Because variability in tissue contaminant residues is potentially high, factors
explaining this variability must be addressed. Age will be an important consideration for
determining tissue contaminant concentrations. Additional factors to consider are
substrate and microclimate.
As with earthworms, there is no species of lichens or of moss that would be
common throughout each region of the United States. Research is, therefore, required to
determine the feasibility of comparing tissue contaminant concentrations across different
species. Issues regarding the retention of trace metals and especially organics within the
lichen or moss need to be addressed.
A problem encountered by investigators using lichen distribution data to
demonstrate patterns of air quality is the fact that large amounts of data are collected and
it becomes difficult to interpret and summarize. Perhaps the best known procedure in
lichen floristic studies is to calculate the Index of Atmospheric Purity (IAP) from
variables describing the epiphytic lichen vegetation around urban areas or point sources
of air emissions. The IAP provides a numerical assessment of the degree of air pollution
based on the number, frequency and tolerance of the lichen species present on trees at
each site. The IAP procedure allows the data to be reduced to a level suitable for
mapping. This approach assumes that the substrate and microclimate of the sampling
sites are similar, which will not be the case for a regional/national monitoring effort.
Additional effort needs to be given to determine if an IAP approach is feasible in an
EMAP context.
Although data on chemical contamination of biological components (honey bees,
earthworms, lichens and/or mosses) is important, it must be weighed against the high
analytical costs to gain such data. Decisions on the number of sampling sites and
contaminants to be analyzed must be made.
A8.9-12
-------
A8.10 Pest Density
A8.10.1 Endpoint: Sustainability
A8.10.2 Indicator category: exposure
A8.10.3 Status: candidate, active
A8.10.4 Description of Indicator
Two components of pest density are considered for use: weed incidence and
density of plant-parasitic nematodes.
Weeds are defined as any plants present in an agricultural field other than the
intended crop plant. The number of weed species present, particularly noxious weed
species, indicates the effectiveness of weed management and cultural stresses, as well
as the potential stress to which the associated crop is subjected. In conjunction with
information on type and amount of herbicides applied and on tillage practices,
incidence of weeds provides a measure of efficiency of weed management. The
presence of certain noxious weeds, (e.g., sicklepod, witchweed) also indicates
potential for a high degree of continued stress on the system.
Plant-parasitic nematodes are stylet-bearing nematodes that depend on a
host plant for nutrition at one or more life stages. Parasitic nematodes are present in
most agricultural soils in the regions. Nematodes are currently of most importance in
the southeast, California and Hawaii. Specific nematode species are also important in
the midwest. Nematodes have a direct impact on yield of agronomic and
horticultural crops and the population levels of plant-parasitic nematodes can be
related to a crop damage threshold or nematode hazard index. In conjunction with
information on type and amount of nematicides applied and cultivars selected for use,
nematode population levels provide a measure of efficiency of pest management.
Increasing population density of nematodes in agricultural soils indicates a definite
decline in agroecosystem health.
Measures of weed prevalence and nematode population density, as indicators
of pest density, provide a direct measure of exposure of crop plants to a specific
biotic stresses. Yield and crop production efficiency can be affected by the density of
these pests. Additionally, presence or absence of particular species of weeds and
nematodes in conjunction with data on pesticides applied and tillage/rotational
practices indicates efficiency of management of pests, and, thus, the health of the
agroecosystem.
A8.10.5 Data Needs
Source of data
A8.10-1
-------
abundance and diversity of plant parasitic nematode genera NASS sample?
abundance and diversity of weed genera EMAP sample?
Ancillary data:
type, rate, and frequency of herbicide use NASS question
type, rate, and frequency of nematicide use NASS question
use of soil fumigants
crop variety NASS question
crop rotation history NASS question
temp and rainfall during growing season
stage of crop maturity at sampling
soil physical and chemical parameters
tillage practices
A8.10.6 Index Period
Weeds The most appropriate time for the survey of prevalence of
agricultural weeds would be early in the growing season (spring to early summer),
after crop emergence. The weeds that are present early in the crop season are those
that will have the largest impact on growth and productivity of the crop. Also, most
weed seedlings are easily recognized and are most easily noticed before crop plants
become too large. The survey can be conducted once per year.
