Explorers from competing teams race to find a mysterious lost city in the heart of Africa. The American team is continuously in touch with its Houston home base through satellite communications. In flight, team leader Karen Ross displays a map of Africa on her computer screen and notes the multicolored lines suggesting different routes from city to city and into the rain forest. Each pathway is accompanied by a precise estimate of travel time to the final destination. Zooming in on the target area, she switches to satellite images and interprets them in shades of blue, purple, and green. At each checkpoint, the team reports its progress and gets a revised estimate of arrival time.
Beset by difficulties, the expl orers ask for a faster route, but the computer says the alternative is too dangerous. A simulation model with data representing geology, terrain, vegetation, weather, and many other geographic factors predicts local hazards, including the impending erupti on of a nearby volcano. The Americans take the faster route anyway and beat the odds.
his fictional account of emerging geographic information system (GIS)
technologies comes from Michael
Crichton's 1980 novel Congo, which was made
into a 1995 movie. The same technologies were highlighted in Clive Cussler's
1988 techno-thriller Treasure. In reality, GIS technology began more than a
quarter of a century ago at key universitie
s and government laboratories in the
United States and Canada. Since 1969, Oak Ridge National Laboratory has been
among the leading institutions in this diverse, now booming field. GIS has been
evolving through new forms
and applications ever since. Consider the following
examples of GIS applications that rival and sometimes exceed Crichton's
futuristic vision.
The roots of our remote sensing and GIS tradition started early at ORNL; more than 20 years ago, ORNL scientists studied some of the first satellited ata from Landsat satellites (then called ERTS). By analyzing computer images of the Cumberland Mountains north of Oak Ridge, we were able to compute and display a three-dimensional perspective view of the coal strip mines in the area and superimpose the n earby streams on the terrain. After developing spatial models, we determined which streams were most likely to receive acid drainage from the strip mines. The visual impacts of strip mining on Oak Ridge residents were also predicted.
With or without satellite imagery, GIS is a powerful tool. In 1990, when the United States and other nations responded to Saddam Hussein's invasion of Kuwait, military leaders mounted the largest a nd most rapid deployment of military personnel and equipment ever attempted. The massive logistics were processed on the Airlift Deployment Analysis System (ADANS) developed at ORNL. ADANS, operating on networked computers, draws on a variety of logistic and spatial technologies to efficiently schedule the transport of U.S military troops and equipment to trouble spots anywhere in the world. Since 1990, ADANS has been used to deploy military personnel and equipment not only to the Persian Gulf but also to Somalia, Rwanda, and Haiti.
In 1995, at ORNL's World Wide Web Showcase, Peter Pace showed a colorful high-resolution image of ORNL buildings and the roads, streams, and forested areas of the surrounding reservation. The view on his computer screen w as constructed from a series of aerial photographs that had been scanned and converted to form a digital image. Various computer techniques were used to enhance and blend a series of images, eliminating unwanted elements and bringing out important details . Special photogrammetric techniques were used to remove distortions from the digitized photos. Each pixel (tiny rectangular element) on the screen represents 0.25 square meter (m2) on the ground. Spatial registration of geographical features in the image is sufficiently accurate that a highly detailed map can be overlaid on the image. Pace zoomed in on a cooling tower and magnified it enough to see the blades of a fan. He printed out an image of the cooling tower alone. He and other ORNL researchers arep reparing geographical data and imagery developed at Oak Ridge for distribution to selected users of the World Wide Web through Netscape, a navigational tool for accessing still and animated images as well as audio and text from the Internet.
Recent g rowth of GIS markets has been phenomenal. In 1994, GIS was listed under "Whole Systems" in the Whole Earth Catalog. Tens of thousands of people and organizations--universities, research centers, municipal planners, tax assessors, corporations, and resourc e managers--have come to depend on GIS for geographic data collection, analysis, and display. The commercial GIS industry, which started in the early 1980s, is now estimated to be worth $3.5 billion.
Today's rosy picture sharply contrasts with the si tuation in 1969 when GIS first began at ORNL. At that time only a few centers--principally Environment Canada, the U.S. Geological Survey, < a href="http://www.hup.harvard.edu/">Harvard University, and ORNL--shared a common interest in solving the riddle of geographic analysis. Along with scientists from these centers and a few leading research universities, early members of ORNL's GIS and Computer Modeling (GCM) Group, led by Richard Durfee, contributed many of the developments that made the current boom possible.
These contributions include fundamental development of early geographic computational techniques that supported and accel erated the growth of a commercial industry; development and integration of key GIS data bases and methodologies; and use of geographic and spatial analysis to provide information to help policymakers make decisions on national issues, such as development of energy sources and protection of water resources and fish populations, and to help government agencies assess natural resources and environmentally contaminated sites needing remediation.
Many people think of GIS as a computer tool for making maps. Actually, it is a complex technology beginning with the digital representation of landscapes ca ptured by cameras, digitizers, or scanners, in some cases transmitted by satellite, and, with the help of computer systems, stored, checked, manipulated, enhanced, analyzed, and displayed as data referenced to the earth. This spatial information includes earth coordinates and geometric and topological configurations to portray spatial relationships between features such as streams, roads, cities, and mountains. GIS is "a digital representation of the landscape of a place (site, region, planet), structured to support analysis." Under this broad definition, GIS conceivably may include process models and transport models as well as mapping and other spatial functions. The ability to integrate and analyze spatial data is what sets GIS apart from the multitude of graphics, computer-aided design and drafting, and mapping software systems.
