&EPA
United States
Environmental Protection
Agency
Remote Sensing and Geographic
Information Systems for Decision
Analysis in Public Resource
Administration: A Case Study of
25 Years of Landscape Change in
a Southwestern Watershed
054LEB02.RPT • 6/4/02
-------
EPA/600/R-02/039
June 2002
Remote Sensing and Geographic
Information Systems for Decision Analysis
in Public Resource Administration:
A Case Study of 25 Years of Landscape
Change in a Southwestern Watershed
William G. Kepnerand C.M. Edmonds
U.S. Environmental Protection Agency
Office of Research and Development
National Exposure Research Laboratory
P.O. Box 93478
Las Vegas, NV89193
Christopher J. Watts
Institute del Medio Ambiente y el Desarrollo Sustentable del Estado de Sonora
Reyes and Aquacalientes Esq., Col. San Benito
Hermosillo, Sonora, Mexico 83190
-------
"Siempre beo y es ansi que por la mayor parte quando tenemos entre las manos alguna cosa preciosa y
la tratamos sin impedimento no la tenemos ni la preciamos en quanta vale ni entendemos la falta que nos
haria si la perdiesemos y por tanto de continue la bamos teniendo en menos pero despues que la abemos
perdido y carecemos del beneficio de ella abemos gran dolor en el coracon y siempre andamos y
maginatibos buscando modos y maneras como latornemos a cobrar...."
"I have always noticed, and it is a fact, that often when we have something valuable in our possession
and handle it freely, we do not esteem or appreciate it in all its worth, as we would if we could realize how
much we would miss it if we were to lose it. Thus we gradually belittle its value, but once we have lost it
and we miss its benefits, we feel it in our heart and are forever wanting, thinking of way and means to
retrieve it...."
- Pedro de Castaneda
History of the Expedition
October 1596
(A chronicle of Francisco Vasquez de Coronado's expedition in search of the Seven Cities of Cibola in
1540. It is believed that Coronado's party followed the San Pedro north from modern-day Sonora into what
is now southeastern Arizona.)
-------
Acknowledgments
I gratefully acknowledge Dr. William N. Thompson, University of Nevada Las Vegas (UNLV),
College of Urban Affairs; Dr. Karen S. Layne UNLV College of Urban Affairs; and Dr. David K.
Kreamer, UNLV College of Sciences (Water Resources and Management Program) for their helpful
suggestions as reviewers for this report.
Notice
This report has been peer reviewed by the U.S. Environmental Protection Agency (EPA), through its
Office of Research and Development (ORD) and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation by EPA for use.
-------
Table of Contents
Acknowledgments iii
Notice iii
List of Tables and Figures v
Abstract vi
Section 1 - Introduction 1
Section 2 - Review of Related Literature 3
2.1 Landscape as an Integration Concept 3
2.2 Technology and Theory Integration as a Concept for Measuring the Environment 3
2.3 Organizational Framework for Decision Analysis 4
Section 3 - Materials and Methods 7
3.1 Study Site 7
3.2 Image Acquisition and Characterization 7
3.3 Change Detection Analysis 11
Results 12
Conclusions 17
Literature Cited 19
IV
-------
List of
Tables and Figures
Tables
1 Land cover class descriptions for the Upper San Pedro Watershed 10
2 Landscape change statistics 11
3 Proportional land cover extent as total hectares and percent for the Upper San Pedro
Watershed (1973, 1986, 1992, and 1997) 12
4 Percent relative land cover change for the Upper San Pedro Watershed (1973-1986,
1986-1992, 1992-1997, and 1973-1997) 12
5 Landscape change statistics for four land cover classes in the Upper San Pedro
Watershed (1973-1997) 15
Figures
1 Organizational framework for scenario analysis (Steinitz 1996, 1990) 4
2 Landscape change from perennial grassland to mesquite woodland in a semi-arid rangeland, (Santa
Rita Mountains, Arizona). Top photo (1903); bottom photo (1941) 6
3 Location of the Upper San Pedro River Basin, Arizona/Sonora 8
4 Land cover in the Upper San Pedro Watershed (U.S./Mexico). Source: Landsat MSS (5 June
1973, 10 June 1986, 2 June 1992) and Landsat TM (8 June 1997) 9
5 Urban land cover change for the Upper San Pedro Watershed (1973-1986, 1986-1992, and
1992-1997) 13
6 Mesquite land cover change for the Upper San Pedro Watershed (1973-1986, 1986-1992, and
1992-1997) 14
7 Grassland land cover change for the Upper San Pedro Watershed (1973-1986, 1986-1992, and
1992-1997) 14
8 Desertscrub land cover change for the Upper San Pedro Watershed (1973-1986, 1986-1992,
and 1992-1997) 15
9 Conceptual model of vegetation phase transitions in a semi-arid watershed 17
-------
Abstract
Alternative futures analysis is a scenario-based approach to regional land planning that attempts to
synthesize existing scientific information in a format useful to community decision-makers. Typically, this
approach attempts to investigate the impacts of several alternative sets of choices preferred by
representative stakeholder groups relative to selected environmental or economical endpoints. Potential
impacts from each of the scenarios are compared to current conditions of the region in terms of a set of
processes that are modeled within a geographic information system. Future conditions are generally
examined from the perspective of a recent baseline condition (versus empirically determined using a series
of retrospective measurements).
During the past two decades, important advances in the integration of remote imagery, computer
processing, and spatial analysis technologies have been linked to the study of distribution patterns of
communities and ecosystems and the ecological processes that affect these patterns. Because of the 25+
year availability of commercial satellite imagery, it is possible to examine environmental change and
establish models which can narrow the actual choice of possible and probable change scenarios.
This research A) examines the potential to establish reference condition and measure change over large
geographic areas; B) determine trends in environmental condition; and C) model and predict future
landscape scenarios using advanced space-based technologies. Specifically, landscape pattern
measurements were developed from satellite remote sensing, spatial statistics, and geographic information
systems technology for a semi-arid watershed in southeast Arizona and northeast Sonora, Mexico and
evaluated for their use in a decision-making framework.
VI
-------
Section 1
Introduction
The assessment of land use and land cover is an extremely important activity for contemporary land
management. A large body of recent literature (Houghton et al. 1983, Turner 1990, McDonnell and Pickett
1993) suggests that human land-use practices (including type, magnitude, and distribution) are the most
important factor influencing natural resource management at local, regional, and global scales.