Plant-parasitic nematodes - The most appropriate time for the determination
of density of plant-parasitic nematodes would be at or near crop maturity, the
rational for this is that nematodes are detected most efficiently when population
levels are highest, which occurs for nearly all nematodes near crop maturity. Most
indices that relate nematode population density to crop damage or potential crop
maturity. For most agronomic crops, this would be in the fall. The specific data is
not as important as the sampling window. Early September would be best for
northern regions and late September for southern regions. Population density assays
can be performed once per year.
AS.lOJData Collection
Weeds - The evaluator will need a pictorial diagram of the weed species of
interest in the seedlings stage and at one or more advanced stages of development and
a tally sheet for recording the presence/absence of each weed species. The diagram
will be limited to 10-15 specific weed species. It may be possible to develop a
national pictorial diagram or it may be necessary to develop regional diagrams. The
national diagram would be better from the standpoint of reporting and data
summarization. In one possible sampling scheme, the evaluator will begin the survey
at the "left" corner of the sample unit and proceed to walk through the sample unit at
an angle sufficient to intersect the midpoint of the opposite edge of the sample unit.
A8.10-2
-------
At that midpoint the evaluator will return to the "right" corner of the original edge of
the sample unit (as indicated on the aerial photograph). While walking along this V-
shaped path, the evaluator will survey the soil surface and tally the presence of each
weed species on the specified form. Time involved should be no more than 30
minutes for a sample unit.
Weeds - Evaluators such as the NASS enumerators with some knowledge of
farming practices can reliably collect the data. Some training will be needed on weed
identification; however, with the pictorial diagram training required will be minimal.
No further expertise is required since data will be tallied in the sample unit during the
site visit.
Plant-parasitic nematodes The evaluator will need a soil sampling probe
(20-25 cm long and 2.5 cm in diameter), a plastic bucket and a sample container
sufficient in size to contain 500 cm-* of soil. The evaluator will walk across the
sample unit in the same V-shaped pattern as for the weed survey. Individual soil
cores will be collected to a depth of 20 cm at 10 m intervals by vertical insertion of
the probe into soil in the root zone of the crop present. Soil will be shaken from the
soil probe into the plastic bucket as each soil core is taken. When all soil cores have
been taken, the soil in the bucket will be mixed by hand and a portion of
approximately 500 cm^ will be removed from the bucket and placed in the sample
container. The container will be labelled with the sample unit identification, data and
evaluator number. The soil sample will then be stored in an insulated "cooler" until it
is sent by mail to the laboratory for assay. Nematode assays will be performed with
the aid of a wet-sieving assay and microscopic identification procedure at a contract
laboratory in accordance with procedures and quality assurance standard to be
developed.
NASS enumerators will likely collect the soil samples. Training in the
specific methods of sample collection will be needed to ensure that samples are
collected uniformly among sample units. Processing of samples for extraction and
identification of nematodes will require the expertise of a technician with a full
knowledge and familiarity with the extraction process and extensive training in the
identification of nematode genera. Sample processing and nematode identification
must be done in a contract laboratory by trained personnel.
A8.10.8 Estimated Costs
Weeds - Total time for moving through the sample unit and tallying the
presence/absence of each weed species should not exceed 30 minutes. Collection of
ancillary data will require an additional 15-30 minutes.
Plant-parasitic nematodes - Total time for collection of soil cores will be 30
minutes (if combined with the weed survey it may be possible to reduce total time for
A8.10-3
-------
both pests to 45 minutes per sample unit). Preparation and packaging of sample for
shipment should be 10 minutes. Mailing costs should not exceed $1.00 per sample.
Extraction and identification of one sample should cost about $10.00 - $15.00, but
will depend on exact prices negotiated with the contract laboratory.
A8.10.9 Variability
Weeds - Because the information obtained is based on the presence or
absence of specific weed species, the major source of variation or error would be the
lack of detection of a specific species. The evaluator would cover the sample unit in a
V-shaped path to reduce the probability of non-detection. Misclassification is
another possible type of error; however, with proper training and the pictorial
diagram this possibility should be reduced.