Typical sources of geographic data for computer manipulation include digitized maps, field survey data, aerial photographs (including infrared photographs), and satellite imagery. Most image data are collected using remote sensing techniques. Aerial photographs are normally taken with special mapping cameras using photographic film. Most commercially available satellite imagery is collected using multispectral scanners, w hich record light intensities in different wavelengths in the spectrum--from infrared through visible light through ultraviolet light.
Spatial information can be represented in two distinctly different forms. Satellite images, for example, usually app ear as raster data, a gridded matrix in which the position of each data point is indicated by its row and column numbers. Each position on a computer screen or map thus corresponds to the position on the ground measured by the satellite as it passes overh ead. In contrast, cartographic features such as roads, boundaries, buildings, and contour lines usually are represented in vector form. In digitizing a lake, for example, the shoreline can be indicated as a series of points and line segments. In this case , each point is measured in Cartesian (X, Y) coordinates and each line segment is measured as a vector leading from one point to the next. The more points recorded, the more detailed the shoreline will be. Both forms, raster and vector, are essential to s upport environmental restoration projects on the ORNL reservation, for instance, and the software must be capable of rapid conversion from one form to the other.
For such geographic information to be meaningful, it must be accompanied by "metadata" d ocumenting the source, description, specifications, accuracy, time of acquisition, and quality of each data element. As GIS technologies and multitudes of geographic data bases have spread to the desktop in the past decade, metadata have become very impor tant. Good metadata are essential in determining fitness of the geospatial data for each intended use--that is, determining which applications can be accomplished while ensuring the desired quality of results and decisions made from those data.
In this article, we focus primarily on ORNL's role in the development and application of GIS to real-world problems over the past 25 years. Over this time, hundreds of projects and tasks involving GIS have been carried out by several organizations at ORNL involving a number of scientists, managers, and sponsors. It would be impossible to mention them all, but we do recognize and appreciate their significant contributions and collective vision for advancing GIS technologies over the years. In addition to t he Computational Physics and Engineering Division, the examples of collaborating organizations within Martin Marietta Energy Systems have included the Energy Division, the Environmental Sciences Division, Chemical Technology Division, the Environmental Restoration Program, Biology Division, Data Systems Research and Dev elopment, and the Hazardous Waste Remedial Action Program. We highlight several of the larger efforts to illustrate the diversity of applications and techniques. We describe some of the early GIS developments and summarize some of the current systems capa bilities. We offer examples in which GIS has proven useful in research and decision support.
Actually, the term GIS, though first introduced in 1964, was not extensively used until the late 1970s. The first c omprehensive geographic data management system--called the Oak Ridge Regional Modeling Information System (ORRMIS)--was developed in 1974 at ORNL by Durfee. Its purpose was to integrate and support the data management needs of a series of regional analyti c models depicting and forecasting land-use, environmental, socioeconomic, and sociopolitical activities in the East Tennessee region.
Many early ORNL developments in GIS that are commonplace today are remarkable primarily because of their dates. Exam ples from the 1970s and early 1980s include perspective and isometric drawings of cartographic surfaces, integration of remote sensing and statistical techniques with GIS, raster-vector transformation, viewshed calculation, polygon intersections, transpor tation routing models, and true three-dimensional (3-D) imaging.
ORNL has a long heritage of GIS research, development, and application to complex problems ranging from national issues to site-specific impacts. After presenting an overview of GIS tech nology development in ORNL's computing environment, we discuss three eras of GIS history at ORNL--regional modeling and fundamental development (1969-1976), integrated assessments (1977-1985), and issue-oriented research and analysis (1986-1995).
In the past 25 years, GIS software development and applications have migrated from mainframe computers to minicompu ters to personal computers (PCs) to networked UNIX workstations. GIS software is now being modified for use on parallel processors and supercomputers, such as the IBM SP2 and Intel Paragon X/PS machines at ORNL.
In the very early 1970s, a technological feat was the development of a computer-generated 3-D perspective movie by Tom Tucker of ORNL. The movie simulated terrain and population changes over a 40-year p eriod in the Norris, Tennessee, area as Norris Dam began operation as a hydroelectric facility.
Over the years, one of the benefits of these spatial technologies has been their applicability to many different types of problems. One example was the de velopment in the early 1980s of electron microscope tomography for 3-D reconstruction of DNA chromosomes as a collaborative effort led by Don Olins and his colleagues in ORNL's Biology Division in cooperation with the GCM group at ORNL. Adaptation and dev elopment of hardware and software for a commercial remote sensing system, I2S, on GCM minicomputers played a major role in the analysis of electron micrographs and display of chromosome structures. When it was determined that more sophisticated true 3-D d isplays were needed, a special varifocal mirror display was built. Depth visualization was provided by a vibrating mylar mirror synchronized with a monitor mounted above the mirror whose image was reflected to the operator. Data at greater depths were di splayed when the mirror was at a greater deflection, thus varying the focal length to correspond to the appropriate depth. This occurred at a rate of 60 times per second, so the observer saw a continuous 3-D image.