Traditionally, western U.S. land management has been pursued within small localized areas, such as
grazing allotments, or within political jurisdictional boundaries, such as National Park Service units and
National Forest systems. Through much of the past century, forests and rangelands have been managed to
assure production of timber, livestock, water, minerals, and recreational opportunities, with the primary
focus on outputs rather than on the environmental condition left behind.
Today's environmental managers, urban planners, and decision-makers are increasingly expected to
examine environmental and economic problems in a larger geographic context to 1) understand the scales at
which specific management actions are needed; 2) conceptualize environmental management strategies; 3)
formulate sets of alternatives to reduce environmental and economic vulnerability and uncertainty in their
evaluation analyses; and 4) to prioritize, conserve, or restore valued natural resources, especially those
which provide important economic goods and services.
To manage natural resources effectively, managers and decision-makers need a means to
1) characterize the environment at different hierarchical spatial and temporal scales; 2) identify patterns
and processes important at different scales; and 3) compare these patterns and processes to a set of
reference conditions (Kaufmann et al. 1994).
A scenario-based approach to regional land planning offers an organizational basis to explore decision
analysis and opportunities for public resources. Scenario planning was initially used by the military after
the Second World War and since has been tested in a variety of geographical settings to assist stakeholders
and policy makers in shaping future use of land and water resources (Schwartz 1996, Steinitz 1990).
Scenario analysis offers several advantages over other assessment frameworks including the ability to
intentionally investigate several "futures," i.e. different points of view, at one time. The most important
reasons for employing scenario analysis relate primarily to the potential benefits of evaluating all aspects of
the local decision-making processes. For example, for land owners interested in protecting their property
rights, scenario analysis can be used to understand the range of potential impacts to their lands that may be
caused by regional change relative to the type, location, and magnitude of proposed management actions or
policy.
-------
Additionally, for elected officials and public administrators, scenarios can be used to test current
planning ideas in terms of public perceptions or presumed changes in human demography. Thus scenarios
can be used to test the resilience of plans against assumptions about the stability and growth into the future.
Lastly, the use of scenarios also allow members of an entire community to assess the relative impacts
of several alternative sets of choices for a desirable future environment. Scenario analysis thus requires
that scenarios must be possible, credible, and relevant to be useful in decision-making processes.
The purpose of this research is to develop representative (reference) and change models which can aid
in the administration of public natural resources by assessing spatial and temporal changes in land use and
land cover at a watershed scale. Subsequently, it is anticipated that through the use of satellite remote
sensing and geographic information systems technology that it will be possible to characterize resource
stability relative to cumulative environmental stress and model and predict future outcomes based on multi-
year trend information.
It is the hypothesis of this project that landscape composition and pattern measures are diagnostic of
environmental condition and can be measured using space-based technologies for decision-making
processes in public natural resource management. Secondly, it is believed that a set of landscape
characteristics measured over time can be established for reporting status and detecting trends in resource
vulnerability to human-induced and natural disturbance. Vulnerability for the purpose of this study
location has been defined as any serious risk to maintaining a desired state in which community diversity,
productivity, and resistance to disturbance are sustained (CEC 1999).
The following sections include a review of pertinent literature relative to 1) performing large-scale
environmental assessment and incorporating science into a decision-making process; 2) methodology and
materials utilized to remotely measure the environment and analyze very large spatial data sets; 3)
demonstration of the combination of technologies to assess changes in a selected location in the semi-arid
Southwest, and 4) application of results within a decision-making framework to solve complex problems
related to the environment and the people who depend on it.
-------
Section 2
Review of Related Literature
The combination of landscape ecology, advanced technology, and decision analysis provide a unique
basis for measuring and interpreting large-scale environmental change. The approach discussed and tested
within this report is largely dependent on the integration of natural and social science to interpret landscape
pattern metrics relative to specific endpoints such as regional or watershed vulnerability.
2.1 Landscape as an Integration Concept
Landscapes are conceptual units for the study of spatial patterns on the physical environment and the
influence of these patterns on important environmental endpoints. Hence, landscapes provide the spatial
context for ecosystem dynamics and integrity (O'Neill 1999). Landscape composition and pattern affect
key ecological transfer processes which govern the movement or flow of energy, nutrients, water, and biota
over time and operate at many scales (Forman and Godron 1986). Hierarchy theory provides the context
for integrating multiple scales of information related to the operation of ecological processes (O'Neill et al.
1986). In simple terms, it states that landscapes are organized into patterns within a hierarchy of spatial
and temporal scales. Natural and human-induced disturbances occur across a range of spatial and
temporal scales and serve to either maintain landscape patterns or initiate phase transitions into new
patterns. A landscape framework provides the context 1) to investigate changes in composition, pattern
distribution, and process function; 2) to compare conditions across mixed landscapes; and 3) to assess
cumulative sources of environmental perturbation (Jensen and Everett 1994).
Land use decisions are generally made at an individual landowner or local scale level, however, the
impacts are often manifested cumulatively as change in spatial pattern on the landscape (O'Neill 1988).
For example, changes in spatial pattern and composition have been implicated in the decline of biological
diversity, ecosystem sustainability, and the ability to recover from disturbance at a number of scales
(Flather et al. 1992, O'Neill et al. 1988). This is important because individual land use can result in an
additive response condition which impacts ecological processes on a broader scale. In terms of policy,
although individual actions occur on a local scale, they are often administratively governed at the greater
landscape level of organization, i.e. natural resources are managed by watershed, forest service regions, or
within political units such as states and counties.
2.2 Technology and Theory Integration as a Concept for Measuring the Environment
During the past decade, important advances in the integration of remote imagery, computer processing,
and spatial analysis technologies have been linked to the study of distribution patterns of communities and
ecosystems, ecological processes that affect these patterns, and changes in pattern and process over time.
O'Neill et al. 1997 argue that a landscape approach is practical within current technologies for monitoring
environmental quality over large regions and it may represent a less expensive approach than using
traditional fine-scaled ground-based surveys. Although not all environmental perturbations can be
-------
explained or measured via alterations of land cover, this approach at least supplements existing
technologies and improves our ability to measure and understand change and trend over time.
Earth observing satellite imagery is globally available via the Advanced Very High Resolution
Radiometer (AVHRR). AVHRR imagery (1.1 km2 pixel resolution) has been used to estimate current
vegetation for the United States (Loveland et al. 1991). Improved spectral and spatial resolution imagery is
commonly available from commercial and government vendors. It is now clearly possible to map natural
resource features at the 60-meter (e.g. Landsat Multi-spectral scanner), 30-meter (Landsat Thematic
Mapper) and 10-meter (SPOT) scales of pixel resolution.