Variation in number of major species that occur within fields or sample units
will occur across the country. The occurrence of some weeds is regional. Selection
of major, widely-occurring species of weeds will give a basis of comparison across
all regions, whereas selection of weeds of particular importance in major regions of
the country will give regionally important information.
Plant-parasitic nematodes Due to the aggregated spatial pattern of most
plant-parasitic nematodes, a great deal of variation is expected with a sampling unit.
Coefficients of variation among individual samples within fields can often exceed
100%. For this reason, numerous soil cores are usually taken, the soil mixed and a
portion of the mixture selected as the sample to represent a field. This tends to give a
mean value for the field and it is upon such values that damage thresholds are based.
Variation among repeated samples taken from bulked soil samples, which represent a
mixture of many soil cores, should be small. Across the country extreme variation is
expected with respect to occurrence of species and population densities. Nematode
population densities may rang from 0-1000+ individuals per 500 cm^ soil. In most
reporting of population densities and in most analyses, population densities are
transformed as log (x+1) where x is the population density obtained.
A8.10.10 Interpretation
Weeds - Number of weed species present will be divided into at least two
categories: noxious and non-noxious. Weed scientists are currently working to
obtain calibration information that will relate number of species present to anticipated
yield reductions for certain crops. Until such calibration information is available for
a sufficient number of crops, it may be possible to use an index of biological
diversity based simply on number of species present. Alternatively, it may be
possible to simply use actual species number within each category as a value.
A8.10-4
-------
Amount and type of herbicides applied will be needed as ancillary
information to interpret the weed prevalence indicator. Applications of most
herbicides are made in anticipation of the occurrence of specific types of weeds and
thus the need to apply such herbicides is also a type of weed indicator.
Plant-Parasitic nematodes Estimates of population density of plant-parasitic
nematodes will be interpreted in relation to established economic threshold values.
Classification of sample units will be on the basis of hazard of future damage for the
same crop as is currently in the field. Population densities can be divided into two or
three categories: low and high hazard (corresponding to nominal and subnominal,
respectively) or low, moderate and high hazard, it may be possible to present the
hazard as a probability of damage or a damage potential index based on a numerical
scale from 0 to 10 that includes more than one plant-parasitic species. The presence
of some species alone may be sufficient to classify the sample unit as subnominal or,
in other cases, the presence of a combination of species at certain population densities
may result in a classification as subnominal. An index of species diversity and
abundance may be a possibility for classification purposes.
Amount and type of nematicides applied will be needed as ancillary
information for the interpretation of population densities. Also, the perceived need to
apply nematicides may in itself serve as an indicator of anticipated problems due to
nematodes.
If indices are developed for weediness and nematode diversity and population
density, reporting can be done as cumulative distributions. If particular species of
weeds or nematodes are important within a specific region, maps of occurrence will
be important in reports. Tables of mean nematode population density by specific
nematode density categories would also be useful. For both the weed and the
nematode indicator, crop hazard values could be presented as proportion of sample
units in specific hazard categories - either in a bar chart or in a cumulative frequency
distribution.
A8.10.ll Future Work
Weeds A list of the most important weeds needs to be assembled by region
and the specific weeds to be monitored selected. Pictorial diagrams need to be
assembled for use by evaluators and tally forms developed. Also, the training manual
and precise protocols need to be written. Actual survey work should be relatively
easy to implement.
The exact relationship between weed prevalence and crop yield is not fully
established. Further research is needed to fully calibrate weed prevalence as an
indicator of crop yield. It may be necessary to develop an index of weediness based
upon number of noxious and non-noxious species present. Also, work is needed to
establish how herbicide use and weed prevalence will be used together in EMAP.
A8.10-5
-------
Plant-parasitic nematodes - A list of the most important nematodes needs to
be assembled by region and the specific nematodes to be assayed for in an EMAP
context determined. Threshold values are available for most plant-parasitic
nematodes; however, no comprehensive publication is available that summarizes all
of these thresholds. It may be necessary to develop an overall nematode hazard index
for EMAP that will reflect species diversity and population densities of important
nematodes species.
Several assay procedures are available for some species of nematodes. Also,
the selection of different nematode species for inclusion in the indicator will require
the use of several specific assays (e.g., wet-sieving, sugar flotation, mist extraction).
