The USGS Land Cover Institute (LCI)
DEFINING THE TEMPORAL AND GEOGRAPHIC LIMITS
Johnson Controls World Services
NASA Ames Research Center, MS242-4
Moffett Field, CA 94035
Establishing the proper temporal scale requires that one capture the time periods that are critical to characterize the growth experienced in an area. Several factors, including population peaks, significant historical events, and data availability, should be considered together before determining the temporal range and increments for a study.
The geographic extent must not only encompass the areas of significant change but also the landscape that is most threatened by future human impacts. Several factors to consider when defining the geographic limits of a study include growth patterns, development trends, local environmental issues and concerns, data needs of local organizations, and the practical constraints of schedules and resource requirements to map the area defined.
The appropriate selection of temporal and geographic limits is essential to insure efficient mapping and relevance of the database for future studies.
Temporal urban mapping studies have demonstrated that historic maps can be integrated with remotely sensed data and related geographic information to successfully map urban land cover changes for large metropolitan areas (Acevedo et. al., 1996; Crawford et. al., 1996; Bell et.al., 1995). From these studies, we have found that selecting appropriate geographic and temporal limits is essential to insure the relevance of the database to address local issues and problems resulting from unplanned growth.
Temporal urban mapping is a U.S. Geological Survey (USGS) program that seeks to develop a database documenting land transformation in urban regions. Data layers may include urban or built-up land, transportation routes, wetlands, agricultural land, forest land, and population density. The spatial patterns and interactions between the physical and socio-economic variables contributing to urbanization are easily illustrated by such a temporal database. USGS resources such as the 100 year archive of topographic maps and the 25 year archive of Landsat data, provide the primary information base for the development of this database. Combined with other information, the database becomes a multiscale and multiresolution record of regional patterns of urban growth.
Currently, two temporal databases have been developed, San Francisco-Sacramento and Baltimore-Washington (Figure 1). Methodology for the extraction and interpretation of information from various data sources were defined during the development of the two databases (Crawford et.al., 1996; Clark et. al., 1995; Bell et. al., 1995). However, methodology for defining the geographic and temporal limits were not developed. This paper describes a methodology for identifying the geographic and temporal limits, in an effort to design an efficient temporal mapping process. The two completed projects, San Francisco-Sacramento and Baltimore-Washington, will be critiqued and two new areas (Chicago and Portland) will be illustrated to emphasize how this methodology for creating the limits, geographic and temporal, can be employed in any metropolitan region.
Urban growth rates show no signs of slowing, especially when viewed at the global scale. Cities throughout the world have gone from small, isolated population centers to large, connected economic, physical, and environmental features of our planet. There has been a continuous demographic shift of the world's populations in the past fifty years. In 1950, 30 percent of the world's population lived in cities. In 1995, 45 percent of the world's population lived in cities. It is anticipated that by the year 2025, 60 percent or more of the world's population will live in cities (Gardner 1996). The rate of urban growth is now recognized as a significant global research problem.
Urban sprawl is a pressing issue for all metropolitan areas. Economic growth drives urban expansion in the form of construction of commercial businesses, houses, roads, and recreational space. The result of urban sprawl includes a loss of natural vegetation and open spaces, as well as a general decline in the spatial extent and connectivity of wetlands, wildlife habitat, and agricultural lands.
Tracking the development of metropolitan areas over time indicates how small towns located in and around agricultural areas are becoming large developed cities and continuous urban corridors. A good example of the conversion of farmland to urban uses is seen throughout California's Central Valley as the population continues to expand and is expected to triple by 2040. Gardener (1996) suggests that this conversion of agricultural lands to urban uses is a very serious issue due to the fact that greater than 50 percent of U.S. agricultural production comes from counties on the edge of cities. Loss of farmland is a long-standing trend, for
example, Los Angeles County with 9 million people today was considered the most productive agricultural county in the United States at the end of WWII. Santa Clara County, in California, once acclaimed for apricots, prunes, and cherries production is better known today for its technological industry, defined by its new name - Silicon Valley (Gardner, 1996).
FIGURES 1 a&b
1a. A time-series representation of the San Francisco-Sacramento temporal database. The urban land is shown as red polygons on a shaded relief background.
1b. A time-series representation of the Baltimore-Washington temporal database. The urban land is shown as red polygons on a shaded relief background.