2.3 Organizational Framework for Decision Analysis
Landscape architecture involves several areas of theory all of which influence design. Much of the
contemporary thinking in regard to landscape design analysis has been outlined in various studies
performed by the Harvard University Graduate School of Design (Steinitz 1996, 1993, 1990) in which
potential impacts from a number of wide-ranging scenarios are compared to current conditions of a region
in terms of a set of processes that are modeled in a geographic information system (GIS). Alternative
future landscape analysis involves describing the patterns and significant human and natural processes
affecting a geographic area of concern, constructing GIS models to simulate these processes and patterns,
creating changes in the landscape by forecasting and by design, and evaluating how the changes affect
pattern and process using models. The organizational framework for the analysis identifies six types of
question or levels of inquiry (Steinitz 1990).
The six levels of inquiry (and the associated models) are listed below in the order in which they are
usually applied (Figure 1):
1. How should the state of the landscape be
described in terms of content, boundaries,
space, and time? (Representation Models)
2. How does the landscape operate? What are
the structural and functional relationships
among its elements? (Process Models)
3. How does one judge whether the current
state of the landscape is working well?
(Evaluation Models)
4. By what actions might the current
representation of the landscape be altered,
e.g. by conservation or development?
(Change Models)
4b. How might the landscape be changed
by current projected trends?
(Projection Models)
4c. How might the landscape be changed
by designed action? (Intervention
Models)
Recogn
Conte
How should the
landscape be described?
How does the landscape
operate?
Is the current landscape
working well?
How might the landscape
be altered - by what actions,
where, and when?
What predictable differences
might the changes cause?
How should the landscape
be changed?
ize Perform
xt Study
"> f>
Representation
Models
Process
Models
Evaluation
Models
Change
Models
Impact
Models
Decision
Models
U
Specify
Method
^
Implement
Decision
Figure 1. Organizational framework for scenario
analysis (Steinitz 1996, 1990).
-------
5. What predictable differences might the changes cause? (Impact Models)
6. How is a decision to change (or conserve) the landscape to be made? How is comparative
evaluation to be made among alternative courses of action? (Decision Models)
In practice, the organizational framework works in reverse, i.e. to be able to decide whether to propose
or make a change one needs to know how to compare and evaluate the alternatives. To be able to evaluate
the alternatives one needs to predict the comparative impacts from simulated changes. To be able to
simulate change, one needs to know what changes to simulate. To be able to consider changes, one needs
to evaluate how well the current situation is performing. To be able to evaluate the current situation, one
needs to understand how it works. Lastly, to understand how it works, one needs representational
information to describe the current state.
The steps outlined above include components which determine the reference (or historic) conditions of
the analysis area. Historic reference conditions are useful in managing the environment by telling which
processes or functional parts need to be preserved. If only current conditions are the criteria used to make
management decisions, there is no basis to determine whether management practices or impacts will lead to
environmental outcomes that fall within the historic range of variability (Covington and Moore 1992).
Preferably it is desirable to directly evaluate undisturbed environments to determine reference
condition. However, in reality most natural environments have been impacted and modified by both
modern and aboriginal humans (Swanson et al. 1993). Secondly, it is much easier to evaluate spatial scales
than temporal scales because we can directly observe the present, however, evaluating changes through
time is fundamental to predicting potential future conditions.
Historically, it has not been possible to of compare conditions across large landscapes or assess
cumulative sources of environmental perturbation. Ideally, historical documents and inventories should
provide a significant portion of information for understanding reference condition, however, historical
references or reconstructions are generally quite limited (Maser 1990).
As an example, vegetation change in the American West has been a subject of concern throughout the
twentieth century (Humphrey 1958, Branson 1985, Grover and Musick 1990, Bahre 1991, and Bahre and
Shelton 1993). The information for vegetation change has largely been derived from archival literature and
photography. Most of the evidence for vegetation change is actually provided from a series of matched
photographs beginning in the late 1800s and early 1900s (Figure 2). However, there are serious drawbacks
in using this technique to assign change over this period of history.
As some authors (Bahre 1991) point out, the field of view in ground photographs is usually oblique and
covers little total area which limits their usefulness in determining change in plant occurrence over large
regional areas. Secondly, the historic photographic series are usually separated by large periods of time
and they are often captured more than a decade after the sites were first disturbed by human activity.
Lastly, the change photography has largely been used for qualitative comparisons and little progress has
been made in quantifying and characterizing vegetation change, especially in regard to determining which
systems are most resilient or vulnerable. Although several studies have addressed specific aspects of
vegetation change in the Southwest, few have attempted to synthesize the cumulative impacts over large
regional or watershed areas.
-------
Important advances in the integration of remote imagery, computer processing, and spatial analysis
technologies have been coupled to landscape ecology theory to study the distribution patterns of
communities and ecosystems, human and environmental processes that affect these patterns, and changes in
pattern and process over time. The work provided from this research is intended to contribute to our ability
to characterize large assessment areas (representative model) and provide predictive inference (change
model) for alternative future scenarios which can lead to a comparative analysis of impacts relative to
alternative courses of management action (decision model).
'"'''-.ffCiJSIiiSt^
rysj£3§iS
Figure 2. Landscape change from perennial grassland to
mesquite woodland in a semi-arid rangeland, (Santa
Rita Mountains, Arizona). Top photo (1903); bottom
photo (1941).
-------
Section 3
Materials and Methods
The application of several advanced technologies to assess spatial and temporal changes was tested in a
moderately sized Southwestern watershed described below. The source data were drawn off a series of
Landsat satellite platforms beginning in 1973. The case study area was selected for a variety of reasons
including data richness, circumstantial information related to change, and stakeholder involvement. It
represents one of the first attempts to examine large scale change over a quarter century of time using large
datasets acquired from remote earth-orbiting sensors.
3.1 Study Site
The study location is the upper San Pedro River basin which originates in Sonora, Mexico and flows
north into southeastern Arizona (Figure 3). The San Pedro River is an international basin with
significantly different cross border legal and land use practices (CEC 1998, USBLM 1998, Tellman et al.
1997). The watershed embodies a variety of characteristics which make it an exceptional outdoor
laboratory for addressing a large number of scientific questions in arid and semi-arid hydrology, ecology,
meteorology, and the social and policy sciences. The Upper San Pedro Watershed represents a transition
area between the Sonoran and Chihuahuan deserts and topography, climate, and vegetation vary
substantially across the watershed. Elevation ranges from 900 - 2,900 m and annual rainfall ranges from
300 to 750 mm. Biome types include desertscrub, grasslands, oak woodland-savannah, mesquite
woodland, riparian forest, coniferous forest, and agriculture. The upper watershed encompasses an area of
approximately 7,600 km2 (5,800 km2 in Arizona and 1,800 km2 in Sonora, Mexico).