The exact procedures for sampling handling, assay and reporting must be established
and contract lab(s) found to perform the assays.
A8.10-6
-------
A8.ll Density and Diversity of Beneficial Insects
A8.11.1 Endpoint: Sustainability
A8.11.2 Indicator category: response
A8.11.3 Status: candidate, on hold
A8.11.4 Description and rationale
Insect predators and parasites play an important ecological role in regulating
natural populations of phytophagous insects. Data on the abundance and diversity of
beneficial insects might also be used as an indicator of diversity in the
agroecosystem. Insects respond to changes in habitat caused by fluctuating land use,
such as participation in federal, low-management conservation reserve programs.
A8.11.5 Data Needs
number of pests and number of beneficials/ft row EMAP or NASS
sample
number of parasitized pests/ft row EMAP or NASS
sample
Ancillary data:
type, rate and frequency of insecticide use NASS question
participation in CRP or other set-aside programs NASS question
crop variety NASS question?
crop rotation history NASS question
temp and rainfall during growing season NWS?
A8.11.6 Index period
Not yet determined
A8.11.7 Data collection
Not yet determined
A8.11.8 Estimated cost
Not yet determined
A8.11.9 Variability
Not yet determined
A8.ll.10 Interpretation
Not yet determined
A8.11.H Future work
Much of the background work on this indicator has not been done yet due to
lack of resources.
A8.11-1
-------
A8.12 Socioeconomics
A8.12.1 Endpoint: Sustainability; Quality of agricultural landscapes
A8.12.2 Indicator category: Stressor, response
A8.12.3 Status: candidate, on-hold
A8.12.4 Description and rationale
The ARG recognizes the importance of socio-economic factors in the health of
agroecosystems. However, due to limited resources, development of this indicator
has been limited to a few initial approaches, listed below. These ideas will be further
developed in 1991.
Social Structure and Population Shifts - Monitoring the number and type of social
institutions associated with rural farm life is an indicator of cultural changes caused
by shifts in farm policies and management practices. These in turn determine
population shifts, particularly emigration to urban areas. Land use patterns in
agroecosystems and other resource areas are impacted.
Number and Size of Farms - These are indicators of potential land use and
management practices.
Federal Conservation and Cropland Reduction Program Acreage - The number of
acres in these federal programs is thought to impact soil erosion, crop production, use
of agricultural chemicals, nonpoint source loading and wildlife habitat. Agro-EMAP
data can be used to determine the impacts of this policy.
Farm Program Financing Criteria - National aggregate indicator of where
landscaping, drainage, and crop production program money is being spent. Indicator
of land use, crop production and diversity, and soil erosion. Analysis of these criteria
can point to weaknesses in national policy areas.
Federal Commodity Programs - These include a) level of pure support, b) acreage
retirement programs, c) level of grain export subsidy, and d) base acreage system.
Federal commodity programs have generally resulted in increased acreage in
production, increased acreage in those crops which aggravate soil erosion; increased
fertilizer and pesticide use; and reduced the extent of rotation and diversity of crops
planted.
Indebtedness - Indebtedness, particularly if split out by capital vs land vs operation
should indicate how the system is functioning. Increase in indebtedness might
indicate decline in sustainability or expenditures for resource modification. It will
also reflect changes in the social or economic status of farm communities.
A8.12-1
-------
A8.12.5 Data Needs(tentati\e)
Source of data
farm size NASS question
number of farms calculation
rural population Bureau of Census?
social institutions Bureau of Census?
sodbuster/CPR acreage NASS question
ownership NASS question
federal commodity programs NASS question
indebtedness NASS/ERS question?
perceptions of land manager regarding the environment NASS question?
consumption of fossil fuels NASS question?
farm program financing criteria NASS question?
A8.11.6 Index period
Could be part of the NASS enumerative survey conducted in December for
EMAP.
A8.11.7 Data Collection
NASS enumerators will likely be able to collect much of the data required from
the farm operator.
A8.11.8 Estimated cost
The cost would be part of a NASS survey conducted for EMAP and will be
negotiated with NASS.