Another ORNL breakthrough in GIS technology in the mid-1970s was the
development of vector-based algorithms and their eventual integration with
raster-based grid cell systems. The GCM group used these techniques for a
ll
types of water-resource and energy-related studies in collaboration with the
Energy Division. In the late 1970s, ORNL developed transportation data bases
and capabilities for routing hazardous was
tes across the United States. Through
use of GIS technology to match proposed routes with population density, the
health and safety risks of hazardous waste transport could be estimated.
Some of the latest GIS research under way at ORNL involves developing
software for use on parallel-processing supercomputers. Very recent work has
shown that, by significant improvement of algorithms and by using parallel
process
ors on ORNL supercomputers, the transformation and interpolation
(estimation of values between data points) of large GIS data sets can be done
50 to 18,000 times faster than on smaller Sparc workstations. Because of the
explosion in data collection from a
ll types of earth sensor systems,
workstations and supercomputers must be integrated to handle massive volumes of
data.
We are also integrating portable GIS capabilities with GPS in which
relative positions of objects on the earth can be pinpointed i
n real time by
satellite sensors in communication with hand-held devices. As this technology
becomes more commonplace, geospatial data will be collected at an ever
increasing rate. Real-time airborne GPS techniques have already been used in
aerial surveys
of the Oak Ridge Reservation to collect high-resolution aerial
photography with accurate positioning information. Computerized stereo
techniques are being used with special goggles to help generate
orthographic
images (digital images corrected for camera, terrain, and other distortions)
from stereo photography. Also, 3-D subsurface modeling and visualization are
being done for hazardous waste studies.
To provide intelligent and efficient access to large amounts of geospatial
data, work is under way to prepare and load this information on Internet and
World Wide Web servers, which can be accessed by data browsing tools such as
Mosaic and Nets
cape. These capabilities are important to the Oak Ridge user
community and to the success of the National Spatial Data Infrastructure (NSDI)
during the 1990s.
In 1969, the U.S. Congress passed
the National Environmental Policy Act (NEPA), the National
Science Foundation
(NSF) initiated the Research Applied to
National Needs (RANN) program, Ian McHarg published Design with Nature, and
ORNL delved headlong into regional modeling and GIS. Clearly, NEPA was a major
impetus to the other three events.
Before NEPA, research
and development, infrastructural development, and
resource management decisions had been based almost exclusively on engineering
and cost-benefit considerations. Suddenly, NEPA thrust all large enterprises,
including the federal government itself, into a
new legal and ethical milieu in
which comprehensive, interdisciplinary analyses were absolutely essential.
Alvin Weinberg, director of the Laboratory from 1955 to 1973, immediately
recognized the need and sought to diversify the Laboratory's missions.
For GIS, the most important development in the early days at ORNL was the
Oak Ridge Regional Modeling Information System and associated tools that
supported spatial data input and display. The primary purpose was "to provide
the data management capability
for analysis models which forecast the spatial
distribution and ecological effects of activities within a geographical
region." The land-use modeling efforts became the principal impetus to remote
sensing development as well as to the GIS expansions.
Initial GIS software techniques were based on hierarchical grid cell
systems. It became apparent that additional capabilities were needed for
accurate cartographic representation and analysis of vector-based map data. By
the mid-1970s, development of so
phisticated polygonal-based GIS systems at ORNL
were well under way. Our development of efficient storage and computational
techniques for integrating raster-based grid cell and vector-based systems
opened the door to addressing larger and more complex pr
oblems with a national
scope. Incorporation of new algorithms designed by Phil Coleman and Bob Edwards
provided a capability for analyzing and displaying large national data bases.
In the mid-1970s, a
shift in federal policy greatly reduced NSF funding for
the DOE national laboratories. From then on, hardly another penny was received
to support basic research, development, or operation of GIS systems at ORNL.
The GCM group and the Energy Division shifted to applications-driven research,
the funding for which allowed continued development and operations.
We never had the luxury of focusing on a particular technology (remote
sensing or computer
cartography, for instance) to the exclusion of other
technologies. We were then, and are still, comprehensive integrators with
analytical purposes paramount in everything we do. In many respects, this
approach has been advantageous because (1) the integra
ted GIS technologies were
then applicable to a wide range of spatial problems, and (2) the
applications-driven development minimized "ivory-tower" research looking for a
problem to solve.
During the mid-to-late 1970s, the Laboratory played a major role in the
National Uranium Resource Evaluation (NURE) Program. ORNL's Computer Sciences
Division (now the Computational Physics and Engineering Divisi
on), in
cooperation with DOE's Grand Junction Office, was the national repository for
all data collected and analyzed to assess the availability and location of
potential uranium resources for future commercial nuclear power, research
reactors, and other
uses. ORNL staff were responsible for overall data
management, GIS processing, spatial analysis, and mapping. Al Brooks was
director of the Oak Ridge effort to support DOE in surveying the country for
potential uranium resources and estimating possible re
serves. Through a
multitude of subcontractors, DOE conducted both aerial radiometric and
geomagnetic surveys and hydrogeologic ground sampling on a
quadrangle-by-quadrangle basis across the United States.
The aircraft had special sensors to detect ra
dioactive isotopes of elements
such as bismuth, thallium, and potassium as well as magnetic fields.