3.2 Image Acquisition and Characterization
Remote imagery was derived from the Landsat Multi-spectral Scanner (MSS) and Landsat Thematic
Mapper (TM) earth observing satellites (path/row 35/38 and 35/39). Landsat-MSS satellite scenes were
selected from the North American Landscape Characterization (NALC) project (USEPA 1993). The
scenes available in the NALC database (1973-92) and Landsat TM (1997) are from four pre-monsoon
dates for a period of approximately 25 years (i.e. 5 June 1973, 10 June 1986, 2 June 1992, 8 June 1997).
All imagery in the database is coregistered and georeferenced to a 60 x 60 meter Universal Transverse
Mercator (UTM) ground coordinate grid with a nominal geometric precision of 1-1.5 pixels (60-90 m).
Digital land cover maps were developed separately for each year using 10 classes: Forest, Oak Woodland,
Mesquite Woodland, Grassland, Desertscrub, Riparian, Agriculture, Urban, Water, and Barren (Figure 4).
The cover classes are briefly described in Table 1. A decision similar to other studies (Klemas et al. 1993)
was made to classify the images separately prior to change detection analysis because of the difficulty in
normalizing images derived from different satellite sensors. The landscape changes were analyzed in a
geographical information system using ARC/INFO software.
-------
10 0 10 20 30 Kilometers
Figure 3. Location of the Upper San Pedro River Basin, Arizona/Sonora.
-------
S Forest
. - : Oak Woodland
Mesquite Woodland
Grassland
Desertscrub
II Riparian
P Agriculture [_
• Urban
Water
Barren
Clouds ('92 and '97 only)
1973
1986
1992
1997
Figure 4. Land cover in the Upper San Pedro Watershed (U.S./Mexico). Source: Landsat MSS
(5 June 1973, 10 June 1986, 2 June 1992) and Landsat TM (8 June 1997).
The first step in the image classification was using ERDAS IMAGINE 8.3 software procedure
ISODATA to perform an unsupervised classification using bands 1 (green), 2 (red) and 4 (near infrared) to
produce a map with 60 spectrally distinct classes. The choice of 60 classes was based on previous
experience with NALC data and usually gave satisfactory trade-off between the total number of classes and
the number of mixed classes. In this context, it proved helpful to define a larger set of 21 intermediate
classes, which were easier to relate to the spectral information. For example, the Barren class contains
bare rock, chalk deposits, mines, tailing ponds, etc. which have very different spectral signatures. Each
class was then displayed over the false-color image and classes were assigned into one of the 21 land cover
categories or as mixed. The software allows the interactive manipulation of the signatures for each class
which allowed many of the mixed classes to be resolved.
The remaining mixed classes were separated into different categories using a variety of ancillary
information sources, such as the topographic maps (scale 1:50,000) produced by INEGI, the Mexican
National Institute of Statistics, Geography and Information, and by the U.S. Geological Survey (scale
1:24,000). The land use information used varied depending on the image being analyzed. Thus the
classification of the 1997 image relied heavily on field visits to establish ground control. Five 3-day site
visits were carried out from September 1997 to June 1998 to collect specific land cover data with the aid of
Global Positioning System equipment which were incorporated into successive iterations of the
classification process.
-------
Table 1. Land cover class descriptions for the Upper San Pedro Watershed.
Forest
Vegetative communities comprised principally of trees potentially over 10 m
in height and frequently characterized by closed or multi-layered canopies.
Species in this category are evergreen (with the exception of aspen), largely
coniferous (e.g. ponderosa pine), and restricted to the upper elevations of
mountains that arise off the desert floor.
Oak Woodland
Vegetative communities dominated by evergreen trees (Quercus spp.) with a
mean height usually between 6 and 15 m. Tree canopy is usually open or
interrupted and singularly layered. This cover type often grades into forests
at its upper boundary and into semi-arid grassland below.
Mesquite Woodland
Vegetative communities dominated by leguminous trees whose crowns cover
15% or more of the ground often resulting in dense thickets. Historically
maintained maximum development on alluvium of old dissected flood plains;
now present without proximity to major watercourses. Winter deciduous and
generally found at elevations below 1,200 m.
Grassland
Vegetative communities dominated by perennial and annual grasses with
occasional herbaceous species present. Generally grass height is under 1 m
and they occur at elevations between 1,100 and 1,700 m; sometimes as high
as 1,900 m. This is a landscape largely dominated by perennial bunch
grasses separated by intervening bare ground or low-growing sod grasses and
annual grasses with a I ess-interrupted canopy. Semi-arid grasslands are
mostly positioned in elevation between evergreen woodland above and
desertscrub below.
Desertscrub
Vegetative communities comprised of short shrubs with sparse foliage and
small cacti that occur between 700 and 1,500 m in elevation. Within the San
Pedro river basin this community is often dominated by one of at least three
species, i.e. creosotebush, tarbush, and whitethorn acacia. Individual plants
are often separated by significant areas of barren ground devoid of perennial
vegetation. Many desertscrub species are drought-deciduous.
Riparian
Vegetative communities adjacent to perennial and intermittent stream
reaches. Trees can potentially exceed an overstory height of 10 m and are
frequently characterized by closed or multi-layered canopies depending on
regeneration. Species within the San Pedro basin are largely dominated by
two species, i.e. cottonwood and Goodding willow. Riparian species are
largely winter deciduous.
Agriculture
Crops actively cultivated (and irrigated). In the San Pedro River basin these
are primarily found along the upper terraces of the riparian corridor and are
dominated by hay and alfalfa. They are minimally represented in overall
extent (less than 2%) within the basin and are irrigated by ground and
pivot-sprinkler systems.
Urban
(Low and High
Density)
This is a land cover dominated by small ejidos (farming villages or
communes), retirement homes, or residential neighborhoods (Sierra Vista).
Heavy industry is represented by a single open-pit copper mining district near
the headwaters of the San Pedro River near Cananea, Sonora (Mexico).
Water
Sparse free-standing water is available in the watershed. This category would
be mostly represented by perennial reaches of the San Pedro and
Babocomari rivers with some attached pools or represses (earthen
reservoirs), tailings ponds near Cananea, ponds near recreational sites such
as parks and golf courses, and sewage treatment ponds east of the city of
Sierra Vista, Arizona.
Barren
A cover class represented by large rock outcropping or active and abandoned
mines (including tailings) that are largely absent of above-ground vegetation.