Decisions about transportation, housing, industrial development, and recreation all affect land use. The actual amount of cropland loss and other land changes due to urbanization will be determined through development choices made by local and national governments.
A temporal urban database provides the baseline information to understand, model, and predict regional patterns of urbanization by documenting human-induced land transformation. Temporal database development relies on modern mapping techniques such as remote sensing and geographic information systems (GIS) to capture information from both historic and modern records. These data accurately references changing geographical features that planners, researchers, and modelers can use as a tool in meaningful inquiry.
Previous temporal mapping projects were initiated without clearly defined geographic extents. For example, the extent of the San Francisco project study area oscillated over time causing the project to take longer to complete than originally scheduled. The original extent of the project area was the footprint of the two Landsat Thematic Mapper (TM) scenes covering San Francisco and Sacramento. This original extent cut major cities, like Modesto, in half and excluded important areas that were extensions of the urban corridor.
Temporal limits for both completed projects were not even considered until each study was well underway. In San Francisco, 1900 began as the earliest date mapped. It was then switched to 1850 to document development near the start of the California gold rush. It was finally switched to 1800 to document the landscape when the California Missions were first being settled and wetland areas were still undisturbed. In the Baltimore-Washington study, temporal limits were not established until the city of Baltimore had been mapped with all data sources available. Source materials were not as abundant for the entire region which required several dates to be eliminated in the final database.
Geographic limits for the Baltimore-Washington project defaulted to the footprint of two 1:250,000-scale quadrangles for that area. These limits luckily encompassed a large portion of the urban corridor between the two major metropolitan areas, but growing commuter cities like York, Pennsylvania were on the edge of the study area and only half the city was mapped. Failure to consider local concerns, like the nutrient and sedimentation loads flowing into the Chesapeake Bay due to development, limited usefulness of the data to researchers in the Chesapeake Bay Ecosystem Program, who needed a temporal database for the entire watershed.
USGS's commitment to expand temporal urban mapping activities to other metropolitan regions increases the need to develop a methodology for efficient selection of geographic and temporal limits. Well established methodology will translate into effective results and insure relevance of the database for future studies.
ESTABLISHING TEMPORAL RANGE AND INCREMENTS
Establishing the proper temporal scope for an urban mapping study requires that one capture the time periods that are critical to characterizing the growth experienced in an area. Several factors, including population peaks, significant historic events, and data availability, should be considered together before determining the temporal range and increments for a study.
Population data from the U.S. Census Bureau provides a simple data source to use for identification of population thresholds and growth trends that portend periods of significant human impacts and land degradation. Graphical analysis of population provides an understanding of the major development periods for a specific region. Population statistics for each county in the U.S., from 1900 to 1990, are available from Census Bureau internet locations.
Historic events are another indicator that can be used to provide insight about the development of a particular area. A record of significant historic events should be developed for each metropolitan region to be mapped. A time-line of these events will provide insight into the dates or time periods which significantly impacted or changed the future of a region. Major global events such as the industrial revolution, World Wars I and II, and the heated economy of the 1980's, are easily recognized for their impact on urban growth. Other more localized events such as changes in agricultural practices, changes in labor practices, changes in industrial production, changes in the economy of a region, should all be used as indicators of urban land use change that may need to be documented in a temporal database.
The availability of topographic maps and to a lesser degree satellite data, should be established at the onset of a temporal urban mapping study to determine if resources exist and if they are accessible to those compiling the database. Two main archives of topographic maps have been identified: USGS Library and Cartographic Information Center located in Reston, Virginia, and the Library of Congress, Map Library located in Washington, D.C. These two archives contain maps for the entire U.S. and are accessible to a temporal urban mapping study. Experience in the San Francisco-Sacramento and Baltimore-Washington studies indicate that most metropolitan areas will have local archives containing regional maps.
USGS is a reliable source of topographic maps because they have completed numerous topographic mapping efforts throughout the past century. Between 1880 and 1920 USGS was focused on completing the first in a series of topographic maps for most urban areas in the U.S. The majority of these early maps were produced at 1:62,500-scale and covered a 15-minute quadrangle. Limited funds and outdated methodology brought USGS mapping efforts almost to a stand still between 1930 and 1950. The Army Map Service (later renamed the Defense Mapping Agency) produced the majority of maps during this time. By the 1950's the USGS resumed topographic mapping and between 1950 and 1970 remapped most urban areas at 1:24,000-scale with the 7.5-minute topographic series of maps. Since then one or two map revisions have been completed for most major metropolitan regions (Thompson 1979). A full collection of USGS maps can be found at the USGS Library and Cartographic Information Center.