One example of highly specialized GIS work at ORNL was Ed Ti
nnel's
development, in cooperation with Bill Hinze of Purdue
University , of spatial
filtering, interpolation, and contouring techniques to convert one-dimensional
flight line data into meaningful maps of regional
magnetic data. The purpose was to use these data to help study geologic features
and identify magnetic anomalies that might indicate the presence of mineral
deposits. These maps were also provided to the U.S. Geological Survey for
publication. This was o
ne of the earliest projects that required the handling
of massive amounts of spatial, tabular, and textual information of many
different types. During this time specialized GIS hardware systems were
implemented to provide new ways of digitizing and displa
ying large amounts of
geographic data.
Multiple energy assessments were early examples of policy analysis using
GIS. A flurry of activity began each time President Jimmy Carter
proposed a new
National Energy Plan. Econometric models were run by the Energy Information
Administration to project, as far as the year 2000, energy demand and fuel use
by type in each major region of the country. These regional projections were
passed t
o ORNL, where energy demand was disaggregated by Dave Vogt to Bureau of
Economic Analysis Regions and supply was allocated to counties. Around 1980 Ed
Hillsman and others of the Energy Division proje
cted electrical generation from
each existing plant and simulated construction or retirement of different
plants by fuel type to determine if the president's goals would be met. Dobson
and Alf Shepherd projected the amount of water needed for energy produ
ction and
compared it with the amount of water available in each basin in the United
States. ORNL's projections of electrical generation for
different areas were passed to other national laboratories (Argonne,
Brookhaven, Los Alamos, and the Solar Energy Research Institute), which used
the information to evaluate effects on air quality, water quality, and labor
supply. All results were reported to DOE, which conducted policy analysis of
the feasibility of each proposed p
lan. Results of one GIS assessment of the
projected water consumption by energy facilities in the Ohio River Basin were shown to President Carter in a live presentation using a
graphics station when he visited ORNL in 1978.
In short, as early as the 1970s the nation's energy system and many
pertinent physical and cultural features were simulated through GIS in linkage
with econometric models, location-alloc
ation models, environmental assessment
models, and spatial data bases. The principal output was by county, but many of
the data bases and computations covered details finer than the county level.
For example, the data bases included population at the Enum
eration District
level, all power plants over 10 megawatts in generating capacity and all U.S. Geological Survey stream gauging station records. The
models were as sophisticated as any i
n use at that time with or without GIS.
Another major multiyear effort involving ORNL researchers in the early
1980s was the development of a national aband
oned mine lands inventory for the
Office of Surface Mining (OSM) of the Department of the Interior. This effort,
headed by Bob Honea, was based on federal legislation mandating that abandoned
mine lands be
reclaimed to protect human health, safety, welfare, and the
environment, using funds collected as taxes on mining operations. A national
inventory of abandoned mine lands was necessary to determine the affected areas
in urgent need of reclamation and to e
stablish priorities for reclamation of
other sites. The effort was initially viewed as a technology-based project
involving heavy use of remote sensing, GIS, record-based information systems,
and statistical tools.
It was anticipated that analysis of
Landsat satellite imagery would be a
key ingredient for identifying detailed impacts from the disturbed, abandoned
lands. However, an interesting turn of events made the project much more
difficult than expected. When attempts were made to use results fr
om satellite
analyses to meet the mandates in the legislation, we found that the worst
threats to human health and safety (e.g., open mine shafts, acid drainage,
polluted water supplies) could not be determined from satellite data. Major
environmental imp
acts could be addressed by analyzing satellite images, but
health and safety impacts and reclamation cost estimates required field data
collection and field assessment efforts. Thus, a major field collection effort,
which included on-site interviews with
affected populations, was carried out in
conjunction with the state agencies of all the coal-mining states. Unique information handling techniques were devised to standardize and
computerize textual, tabular, temporal, and spatial data from forms and maps
that could then be linked with GIS for spatial aggregation, statistics, and
mapping. Don Wilson was responsible for overseeing the computerization of all
this information and development of a consolidated data base. These results
could then support asses
sments at the state, regional, and national levels to
aid OSM in allocating reclamation funds and overseeing mitigation of the
severest problems.
Methodologies developed at ORNL
for one application were readily adapted
and applied to other problems. For example, our initial demographic work of the
late 1970s was extended to compute detailed population distributions for any
place or region in the United States.
The technique was used by Phil Coleman and Durfee to compute population
distributions around all nuclear power plants in the United States. Our
results, including the calcul
ation of population exclusion zones, enabled the
Nuclear Regulatory Commission to assess these exclusion areas--regions where
additional nuclear power plants should not be built because too many people
live or work there--to help make planning and licensi
ng decisions.
Starting in the mid-1980s, the emphasis shifted again, this time in a very
positive direction, as GIS became an important tool in topical research on
scientific issues
of national interest, as illustrated in these four
examples.
Lake Acidification and Acid Precipitation. Acid precipitation can cause
water in lakes to acidify, potentially reducing fish populations. Lake
acidification and other environmenta
l issues that may be related to acid
precipitation were major themes of GIS work at ORNL in the late 1980s. The
Environmental Sciences Division (ESD) was involved prominently in the National
Acid Precipitation Asse
ssment Program (NAPAP), especially the National Surface
Water Survey. Through extensive collaboration with U.S. Environmental Protection Agency (EPA) laboratories and numerous universities and private
firms, Dick Olson, Carolyn Hunsaker, and other ESD personnel collected,
managed, and analyzed massive geographic data bases for lakes and watersheds
throughout the United States. The goal was to characterize contemporary
chemistry, temporal variability,
and key biological resources of lakes and
streams in regions potentially sensitive to acid precipitation.