10
-------
3.3 Change Detection Analysis
Mouat et al. (1993) review remote sensing techniques for detecting change by analyzing multi-date
imagery. The San Pedro digital land cover maps were transferred into UTM map projection coordinates
and incorporated into a geographical information system for change analysis. Change was analyzed using
landscape statistical software to produce landscape statistics, including actual total extent. Image
enhancement in ARC/INFO allows mathematical treatment of the composite images and to display change,
either as gain, loss, or no change. This technique has been very useful in identifying semi-arid areas which
have undergone change relative to human-induced and natural environmental stress (Pillon et al. 1988) and
was employed for this research.
Landsat-MSS 1973 was used for the baseline condition. Change between time intervals, i.e. 1973,
1986, 1992, and 1997 was measured and the discrete landscape metrics were described (Table 2).
Landscape statistics that describe shape and size were used to assess dominance, fragmentation, and rates
of conversion in an effort to determine sensitive measures for resistance to change (= landscape resilience).
Sample size was 2,100,407 pixels (60-m resolution) per digital image map.
Table 2. Landscape change statistics.
Statistic
Dominance
Connectivity
Total # of patches
Largest patch size
Avg. Patch size
Description
Area-based metric which indicates the extent to which the landscape is
dominated by a single land cover type.
Percentage of edges that are of the same land cover class. Higher value
indicates lower patchiness. Only calculated for individual land cover classes.
Number of polygons of a single land cover type.
The size of the largest contiguous polygon of a single land cover type.
Average patch size. Overall average is not area weighted.
11
-------
Results
Results for land cover extent (total hectares and percent by class) by sample year and relative change
for each interval period are presented in Tables 3 and 4. Results vary over the 25-year period, however,
certain land cover types, i.e. forest and oak woodland have changed little over this period relative to other
classes.
Table 3. Proportional land cover extent
Watershed (1973, 1986, 1992,
as total hectares and percent for the Upper San Pedro
and 1997).
Forest
Oak Woodland
Mesquite
Grassland
Desertscrub
Riparian
Agriculture
Urban
Water
Barren
Clouds
1973
Hectares
7446
93612
20821
312850
296330
8665
8775
3205
264
4177
0
%
0.98
12.38
2.75
41.37
39.19
1.15
1.16
0.42
0.03
0.55
0.00
1986
Hectares
7437
93464
106968
267321
243502
8852
11507
10002
294
6799
0
%
0.98
12.36
14.15
35.35
32.20
1.17
1.52
1.32
0.04
0.90
0.00
1992
Hectares
7045
88894
105192
265231
235480
8889
14859
12574
337
6792
10850
%
0.93
11.76
13.91
35.08
31.14
1.18
1.97
1.66
0.04
0.90
1.44
1997
Hectares
7071
90270
101602
263432
229953
9218
14530
16494
415
6769
16388
%
0.94
11.94
13.44
34.84
30.41
1.22
1.92
2.18
0.05
0.90
2.17
Table 4. Percent relative land cover change for the Upper San Pedro
Watershed (1973-1986, 1986-1992, 1992-1997, and 1973-1997).
Forest
Oak Woodland
Mesquite
Grassland
Desertscrub
Riparian
Agriculture
Urban
Water
Barren
1973-1986
-0.12
-0.16
413.75
-14.55
-17.83
2.16
31.13
212.07
11.36
62.77
1986-1992
-5.27
-4.89
-1.66
-0.78
-3.29
0.42
29.13
25.71
14.63
-0.10
1992-1997
0.37
1.55
-3.41
-0.68
-2.35
3.70
-2.21
31.18
23.15
-0.34
1973-1997
-5.04
-3.57
387.98
-15.80
-22.40
6.38
65.58
414.63
57.20
62.05
12
-------
Five of the ten land cover types represent rare (<2% total extent) classes in the study area. Although
urban land cover represents close to 2 per cent of the land cover, growth of this cover type, particularly in
Arizona, has been rapid and has increased from 3,205 total ha in 1973 to 16,494 ha in 1997; a relative
increase of 415 percent for this period (Table 4). The major surge in urbanization occurred within the first
13-year period from 1973-1986 when urban cover increased three times from the 1973 baseline (Figure 5).
Mesquite woodland, a native tree life-form, has encroached upon the entire watershed. Mesquite total
extent increased five-fold between 1973 and 1986 from 20,821 to 106,968 ha (Table 3, Figure 6). The
baseline extent of mesquite for the watershed in 1973 was 2.75 percent and by 1997 it represented 13.44
percent of the total land cover.
Major decreasing cover types included desertscrub and grassland. Although grassland dominates the
San Pedro landscape for each of the four sample periods, its total extent has steadily declined. Almost
50,000 ha of perennial and annual grasses were lost between 1973 and 1997. The major decrease for this
cover type occurred between 1973 and 1986 (45,529 ha lost) whereas 2,090 ha and 1,799 ha were lost the
following periods between 1986-1992 and 1992-1997, respectively (Figure 7).
Desertscrub had an identical trend as grasslands. Desertscrub (Sonoran and Chihuahuan species)
represents the second most dominant land cover type within the study area. Over 66,000 ha of desertscrub
were lost over the 25-year period. Similar to grasslands, most of this loss (80 percent) occurred during the
first 13 years between 1973 and 1986 (Figure 8).
Loss
No Change
Gain
1992-1 §§7
Figure 5. Urban land cover change for the Upper San Pedro Watershed (1973-1986, 1986-1992, and
1992-1997).
13
-------
Loss
No Change
Gain
1973-1986
1986-1992
1992-1997
Figure 6. Mesquite land cover change for the Upper San Pedro Watershed (1973-1986, 1986-1992, and
1992-1997).
Loss
No Change
Gain
1973-1986
I ..—.—
1986-1992
1992-1997
Figure 7. Grassland land cover change for the Upper San Pedro Watershed (1973-1986, 1986-1992,
and 1992-1997).
14
-------
1973-1986
1986-1992
1992-1997
Loss
No Change
Gain
Figure 8. Desertscrub land cover change for the Upper San Pedro Watershed (1973-1986, 1986-1992,
and 1992-1997).
Landscape statistics that describe shape and size were used to assess dominance, fragmentation, and
conversion matrices for selected cover types and are presented in Table 5.
Table 5. Landscape change statistics for four land cover classes in the Upper San Pedro Watershed
(1973-1997).
Area (ha)
% Cover
# of Patches
Largest Patch
(ha)
Ave Patch Size
Connectivity
Grassland
1973
312,850
41.37
50,715
126,258
6.18
0.62
1997
263,432
34.84
58,142
53,173
4.54
0.56
% Rel.