A8.11.9 Variability
Not determined yet
A8.ll.10 Data interpretation
Not determined yet
A8.ll.ll Future Work
Most of the background work on this indicator is not yet done.
A8.12-2
-------
A8.13 Livestock Production
A8.13.1 Endpoint: Sustainability; Contamination of natural resources
A8.13.2 Indicator category: response, stressor
A8.13.3 Status: candidate, on-hold
A8.13.4 Description and rationale
The agroecosystem should monitor the condition of livestock operations. As with
other aspects of agriculture, the definition of a "healthy" livestock system is difficult to
formulate. Traditional measurements of these systems tend to be economic in nature and
are measured in terms of profits. Although an ecologically oriented set of indicators
should be developed, animal productivity is the single indicator for animal systems in the
preliminary indicator list. Initially, production information will concentrate on poultry
and swine operations since 1991 pilot project for agroecosystems is planned for North
Carolina and swine and poultry are the two most important livestock operations in the
state.
Livestock operations, like crop agriculture, are highly managed with the goal of
maximum production with minimum input. The operator provides a certain amount of
food, water, and labor and expects a high livestock production. A certain amount of
waste is produced and exported to other ecosystems, which is part of the agroecosystem
nonpoint source pollution concerns.
Livestock operations suggest two useful indicators: 1) productivity, and 2) waste
export. Only productivity is addressed in this discussion. Productivity is an economic
barometer which quantifies the amount of product produced. Use of existing measures of
productivity assumes that a productive system is a healthy one. Although that is true
economically in the short-term, it may not be true ecologically speaking. A highly
productive operation which is exporting waste may not be an ecologically healthy
system.
Productivity is controlled by both economic and production criteria. Production
factors include conditions under which the animals are raised, the quality of the feed
(nutrition), and genetic advances. One way to determine the effects of production criteria
on productivity would be to define an ideal environment which would result in maximum
potential productivity. Research has been accomplished to optimize housing,
reproduction and nutrition independently. If the combination of these optima are linearly
additive, it might be possible to define an optimal environment. It would then be
possible to compare the deviations between productivity measurements from livestock
operations and the maximum potential productivity for use as a system health indicator.
A8.13-1
-------
A8.13.5 Data Needs
NASS currently collects annual data on numbers of livestock in the U.S.
Since productivity measurements differ with livestock, there is no universal
indicator of productivity which could be applied across all livestock operations. For
example, days to 230 pounds is a common measurement for swine operations (Mayrose
et al. 1985) but would be meaningless for a poultry producer. The poultry producer
would be more interested in eggs produced or pounds live weight at slaughter.
Reproductive and growth rate statistics for various livestock are also used and available
(USDA 1988).
Growth rate and live weight at slaughter are accepted indexes of productivity for
broilers. Rate of lay serves as the measure for egg- producing poultry. These statistics
are available from the USDA (3) on at least a state basis. It would probably be useful to
form some sort of production-to-input ratio as an indicator.
There are several commonly accepted measures of swine productivity. They
include live weight at slaughter, days to 230 pounds, feed efficiency, breeding efficiency
and pigs per litter (Mayrose et al. 1985).
A8.13.6. Index period
Flexible, could be part of the December survey conducted by NASS for EMAP.
A8.13.7 Data Collection
Data could be collected by NASS enumerators from the farm operator.
A8.13.8 Estimated Cost
Not yet determined
A8.13.9 Variability
Should be low within a given producer operation but probably varies more across
producers.
A8.13.10 Interpretation
Not yet determined
A8.13.ll Future Work
Much of the background work on this indicator needs to be done.
A8.13-2
-------
A8.14. Name: Genetic Diversity
A8.14.1. Endpoint: Sustainability of Commodity Production
A8.14.2. Category: Response / Exposure
A8.14.3. Status: Candidate / On Hold
A8.14.4. Description and Rationale
The status of crop and livestock genetic diversity is an important issue in agriculture. Its
importance increases daily as national and global developments continue to push wild crop
varieties and livestock breeds into extinction. From a purely anthropocentric viewpoint, the
conservation of crop and livestock genetic diversity is of value as a source of raw material for
continued genetic improvement. Loss of genetic diversity could seriously hamper efforts to feed
an expanding population in changing environmental conditions.