Simultaneously, the Energy Division approached the same problem from a
different perspecti
ve. While NAPAP focused on impacts of acid precipitation,
this project focused on watersheds and investigated possible causes of lake
acidification.
In 195
0, a huge storm with heavy rain and 105-mile-per-hour winds blew down
numerous trees in 171,000 hectares of forest in the Adirondack Mountains of New
York. In the 1980s it was observed that several lakes in
the area were acidified, so one hypothesis was t
hat the blowdown of the forest
might be a cause. To determine if a relationship existed between the forest
blowdown and lake acidification, Dobson and Dick Rush of ORNL and Bob Peplies
of East Tennessee State Univers
ity used an approach that combined GIS and
digital remote sensing with the traditional field methods of geography. The
methods of analysis consisted of direct observation, interpretation of
satellite images and aerial photographs, and statistical comp
arison of two
geographical distributions--one representing forest blowdown and another
representing lake chemistry.
Both efforts illustrate an ORNL strength--the
ability to assemble
multidisciplinary teams and multiple organizations to attack complex problems.
GIS, in itself, is an integrating technology because it draws together
different sciences that have a common need for spatial data, visualization, and
anal
ysis capabilities. Such was the case in the acidification studies just
described. Although primary responsibility for these two efforts rested
separately in the Energy and Environmental Sciences divisions, the GCM group
was heavily involved in both efforts. Thus, considerable interaction took place
between the two projects. Since then, ESD, in cooperation with GCM and other
groups, has continued to expand its
GIS capabilities and resources. ESD
scientists now have hands-on access to GIS systems and data bases to support a
multitude of research efforts.
Coastal Change Analysis. For decades, the National Marine Fisheries Service
(NMFS) of the National Oceanic and Atmospheric Administration (NOAA) has been
concerned about declining fish populations in U.S. coastal waters. Suspecting
that these declines
might be caused by losses of habitat, such as saltmarshes
and seagrasses, and increases in pollution resulting from expanding urban and
rural development, as well as agriculture, NMFS initiated a research effort to
solve the technical, institutional, and
methodological problems of large-area
change analysis--methods for determining the time, location, and degree of
changes in large areas to better understand changes in ecosystems and
ecological processes. ORNL has led the technical effort to improve metho
ds for
analyses of changes in uplands and wetlands, detected by satellite sensors, and
to perform prototype satellite change analysis of the Chesapeake Bay.
Integration of these remote sensing and GIS methodologies in a laboratory
environment, in field in
vestigations, in workshop settings, and for
presentations and briefings in policy and management arenas shows how much this
evolving technology is becoming ingrained in all phases of earth-sciences
work.
The Coastal Change Analysis Program (C-CAP)
is developing a nationally standardized data base of land cover and land-cover
change in the coastal regions of the United States. As part of the Coastal Ocean Program (COP), C-CAP
inventories coastal and submerged wetland habitats and adjacent uplands and
monitors changes in these habitats over one to five years. This type of
information and frequency of detection are required to
improve scientific
understanding of the linkages of coastal and submerged wetland habitats with
adjacent uplands and with the distribution, abundance, and health of living
marine resources. Satellite imagery (primarily Landsat Thematic Mapper), aerial
ph
otographs, and field data are interpreted, classified, analyzed, and
integrated with other digital data in a GIS. The resulting land-cover change
data bases are disseminated in digital form for use by anyone wishing to
conduct geographic analysis in the c
ompleted regions.
Although the Chesapeake Bay prototype focused on a single region, its
purpose was to provide a technical and methodological foundation for change
analysis throughout the entire U.S. coast. Four regional worksho
ps (Southeast,
Northeast, Great Lakes, and Pacific) addressed a full range of generic issues
and identified the issues of special interest in each major coastal division of
the United States. Ultimately, the protocol development effort involved more
than
250 technical specialists, regional experts, and agency representatives.
During the summer of 1994, field work was conducted in the Gulf of Maine,
along the Oregon and Washington coast, and in Alaska. The Alaskan study is
especially interesting.
In 1986, the Hubbard Glacier moved, closing the narrow opening between the
glacier and Russell Fiord's Gilbert Point on the coastline of Alaska. The ice
dam later burst as the fiord's water rose, and the narrow opening was restored.
The event was worrisom
e to salmon fishermen because the fiord's alternative
outlet to the sea could destroy the unique stock of sockeye salmon that spawn
in the Situk River. The glacier is poised to move again, and the new, more
permanent ice dam that is expected could cause t
he fiord to empty through the
Situk watershed, drastically altering its ecosystem.
Using satellite images of the Alaskan coastline from various years, we are
identifying changes in the Alaskan coastline that will help predict the impacts
on fisheries
when the glacier closes the gap again. If the Situk River salmon
are threatened, it may be necessary to transplant some of them to less
vulnerable streams.