Change
-15.80
-15.80
+14.64
-57.89
-26.54
-9.68
Desertscrub
1973
296,330
39.19
26,260
201,165
11.3
0.66
1997
229,953
30.41
39,991
37,361
5.76
0.55
% Rel.
Change
-22.40
-22.40
+52.29
-81 .43
-49.03
-16.67
Mesquite Woodland
1973
20,821
2.75
15,558
461 .52
1.34
0.31
1997
101,602
13.44
53,310
3,574
1.91
0.37
% Rel.
Change
+387.98
+387.98
+242.65
+674.34
+42.54
+19.35
Urban
1973
3,205
0.42
418
982
7.86
0.74
1997
16,494
2.18
3,010
4,938
5.55
0.69
% Rel.
Change
+414.63
+414.63
+620.10
+402.82
-29.39
-6.76
15
-------
Mesquite woodland has experienced the most rapid increase in extent during the study period. More
than 80,000 ha of mesquite were gained since the 1973 baseline and it has undergone expansion by
aggregation to form clusters which later coalesced into large woodland patches. The number of mesquite
polygons (patches) and average patch size have increased steadily throughout the study area (Table 5).
Mesquite patches have increased up to 3,574 ha in size and increasingly become more connected, i.e. the
percentage of edges are of identical land cover class, resulting in large stands with closed canopies.
Urban cover has also increased during the study period. Similar to mesquite, urban cover has
increased in the number of patches and largest patch size from 418 and 982 ha to 3010 and 4,938 ha,
respectively. However, average urban patch size and connectivity have actually decreased, likely due to
urbanization of the outlying suburban areas.
The majority of mesquite and urban gain during the 25-year study period were predominantly derived
from desertscrub and grassland cover classes. Subsequently, desertscrub and grassland show a general
trend in fragmentation and actual loss. Total extent for these two cover classes decreases through time and
the number of patches increases. Additionally, the average patch size for desertscrub and grassland
decreases from 11.3 to 5.76 ha and 6.18 to 4.54 ha, respectively and connectivity decreases from the 1973
baseline (Table 5).
16
-------
Conclusions
The methods developed as an outcome of this study have been employed for their capability to assess
the spatial and temporal changes in land use and land cover at a landscape scale and to subsequently
determine an effective means to measure landscape stability over large assessment areas such as
watersheds. The ability to interpret condition and change over large areas has only become feasible with
the availability of remotely sensed data such as Landsat. The advantages of this new approach make it
possible to 1) observe large geographic areas and multi-jurisdictions in their entirety; 2) quantify landscape
pattern and the areal extent of resources; 3) observe changes and trend in large-scale patterns through time;
and to 4) assess cumulative sources of environmental perturbation (Graham et al. 1991, Urban et al. 1987).
Specifically, remote sensing integrated into a
GIS environment provides an ability to characterize
large assessment areas and establish reference
condition. The use of landscape metrics based on
land cover generated from remote sensors provides a
unique opportunity to assess areas of large regional
scale. In terms of the alternative landscape analysis
it fulfills the need to describe the landscape in terms
of content, boundaries, space, and time and thus
provides the representative model for the initial step
of scenario analysis.
Secondly, the results of this research will benefit
decision-makers and natural resource managers who
are principally interested in evaluating present and
past cumulative impacts to a watershed or
formulating alternative management strategies to
sustain environmental health and economical
viability into the future. The pattern measurements
from this research provide predictive inference (a
change model) for measuring and evaluating change.
Thus it serves to answer questions related to how
might the landscape be changed by current projected
trends (Figure 9).
Lastly, the combination of remote sensing, GIS,
and landscape pattern metrics help contribute to the
comparative evaluation to be made among
alternative courses of management and policy action
(i.e. alternative future scenarios) which ultimately
lead to the decision model (Steinitz et al. 2000).
Short-term
Drought
Livestock
Grazing
Fire
Urbanization Suppression
Perennial Grassland/Desertscrub
Fragmentation and Decline in Extent
Mesquite Woodland
Encroachment
and Aggregation
to Form Clusters
Urban Centers
Expand
by Aggregation
\
Individual Clusters
Coalesce to Form
Large Patches
Urban Areas
Coalesce to Form
Large Metropolitan
Complex
Closed Canopy
Woodland
Figure 9. Conceptual model of vegetation phase
transitions in a semi-arid watershed.
17
-------
The principal degradation processes that have occurred throughout the western rangelands involves 1)
changes of vegetative cover, i.e. decrease in above ground productivity and compositional diversity
(primarily manifested by the introduction of exotic annual species or native woody xerophytic shrubs and
trees) and 2) acceleration of water and wind erosion processes. Historically, these have been linked to both
human-induced and natural stressors, i.e. livestock grazing and short-term drought (Grover and Musick
1990). However, rapid urbanization in the arid and semi-arid Southwest, within the last 25 years has
become an important factor in altering land cover composition and pattern. The purpose of this research
was not to determine cause and effect, however, clearly native grassland and desertscrub communities in
the upper San Pedro River basin are rapidly declining in the wake of major phase transformation into
mesquite woodland and a newly urbanized environment (Figure 9).
Collectively, the combination of new technologies with an organizational framework for decision
analysis provides decision-makers with an improved ability to understand the conditions of current and past
environment and provides a better predictor for consequences of future actions.
In the specific example of the Upper San Pedro River (Arizona/Sonora), the area has been recognized
by the U.S. Congress and under international treaty as a site important to the conservation of North
American riparian vegetation and migratory birds (CEC 1999). Much of the current public discourse
relates to policy for preserving the transboundary wildlife species connected to the presumably imperiled
riparian corridor. The riparian habitat, although containing important resource values, represents only 1.22
per cent of the total land cover and the U.S. portion is protected by National Conservation Area status.
Although this cover type is considered the most vulnerable within the watershed, the landscape analysis
indicates that upland land cover types, i.e. grassland and desertscrub, are fast disappearing as a result of
urban development and conversion to mesquite woodland. Hence, this work offers a different perspective
to natural resource managers and policy makers whom are concerned with the preservation of biological
diversity and sustainability for present and future generations.
Future research should explore the application of integrated technologies to assess environmental
condition in other geographies and the integration of science results into a decision analysis framework.
The primary spatial datasets can be made readily available to decision-makers and landscape assessment
tools could be developed to assist in the interpretation of results within a natural resource and urban
planning process.
18
-------
Literature Cited
Bahre, C.J. 1991. Legacy of Change: historic human impact on vegetation in the Arizona borderlands.
University of Arizona Press. Tucson, AZ.