Some would argue that crop genetic material is adequately preserved in germplasm
collections. Yet our seed banks are far from perfect. Less than 50% of the seeds in the collections
contain the minimal passport information (latitude, longitude and altitude of origin). Once seeds
are removed from their natural environment they are no longer subject to natural selection
pressures, but rather to selection pressures imposed by storage programs.
The problem is more acute for livestock because, while cryopreservation of sperm is fairly
reliable, the technology for preservation of ovum and embryos is not fully developed. The only
reliable way to conserve the genetic diversity of these animals is by maintaining (at least)
minimal population sizes of each breed.
A8.14.5. Data Needs
Identification and enumeration of crop varieties in production and livestock breeds is
required in order to develop a crop and livestock genetic diversity index.
A8.14.6. Index Period: Not yet determined
A8.14.7. Data Collection
These data could be obtained from several sources. NASS could add questions about crop
and livestock variety to their surveys. For crop varieties, certified seed sales figures might be
collected from each state's certified seed program. The problem with using certified seed data
is that they cover University seed only; they do not account for private varieties. The severity
A8.14 - 1
-------
of this problem depends upon the crop in question. Most soy bean varieties are University
developed. Corn, on the other hand, is almost completely under the control of private
corporations which are very reluctant to release data.
A8.14.8. Estimated Cost: Not yet determined
A8.14.9. Variability: Not yet determined
A8.14.10. Interpretation
Trends in commodity productivity, soil productivity, pest density, density of beneficial
insects, wildlife populations, agri-chemical usage and non-point source loading may be analyzed
with respect to trends in genetic diversity.
For livestock, one might count the number of breeds in production for each type of
animal. Again, there is the problem of determining how different breeds really are. Another
possibility is to compare the numbers of each breed against the minimal viable population size
from population genetics principles. This could be accomplished by comparing the effective
population size of the breed to a minimal viable population size. We might show how many of
the livestock breeds are above or below a minimal viable population level. This would require
enumeration of breeding male and female for each breed.
The actual form of the indicator(s) is unclear. For crops, one could simply count the
number of varieties of each crop in production. The actual amount (acreage) of each variety in
production would provide another level of detail. The problem with this simplistic approach is
that it does not account for the similarity between varieties. For example, 90% of the varieties
of soy bean come from 6 land races. How different are they? How can that difference be
quantified? There are several approaches possible. Isozyme analysis has been suggested as the
best method to determine if two varieties are really different This is a laboratory-intensive and
expensive procedure. Another suggestion is an analysis of inbreeding coefficients for the
varieties. Some (arbitrary) value of inbreeding coefficient would be chosen to separate unique
from "identical" varieties. This requires knowledge of the pedigree for each variety. Pedigrees
are available with different levels of difficulty depending upon the crop and variety in question.
University seed pedigrees are readily available; private companies are unlikely to supply them.
A8.14 - 2
-------
A8.14.ll. Future Work
Future work for this indicator includes:
o extensive research and development of backround material
o determination of appropriate measures for genetic diversity of crops and livestock
o development of an index of genetic diversity based upon those measures
This work is unscheduled at present.
A8.14 - 3
-------
Appendix 9
Integration Across EMAP Ecosvstems
A9.1 Ecosystem Linkage-Associations
A primary area of concern is related to the Tier 2 sampling design options being
considered by the various ecosystem resource groups; the resource groups of the seven
ecosystems are independent!}' planning Tier 2 sampling designs that are appropriate for their
respective ecosystems but which will provide minimal co-location of sample sites across
ecosystems. The question has been raised as to the extent linkages across ecosystems could be
discussed if the different Resource Groups are not taking data on the same sampling units. This
question is the principal subject addressed in this section.
The primary objective of EMAP is to estimate current status, extent, changes and trends
in seven major ecosystems and to report these conditions separately by ecosystem on a regional
basis with known confidence. In order to accomplish this objective, probability samples of each
ecosystem must be designed so as to provide unbiased estimates with a specified precision and
with minimum cost This objective can be fully met with the samples for the separate
ecosystems being drawn independently and even from different sampling frames, as long as the
frames and the sampling procedures are appropriate for the respective ecosystems. In the context
of EMAP it is important to remember that we are trying to develop some understanding of the
conditions of the ecosystems on a regional basis. As long as the monitoring strategy of EMAP
includes the use of statistically valid frames and sampling protocols, the results should provide
valid answers as to the basic health of the ecosystems.