To verify the accuracy of our interpretations of the satellite data, we
visit the imaged sites. In 1993 and 1994, Ed Bright and Dobson
went to Alaska
to conduct field verification of a 1986 land-cover classification in the
Yakutat Foreland and Russell Fiord. Now, when we do field work, we use a
hand-held GPS device linked directly to a color laptop computer. Commercial
software integrate
s the live GPS location coordinates with raster images
representing land cover and with vector images representing other features such
as roads. The device has more than doubled productivity in the field. We are
currently designing a modeling approach tha
t will link GIS, transport models,
and process models to address the linkage between land-cover change and
fisheries.
Environmental Restoration. To clean up a legacy of environmental
contamination and to comply with environmental regulation
s, U.S. government
facilities must locate, characterize, remove or treat, and properly dispose of
hazardous waste. In the 1980s, ORNL researchers helped develop geographic
workstations, spatial algorithms, 3-D subsurface modeling techniques, and data
base
systems for handling hazardous waste problems at Air Force installations.
Later, this work provided a foundation for supporting environmental restoration
activities at DOE facilities. Since the late 1980s, environmental restoration
has become a major the
me for GIS activities at ORNL. The integration of GIS
with other technologies provides an important resource to support hazardous
waste assessment and management, remediation, and policy formulation for
environmental cleanup at DOE facilities. The locatio
ns of waste areas (i.e.,
surface operable units) across the DOE Oak Ridge Reservation (ORR) are
represented by the bold polygons shown on the following map.
In conducting successful cleanup efforts and meeting regulatory
requirements at these facilities, GIS can assist in many ways. Key aspects
include investigation of the types and characteristics of contaminants; the
location of possible pollutant source
s; previous waste disposal techniques; the
spatial extent of contamination; relationships among nearby waste sites;
current and past environmental conditions, including surface, subsurface, and
groundwater characteristics; possible pollutant transport me
chanisms;
efficient methods for analyzing and managing the information; effective cleanup
strategies; and mechanisms for long-term monitoring to verify compliance.
Three programs that involve significant GIS activities in support
of environmental rest
oration (ER) in Oak Ridge include the Oak Ridge
Environmental Information System (OREIS), the Remote Sensing and Special
Surveys (RSSS) Program, and the GIS and Spatial Technologies (GISST) Program.
The OREIS effort is designed to meet environmental data
management, analysis,
storage, and dissemination needs in compliance with federal and state
regulatory agreements for all five DOE facilities operated by Lockheed Martin Energy Systems. The primary focus of th
is effort has been to develop a
consolidated data base, an environmental information system, and data
management procedures that will ensure the integrity and legal defensibility of
environmental and geographic data throughout the facilities. The informat
ion system is
composed of an integrated suite of GIS, relational data base management, and
statistical tools under the control of a user-friendly interface. The OREIS effort,
previously led by Larry Voorhees and Raymond McCord, is now being directed byD
avid Herr.
The GISST
effort, under Durfee, promotes the development,
maintenance, and application of GIS technology, data bases, and standards
throughout the ER Program. The largest activity currently under way is the
development of base map data, digital orthophotos, and ele
vation models for all
Energy Systems facilities using advanced stereo photogrammetric techniques
based on real-time airborne GPS. When completed, these terrain data will be the
most comprehensive GIS and orthoimage coverages of any DOE reservation. This
p
roject, under the technical direction of Mark Tuttle, is being carried out in
cooperation with the Tennessee Valley Authority. Desktop mapping systems are
being integrated into the daily operations of many Oak Ridge staff devoted to
monitoring and cleanin
g up the ORR. To support these activities, a repository
of the resulting data from this project is being made available to users
networked into a local file server, which will soon be accessible as a World
Wide Web server. These GIS data provide a consist
ent, current, and accurate
base map that can be integrated with all other types of environmental and
pollutant data for analysis and reporting.
The fusion
of all types of spatial data is an important tool for
any environmental activity on the ORR. Through these and
other ER programs, facility data and environmental data bases have been
developed to improve understanding of relationships among pollutant sou
rces,
surface and subsurface pathways, and receptors of environmental contaminants.
Three-dimensional modeling, data management, and contaminant analysis have been
enhanced through integration of computer tools and geospatial data. All these
resources ar
e becoming an integral part of the remediation planning and cleanup
process, supported through communication networks linking scientists,
engineers, and decision makers with analytical software and data bases.
Transportation Modeling and Analys
is. Transportation systems and networks
are crucial to the U. S. economy and way of life. GIS is used increasingly to
plan, develop, and manage transportation infrastructures (e.g., highway,
railway, waterway, and air transport networks) with the goal
of improving
efficiency in construction and operation.
Three main centers heavily involved in transportation modeling and
geographic networks are the Energy Division (ED), the Chemical Technology
Division (CTD), and the Computational Physics and Engineering Division (CPED).