Bahre, C.J. and M.L. Shelton. 1993. Historic vegetation change, mesquite increases, and climate in
southeastern Arizona. Journal of Biogeography 20:489-504.
Branson, F.A. 1985. Vegetation changes on western rangelands. Range Monograph No. 2 Society for
Range Management. Denver, CO.
Brandt, R. 1951. Arizona and its bird life. Cleveland, OH: Bird Research Foundation.
Brown, D.E., C.H. Lowe, and C.P. Pase. 1979. A digitized classification system for the biotic
communities of North America, with community (series) and association examples for the Southwest.
Journal of the Arizona-Nevada Academy of Science 14 (Suppl. 1):1-16.
Buffmgton, L.C. and C.H. Herbel. 1965. Vegetational changes on a semi-desert grassland range from
1858 to 1963. Ecological Monographs 35:139-164.
Clements, F.E. 1916. Plant Succession: An analysis of the development of vegetation. Publication 242.
Carnegie Institution of Washington. Washington, DC.
Commission for Environmental Cooperation. 1998. Advisory Panel Report on the Upper San Pedro River
Initiative: Recommendations and Findings presented to the Commission for Environmental
Cooperation. Montreal, Canada.
Commission for Environmental Cooperation. 1999. Sustaining and Enhancing Riparian Migratory Bird
Habitat on the Upper San Pedro River. Final Draft from the San Pedro Expert Study Team.
Montreal, Canada.
Congalton, R.G. and R.A. Mead. 1983. A quantitative method to test for consistency and correctness in
photo-interpretation. Photogrammetric Engineering ad Remote Sensing 49(1): 69-74.
Congalton, R.G. 1991. A review of assessing the accuracy of classifications of remotely sensed data.
Remote Sensing of Environment 37:35-46.
Congalton, R.G. and K. Green. 1993. A practical look at the sources of confusion in error matrix
generation. Photogrammetric Engineering and Remote Sensing 59(5): 641-664.
Congalton, R.G. and K. Green. 1998. Assessing the Accuracy or Remotely Sensed Data: Principles and
Practices. Lewis Publishers. New York, NY.
19
-------
Covington, W.W. and M.M. Moore. 1992. Postsettlement changes in natural fire regimes: implications
for restoration of old-growth ponderosa pine forests. In Kaufmann, M.R., W.H. Moir, and R.L.
Bassett (Eds.), Old-growth forests in the Southwest and Rocky Mountain Regions USDA Forest
Service, General Technical Report RM-213. Pp. 81-99. Fort Collins, CO.
Dimyati, M., K. Mizuno, S. Kobayashi, and T. Kitamura. 1996. An analysis of land use/cover change
using the combination of MSS Landsat and land use map - a case study in Yogyakarta, Indonesia.
International Journal of Remote Sensing 17:931-944.
Dyksterhuis, E.J. 1949. Condition and management of rangeland based on quantitative ecology. Journal
of Range Management. 2:104-115.
Flather, C.H., S.J. Brady, and D.B. Inkey. 1992. Regional habitat appraisals of wildlife communities: a
landscape-level evaluation of resource planning model using avian distribution data. Landscape
Ecology 7:137-147.
Foreman, R.T.T. and M. Gordon. 1986. Landscape Ecology. John Wiley and Sons. New York.
Graham, R.L., C.T. Hunsaker, R.V. O'Neill, and B.L. Jackson. 1991. Ecological risk assessment at the
regional scale. Ecological Applications 1:196-206.
Grover, H.D. and H.B. Musick. 1990. Shrubland encroachment in southern New Mexico, U.S.A.: an
analysis of desertification processes in the American Southwest. Climatic Change 17:305-330.
Hastings, J.R. and R.M. Turner. 1965. The Changing Mile: an ecological study of vegetation change with
time in the lower mile of an arid and semi-arid region. University of Arizona Press. Tucson, AZ.
Houghton, R.A., J.E. Robbie, J.M. Melillo, B. Moore, G.J. Peterson, G.R. Shaver, and G.M. Woodwell.
1983. Changes in the carbon content of terrestrial biota and soil between 1860 and 1980: a net release
of CO2 to the atmosphere. Ecological Monographs 53:235-262.
Humphrey, R.R. 1958. The desert grassland: a history of vegetational change and an analysis of causes.
Bot. Rev. 24:193-252.
Jensen, J.R., S. Narumalani, O. Weatherbee, and H.E. Mackey. 1993. Measurement of seasonal and
yearly cattail and waterlily changes using multi-date SPOT panchromatic data. Photogrammetric
Engineering and Remote Sensing 59:519-525.
Jensen, M.E. and R. Everett. 1994. An Overview of Ecosystem Management Principles. Pp. 7-16, In:
Jensen ME and PS Bourgeron (eds.), Eastside Forest Ecosystem Health Assessment, Vol. II,
Ecosystem Management: Principles and Applications. PNW-GTR-318. USDA Forest Service,
Missoula, MT.
Kaufmann, M.R, R.T. Graham, D.A. Boyce, Jr., W.H. Moir, L. Perry, RT. Reynolds, R.L. Bassett, P.
Mehlop, C.B. Edminster, W.M. Block, and P.S. Corn. 1994. An ecological basis for ecosystem
management. USDA Forest Service, General Technical Report RM-246. Fort Collins, CO.
20
-------
Kepner, W.G., C.J. Watts, C.M. Edmonds, J.K. Maingi, S.E. Marsh, and G. Luna. 2000. A landscape
approach for detecting and evaluating change in a semi-arid environment. Environmental Monitoring
and Assessment 64: 179-195.
Loveland, T.R., J.W. Merchant, D.P. Ohlen, J.F. Brown. 1991. Development of a land-cover
characteristics database for the conterminous U.S. Photogrammetric Engineering and Remote Sensing
57: 1453-1463.
Ma, Z. and R.L. Redmond. 1995. Tau coefficients for accuracy assessment of classification of remote
sensing data. Photogrammetric Engineering and Remote Sensing 61(4): 435-439.
Marsh, S.E., J.L. Walsh, and C. Sobrevila. 1994. Evaluation of airborne video data for land cover
classification accuracy assessment in an isolated Brazilian forest. Remote Sensing of Environment
48:61-69.
McDonnell, M.J. and S.T.A. Pickett, eds. 1993. Humans as Components of Ecosystems. Springer-
Verlag, New York.
McPherson, G.R. and J.K. Weltzin. 2000. Disturbance and climate change in United States/Mexico
borderland plant communities: a state-of-the-knowledge review. General Technical Report RMRS-
GTR-50. USDA Forest Service, Rocky Mountain Research Station. 24pp.