A second objective of EMAP is to seek associations between human-induced stresses and
ecological conditions that may identify possible causes of adverse effects. This second objective
requires an examination of the interrelationships among ecosystems. It is this objective that
stimulates the concern about ecosystem linkages.
Tier 1 characterization, which will provide 1990 baseline information on each ecosystem
for all 12,600 hexagons, is scheduled to be phased in gradually and completed over some time
period. Tier 2 samples will employ an interpenetrating design which will also be phased in over
(presumably) a four year period. Each ecosystem will, independently, select its own Tier 2
sample based on stratification and sampling protocols appropriate to that ecosystem. While
individual hexagons usually will contain more than one resource, the expected frequency with
which even two ecosystems will sample the same hexagon will be too low to provide reliable
information on cross- ecosystem linkages. Even under the assumption of random geographical
distribution of ecosystems and simple random sampling of hexagons by each ecosystem, any pair
of ecosystems would co-sample only approximately one- fourth of the hexagons, or about 200
hexagons per year on a national basis on a four year rotation. Natural stratification (e.g., range
and dessert lands will not often occur with agricultural lands) and the sampling stratification and
A9 1
-------
protocols unique to the individual ecosystems will greatly reduce this number, except perhaps
with streams and wetlands where the ecosystems naturally tend to co-locate. The overlap would
be much less for combinations of three or more ecosystems. Further, even in hexagons that were
co-sampled, the likelihood will be small that plots sampled by the two ecosystems will be in the
same areas of the hexagon. They may even be in different watersheds.
With this basic sampling background, what can we say in terms of ecosystem linkages?
A9.1.1 Tier 1 associations
Since Tier 1 characterizations will be made on all hexagons for all ecosystems, it is
possible to examine the properties of joint associations for the characterization variables; thus,
some correlative linkages across ecosystems are possible at this level. However, linkages
involving Tier 1 variables are of limited value in assessing cross-ecosystem health relationships.
The correlative linkages at the Tier 2 level are the critical linkages because it is at this tier that
ecosystem "health" indices are being developed. What can we say about linkages across
ecosystems at Tier 2?
A9.1.2 Sample level correlations
EMAP will not be able to obtain valid estimates of correlations and establish linkages
across ecosystems at the sampling unit level. The most meaningful estimates of correlations
among potential cause-effect variables are obtained when the variables are measured at the level
in the system at which the interactions occur. The further the measurements are removed from
the interaction process, the less informative will be the analysis of associations. In the EMAP
context this means that all variables to be correlated should be measured on the individual
sampling units, or at least within the local environmental systems. However, no sampling scheme
currently accepted will sample from the same sampling units across all ecosystems. Conceptually
it would be possible to place constraints on Tier 2 sampling designs so as to ensure adequate
estimates of all relevant inter-ecosystem correlations, but this would greatly complicate all
ecosystem sampling designs, make them inefficient for the primary objective of EMAP, and if
not carefully done introduce biases into the estimates of status for the individual ecosystems.
A9.1.3 Regional spatial patterns
Spatial representation of the Tier 2 data within geographical regions will allow one to
observe spatial patterns and how these patterns change over time. Concordance of the spatial
patterns for key indicators in different ecosystems may provide some suggestion of associations
across ecosystems that warrant further investigation. These observed associations will be much
less informative than direct correlations measured at the process level (and may be misleading)
and must be used with caution.
An example of an association developed from regional patterns might involve use of a
specific pesticide. Correlations could be shown -within the agroecosystem between levels of usage
A9 2
-------
of the pesticide and levels found in soils and pond water (also ground, if we were to sample
ground water). Patterns of high and low levels of usage of the pesticide over some region in the
agroecosystem might be accompanied by similar patterns of contamination from the pesticide in
the other ecosystems. Concordance of "hot spots" with respect to usage and "hot spots" with
respect to contamination of the pesticide would suggest a relationship to be pursued, and
particularly so if the concordance of the regional patterns between ecosystems was consistent
with the correlation observed within the agroecosystem. Comparisons of these regional patterns
are the strongest linkages we should expect to show across ecosystems using the current
suggested sampling designs at Tier 2.