CTD has been primarily s
upporting DOE transportation needs in collaboration
with CPED; ED has been supporting the Department of Transportation; and both ED
and CPED have been supporting the Department of Defense. Collectively, the
three groups have developed detailed representat
ions of highway, railway, and
waterway networks for the United States and military air transport networks for
the entire world. ED, for example, is the developer and proprietor of the
National Highway Planning System and the initial INTERLINE railway rout
ing
model. CTD has had a major responsibility for routing and assessing hazardous
materials on the nation's highway and rail systems for many years (see figure
above). They have enhanced and adapted the INTERLINE and HIGHWAY routing
models to assist in t
his work. CPED has been a major developer of the Joint
Flow and Analysis System for Transportation (JFAST), which is a multimodal
transportation analysis model designed for the U.S. Transportation Command
(USTRANSCOM) and the Joint Planning Community. Operations Desert Shield and Desert Storm (1990-1991) involved the largest
airlift of personnel and equipment from region to region ever accomplished. The
U. S. Air Force's Military Airlift Command, n
ow the Air Mobility Command (AMC),
was responsible for this movement from the United States and Europe to the
Persian Gulf region. Prior to that event, ORNL had worked with AMC to develop
the Airlift Deployment Analysis System (ADANS), a series of schedul
ing
algorithms and tools that enabled AMC to schedule missions to and from the
Persian Gulf more rapidly and efficiently than ever before. ADANS is currently
being used 24 hours per day by AMC to schedule peacetime, exercise, and
contingency missions, as
well as peacekeeping relief and humanitarian
operations. Some of the key members of the ADANS team have included Glen
Harrison, Mike Hilliard, Ron Kraemer, Cheng Liu, Steve Margle, and Irene
Robbins.
All data and algorithms are geographically explicit. The user inputs data
on a station-by-station basis with the textual network edito
r; the graphical
network editor allows the user to establish a network and to enter or to edit
information directly on a world map. With this system, it is easy to determine
how cargo and passengers were moved, how many were moved
as required, and to wha
t aircraft they were assigned.
The JFAST effort, initiated by Brian
Jones, is
designed to determine
transportation requirements, perform course of action analysis, and project
delivery profiles of troops and equipment by air, land, and sea. JFAST was u
sed
in Desert Shield to analyze the airlift and sealift transportation requirements
for deploying U.S. forces to the Middle East and predict their arrival dates
in-theater. These deployment estimates provided input for establishing concepts
and timing for
military operations. Under Brian Jones' direction during Desert
Storm, JFAST was also used to track ships, provide delivery forecasts, and
analyze what-if scenarios such as canal closings and maintenance delays. In
addition to analyzing support for human
itarian efforts such as those in Rwanda
and Somalia, USTRANSCOM and the Joint Planning Community use JFAST to determine
the transportation feasibility of deployment plans.
JFAST incorporates a graphical user interface that makes significant use of
geo
graphic display of transportation data as well as other graphic displays to
aid the planner in understanding the output from the flow models. To assist in
preparing briefings, all JFAST screens and graphic displays can be captured and
inserted directly in
to presentation software while JFAST is running. Data from
JFAST can also be sent directly to other Windows(TM)-compliant applications,
such as spreadsheets and word processing packages.
After a
quarter of a century, how have GIS developments and applications at
ORNL advanced science and served the national interest? ORNL has played an
instrumental role in the GIS revolution by establishing and implementing a
coherent vision that has been welcome
d by scientific, policy, and management
communities. ORNL has advanced the use of GIS within our national
infrastructure. Today, commercial GIS products address many of the technical
needs that required so much of our effort in the past, and the researchf
rontiers have moved on to more complex methodological issues. However, no
single commercial product today will handle all the current needs for GIS and
related spatial technologies. One of our ongoing roles will be integration of
multiple products with in
-house technologies to best meet real-world needs that
arise.
We hope to maintain a leadership position through continued advancement of
hardware and graphics systems, GIS software, and data bases that will more
effectively solve complex spatial problems. We think that knowledge-based
expert
systems will play a role in advancing future development and use of GIS
technologies. We intend to assist the GIS community in improving standards and
quality assurance procedures, and we look forward to assisting in enhancement
of the National Spatial D
ata Infrastructure.
Ultimately, we view GIS as an integrating technology with the potential to
improve all branches of science that involve location, place, or movement.
Consider, for example, that most of the advances in medical imaging have been
bas
ed on visual analysis. Imagine how much greater the potential would be if
the images were enhanced by data structures, models, and analytical tools
similar to those employed in analysis of the three-dimensional earth. We
envision that certain technologica
l thresholds will open the door to entire
fields. For example, true 3-D analysis (more than visualization) and temporal
GIS should provide new insights to geophysicists studying plate tectonics and
the dynamic forces operating beneath the earth's surface.
The single
advancement of linking GIS with environmental transport models and process
models will suddenly enable scientists and professionals in numerous other
disciplines to incorporate spatial logic and geographical analysis alongside
their traditiona
l approaches. As these and other developments take place, a
truly revolutionary new form of science should emerge.
Richard Durfee is head of the Geographic Information Systems and Computer
Modeling Group (GCM) in ORNL's Computational Physics and Engineering Division.
He is also program manager of the GIS and Spatial Technologies Program for
Environmental Restoration supporting DOE's Lockheed Martin Energy Systems
facilities. He is responsible for an advanced GI
S Computing and Technology
Center at ORNL with special facilities for analyzing and displaying all types
of geospatial information. Previously, he was head of the Geographic Data
Systems Group in the former Computer Applications Division and section head
in
the former Computing and Telecommunications Division. He joined ORNL in 1965 as
a member of the Mathematics Division. He has an M.S. degree in physics from the
University of Tennessee. An expert in GIS and remote sensing technologies, he
served on the
initial steering committee for the DOE Environmental Restoration
GIS Information Exchange conferences and for the early DOE Interlaboratory
Working Group on Data Exchange. He was also an early member of the Federal
Interagency Coordinating Committee on Di
gital Cartography. He is coauthor of
many publications and presentations, including a GIS-related presentation to
President Carter during his visit to ORNL in 1978. For more than 25 years, he
has researched and directed a wide range of GIS technologies su
pporting
hundreds of applications for more than 15 different federal agencies.