Maser, C. 1990. On the naturalness of natural areas: a perspective for the future. Natural Areas Journal
10: 129-133.
Miguel-Ayanz, J.S. and G.S. Biging. 1997. Comparison of single-stage and multi-stage classification
approaches for cover type mapping with TM and SPOT data. Remote Sensing of Environment 59:92-
104.
Mouat, D.A., G.G. Mahin, and J. Lancaster. 1993. Remote sensing techniques in the analysis of change
detection. Geocarto International 2:39-50.
Mouat, D.A. and J. Lancaster. 1996. Use of remote sensing and GIS to identify vegetation change in the
upper San Pedro River watershed, Arizona. Geocarto International. 11(2):55-67.
Neilson, R.P. 1986. High resolution climatic and Southwest biogeography. Science. 232:27-34.
O'Neill, R.V., D.L. DeAngelis, J.B. Waide, and T.F.H. Allen. 1986. A hierarchical concept of
ecosystems. Princeton University Press. Princeton, NJ.
O'Neill, R.V., B.T. Milne, M.G. Turner, and R.H. Gardner. 1988. Resource utilization scales and
landscape pattern. Landscape Ecology. 2:63-69.
O'Neill, R.V., C.T. Hunsaker, K.B. Jones, K.H. Riitters, J.D. Wickham, P.M. Schwartz, LA. Goodman,
B.L. Jackson, and W.S. Baillargeon. 1997. Monitoring environmental quality at the landscape scale:
using landscape indicators to assess biotic diversity, watershed integrity, and landscape stability.
BioScience 47:8, 513-519.
21
-------
O'Neill, R.V. 1999. Ecosystems on the landscape: The role of space in ecosystem theory. In: Jorgensen,
S. (ed) Handbook of Ecosystem Modeling.
Pillon, P.G., P.J. Howarth, and R.A. Bullock. 1988. An enhanced classification approach to change
detection in semi-arid environments. Photogrammetric Engineering and Remote Sensing 54(12): 1709-
1716.
Ramsey, E.W. and S.C. Laine. 1997. Comparison of Landsat Thematic Mapper and high resolution
photography to identify change in complex coastal wetlands. Journal of Coastal Research 13:281-292.
Schlesinger, W.H., J.F. Reynolds, G.L. Cunningham, L.F. Huenneke, W.M. Jarrell, R.A. Virginia, and
W.G. Whitford. 1990. Biological feedbacks in global desertification. Science 247: 1043-1048.
Simpson, G.G. 1964. Species density of North American recent mammals. Systematic Zoology 13:57-73.
Schwartz, P. 1996. The art of the long view: paths to strategic insight for yourself and your company.
Currency Doubleday, New York, New York. 272 pp.
SPSS Inc. 1998. SYSTAT. Version 8.0, Chicago, Illinois.
Steinitz, C. 1990. A framework for the theory applicable to the education of landscape architects (and
other environmental design professionals). Landscape Journal 9:2. Pp. 136-143.
Steinitz, C. 1993. A framework for theory and practice in landscape planning. GIS Europe.
Steinitz, C., M. Binford, P. Cote, T. Edwards, Jr., S. Ervin, R.T.T. Forman, C. Johnson, R. Kiester, D.
Mouat, D. Olson, A. Shearer, R. Toth, and R. Wills. 1996. Biodiversity and Landscape Planning:
Alternative Futures for the Region of Camp Pendleton, California. Graduate School of Design,
Harvard University for the Biodiversity Research Consortium. Pp. 142.
Steinitz, C., R. Anderson, H. Arias, S. Bassett, M. Cablk, M. Flaxman, T. Goode, R. Lozar, T. Maddock,
D. Mouat, W. Rose, R. Peiser, and A. Shearer. 2000. Summary Report: Alternative Futures for the
Upper San Pedro River Basin Arizona, U.S.A. and Sonora, Mexico. Harvard University, Graduate
School of Design, Cambridge, MA. Pp. 19.
Story, M. and R.G. Congalton. 1986. Accuracy assessment: a users perspective. Photogrammetric
Engineering and Remote Sensing 52(3): 397-399.
Swanson, F.J., J.A. Jones, D.A. Wallin, and J.H. Cissel. 1993. Natural variability- implications for
ecosystem management. In Jensen, M.E. and P.S. Bourgeron. Eastside forest ecosystem health
assessment, Vol. II, Ecosystem management: principles and applications. USDA Forest Service,
Northern Region, Missoula, MT. Pp. 85-99.
Tellman, B., R. Yarde, and M.G. Wallace. 1997. Arizona's Changing Rivers: How People Have
Affected the Rivers. Water Resources Research Center Issue Paper 19, University of Arizona, Tucson,
AZ.
22
-------
Turner, B.L., ed. 1990. The Earth as Transformed by Human Action. Cambridge University Press with
Clark University. Cambridge, MA.
Urban, D.L., R.V. O'Neill, and H.H. Shugart Jr. 1987. Landscape ecology. BioScience 37:119-127.
U.S. Bureau of Land Management. 1998. The Upper San Pedro River Basin of the United States and
Mexico. A resource directory and an overview of natural resource issues confronting decision-makers
and natural resource managers. BLM/AZ/PT-98/021. Arizona State Office, Phoenix.
U.S. Environmental Protection Agency. 1993. North American Landscape Characterization (NALC)
Research Brief. EPA/600/S-93/0005.
U.S. Environmental Protection Agency. 2000. Environmental Planning for Communities: A Guide to the
Environmental Visioning Process Utilizing a Geographic Information System. EPA/625/R-98/003.
Westoby, M., B. Walker, and I. Now-Meir. 1989. Range management on the basis of a model which does
not seek to establish equilibrium. Journal of Arid Environments 17:235-239.
Whittlesey, S.M., R. Ciolek-Torrello, and J.H. Altschul. 1997. Vanishing River, Landscapes and Lives of
the Lower Verde Valley. SRI Press, Tucson, AZ.
Wilson, T.B., R.H. Webb, and T.L. Thompson. 2001. Mechanisms of range expansion and removal of
mesquite in desert grasslands of the southwestern United States. General Technical Report RMRS-
GTR-81. USDA Forest Service, Rocky Mountain Research Station. 23pp.
23
-------
&EPA
United States
Environmental Protection
Agency
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HERE
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Office of Research and Development
National Exposure Research Laboratory
Environmental Sciences Division
P.O. Box 93478
Las Vegas, Nevada 89193-3478
Official Business
Penalty for Private Use
$300
EPA/600/R-02/039
June 2002
------- |