A9.2 Ecosystem Linkage - Correlations/Cause and Effect
The first section of this appendix addressed the issue of what information on cross-
ecosystem linkages could be expected from the Tier 1 and Tier 2 data. For the primary objective
of estimating current status, extent, changes and trends in the nation's seven major ecosystems
on a regional basis, it was argued that valid data on status, extent, changes and trends would be
obtained for each ecosystem as long as the independent ecosystem sampling frames and sampling
procedures were appropriate for the respective ecosystems. The second EMAP objective, to seek
associations between human-induced stresses and ecological conditions that may identify possible
causes of adverse effects, requires examination of the interrelationships among indicators both
within and among ecosystems. In section A9.1 it was argued that, while concordance of spatial
patterns for indicators from different ecosystems may provide suggestions of associations to be
further studied, valid estimates of sample-level correlations across ecosystems could not be
obtained from the Tier 2 data. This requirement to measure cross-ecosystem correlations has
been the subject of considerable discussion and debate among the members of the various
ecosystem teams. This section addresses some of these linkage issues and concerns and in
particular suggests procedures for obtaining valid cross-ecosystems correlations.
In order to obtain valid estimates of correlations between potential cause-and-effect
variables, it must be possible to measure both variables on the same sampling unit. Tier 2 will
measure indicators which will have some potential as cause-and-effect variables. However, it is
doubtful that enough sample points from different ecosystems will be co-located, through random
selection, to provide sufficient sample sizes for valid estimates. Even when ecosystem sample
points are co-located in the same hexagons, the sampling points will usually not be co-located
at the sample unit level.
Since neither Tier 1 nor Tier 2 samples, for different reasons, provide a reasonable
framework for estimates of statistical correlations between ecosystems, it is our contention that
the linkage problem is best approached in the following way. Each Ecosystem Resource Group
must first identify the linkages with other ecosystems that are key tc understanding the "health"
of their system. Once these linkages are identified, approaches are suggested for studying the
linkages. The approach will depend on the degrees of interaction needed between the
ecosystems.
A9 3
-------
The three approaches, briefly developed below, would provide different levels of viewing
linkages across ecosystems. These could be further developed, if EMAP management is
interested.
A9.2.1 Limited Information
If one ecosystem requires limited, but important, information (indicator data) from
another ecosystem, the first ecosystem could adopt the indicator used by the other
as one of their own and obtain the data on that indicator from the appropriate area
within or adjacent to their sample sites. For example, pesticide applications on
surrounding lands in the watershed, whether they be forests, range, or agricultural
lands, might be defined as an important indicator for the aquatic systems.
Similarly, agroecosytem is intending to define an indicator of wildlife activity in
the areas adjacent to agricultural fields. It is important that cross-ecosystem
indicators be defined and measured the same way in all ecosystems.
A9.2.2 Regional Problem
If a regional problem were identified that appeared to involve several ecosystems,
a Tier 2 intensive sampling protocol could be designed to sample areas where the
ecosystems are co-located. The number of sampling units would have to be
determined for the area of interest and the sampling protocol coordinated between
the ecosystems. Statistical correlations could be determined and discussion of
possible cause-effect relationships could be developed. However, cause-effect
relationships could not be proven. We have defined this as a Tier 3 program.
A9.2.3 Potential Cause-Effect Relationships
For those linkage studies that require a higher level of data collection and
coordination between ecosystems, we recommend initiation of a selected suite of
tier four research studies aimed at understanding specific linkage systems. The
study design in these cases would select sample (or research) sites and collect data
across ecosystems as necessary to understand the system. For example, it is
anticipated that such a specialized linkage study would be necessary to understand
the stream loading (pesticide, fertilizer, sediment) process and its effect on fish
populations. This approach emphasizes the process to be studied and not the
individual ecosystems, and would appear to provide the greatest promise for
understanding the more complicated linkage problems.
A9 4
------- |