Regional Modeling and GIS Developmen
t (1969-1976)
Integrated Assessments (1977-1985)
The first
seeds of the new order were sown in 1975 when Richard Durfee and
Bob Honea used ORRMIS tools for predictive modeling of coal strip mining and
associated environmental problems. Results of this work were presented to
Robert Seamans, head of the Energy Res
earch and Development Administration
(ERDA), predecessor to DOE. Soon afterward, we became heavily involved in
siting analysis. In 1975 and 1976, ORNL systems were used, along with data from
the Maryland Automated Geographic Information System, to support
conflict
resolution in power plant siting. By the late 1970s, these systems were heavily
involved in decision support for federal energy policy and resource management.
ORNL employed GIS extensively to evaluate the environmental impacts of various
propos
ed National Energy Plans. Later, we predicted the amount of coal that
could be produced from federally leased lands and evaluated the impacts on
energy supply of designating certain lands as wilderness areas, thus protecting
them from exploration for and
extraction of oil, gas, and uranium. 
Issue-Oriented Research and Analysis (1986-1995)
Land cover change analysis has been completed for the Chesapeake Bay based
on Landsat Thematic Mapper (TM) data. The resulting
data base consists of land cover by class for 1984,
land cover by class for
1988 and 1989, and a matrix of changes by class from 1984 to 1988-89. We found
that, contrary to popular opinion, marshland in the Chesapeake Bay region
increased slightly during the period. However, both forested wetlands and
upl
and forests declined significantly, while land development expanded rapidly.
At greater detail, we observed the formation of a new barrier island and
recorded lateral movement of portions of its tip by almost a kilometer.
In studying satellite images, we have
looked for changes in land cover
from 1986 on and tried to quantify these changes on a regional basis. For
example, we have looked at changes in the size and shape of woodlands,
wetlands, grasslands, and bare ground over a period of years to characterizec
oastal changes. We are trying to model the direct relationship between
land-cover changes and ecological processes.
The RSSS Program under Amy King supports ER site characterization,
problem identification, and remediation efforts through the collection and
analysis of data from aircraft and other re
mote sensors. One example has been
helicopter radiometric surveys to determine gamma radiation levels across
mapped areas of DOE facilities. GIS and remote sensing techniques also aid in
the interpretation and visualization of airborne multispectral scan
ner data,
thermal imagery, infrared and natural color photography, and electromagnetic
and magnetic survey analyses. The following map shows examples of these types of processed information.
Integrated results from such analyses are useful in locating po
tentially
contaminated and affected areas, as well as possible underground structures
that may be pertinent to hazardous waste burial and migration. Another example
has been the delineation of waste trenches in burial ground areas that may be a
source of
waterborne contaminants requiring remediation. The RSSS Program is
also responsible for surveys of environmentally sensitive areas on the ORR.
ORNL's Role in the GIS Revolution
For the future, we envision linkages of GIS with environmental transport
models and process models traditionally used by biologists, e
cologists, and
economists; implementation of GIS and digital remote-sensing techniques on
supercomputers; 3-D GIS visualization and analysis; temporal analysis in a
spatial context, and improved statistical analysis capabilities for geographic
and other s
patial data. The use of supercomputers will become even more
important as new data collections--for example, the next generation of
high-resolution satellite imagery--inundate the scientific community with
terabytes of information. Justification for colle
cting and using these data
will depend on the ability to extract meaningful information using
supercomputer technology. We are currently addressing these and other
technological issues, such as GIS animation, telecommunications, and real-time
GPS and vide
o linkage with GIS.
BIOGRAPHICAL SKETCHES
Jerome E. Dobson (left) is
a senior res earch staff member in ORNL's
Computational Physics and Engineering Division.
He currently serves as chairman of the Interim Research Committee of the University Consortium for Geographic Information Science, scientific editor
of the International GIS Sourcebook, and
a contributing editor and member of the editorial advisory board of GIS World.
He holds a Ph.D. degree in geography from the
University of Tennessee. He
joined the ORNL staff in 1975. He is a former chairman of the Geographic
Information Systems Specialty Group, Association of American Geographers, and
member of the Steering Committee of the
National Committee for Digital
Cartographic Data Standards. He previously served as leader of the Resource
Analysis Group in ORNL's Energy Division,
as visiting associate professor with the Departme
nt of Geography at Arizona State University, as a member of the
editorial board of The Professional Geographer, and as a member of the Steering
Committee of the Applied Geography Conferences. Dobson was co-founder and first
chairman of the Energy Specialt
y Group of the Association of American
Geographers. He proposed the paradigm of automated geography, and he was
instrumental in originating the The National
Center for Geographic Information and Analysis and in est
ablishing the Coastal Change Analysis Program (C-CAP) of the
National Oceanic and Atmospheric Administration. Employing geographic
information systems (GIS) and automated geographic methods, he has proposed new
evidence and theory regarding the mechanisms
responsible for lake acidification
and regarding continental drift and plate tectonics.
Where to?
[ Next article | Search | Mail | Contents | Review Home Page | ORNL Home Page ]