The Library of Congress, Map Library also houses all of the USGS maps produced in addition to many other map series produced throughout the last two centuries. An important record of urban growth and development in the U.S are the Sanborn Map Company fire insurance maps and atlases for 12,000 American cities and towns, between the years 1876 and 1961. Also available at the Library of Congress are several different collections of panoramic maps. The Ruger Collection, one example of panoramic maps, dating from the 1860's and 1870's, produced a "bird's-eye" view of settlements in twenty states from New Hampshire to Alabama.
ESTABLISHING GEOGRAPHIC EXTENT
The geographic extent of a temporal urban study must not only encompass the areas of significant change but also the landscape that is most threatened by future human impacts. The area considered threatened may include future urbanization corridors, agricultural land, sensitive wildlife or wetlands habitat, watershed boundaries, drinking water supplies, estuaries, forestland, and urban recreational areas. There are several factors to consider when defining the geographic limits of a study including historical growth trends, local environmental issues and concerns, and the practical constraints of schedules and resource requirements to map the area defined.
Urban Growth Trends
Defining the location and extent of a temporal urban mapping study requires a preliminary analysis of the spatial patterns and rates of recent urban development in order to anticipate future growth patterns. Before a temporal mapping database is initiated a couple of readily available datasets can be used to analyze growth changes over the past 20 years (Figure 2). The National Oceanic and Atmospheric Administration (NOAA) has produced a `city lights' dataset from data captured by the Defense Meteorological Satellite Program Operation Linescan System in 1995 (Elvidge et. al., 1996). The USGS also has a digital land use and land cover (LULC) database of maps and associated overlays at 1:250,000-scale (Loelkes, 1977; Fegeas et al.,1983). Processed by the Geographic Information Retrieval and Analysis System (GIRAS), these data were derived primarily from 1970's photography. Both datasets provide sufficient information to extrapolate a 20 year growth trend. From this information, potential study boundaries can be established which reflect anticipated future growth directions that encompass most areas susceptible to urban sprawl.
In every region throughout the U.S., there are a host of environmental issues defined by the physical environment. Understanding environmental issues helps to focus importance on specific regions or land cover types. These issues are often caused or exasperated by urban development. Since data from a temporal urban mapping study can provide a valuable resource to a local environmental research project, local research study limits should be incorporated into the geographic limits for the temporal urban mapping study. In the San Francisco-Sacramento study, the major environmental research projects are water quality in the San Francisco Bay, impacts of recreational development in the Sierra Nevada mountains, and the protection of agricultural land in the Central Valley. The areas critical to these research projects were incorporated into the final geographic limits of the study area.
In establishing a geographic extent one must also consider the practical constraints of schedules and the resources required to complete the mapping of an area. It is important to recognize that if the study area is too large it will take too long to complete and gather together all of the resources needed to fully develop the database will be difficult. The number of individuals it will take to complete the project should also be taken into consideration.
Appropriate temporal and geographic limits for a temporal urban mapping study can be derived from a variety of factors. Time-lines of significant events and graphs of population change over time draw attention to specific time periods in the history of a metropolitan region. Time-lines also exhibit when a city became a sizable metropolitan area which helps to determine the earliest point in time to map. Once the important time periods have been identified, identifying resource availability confirms those time periods that can actually be mapped. The procedures used to define the geographic limits involve understanding the current spatial growth patterns, development trends, local environmental issues, public concerns, and data needs of local users. These criteria help focus the geographic extent of the study area which must eventually involve consideration of the practical constraints of mapping a large area over time.
Defining Chicago's temporal and geographic limits.
The Chicago metropolitan region has grown dramatically since Chicago's incorporation in 1830. With a population already at 400,000 in 1871, the area continues to increase, and was over 8.0 million in 1990. Graphing population figures for the region showed significant upward population swings in 1920 and 1950 (Figure 4).
The compilation of a land use history for the Chicago region documents some of the significant historical events which have contributed to development of the area. Major historical changes in modes of transportation, strongly support the need to map development during the transition from canals to railroads (1850's), at the turn of the century (electric railroads), and when the automobile began to dominate (1940's) (Figure 5).
A preliminary inventory of the map and satellite data archives concluded that sufficient source materials would be available and would not limit the ability to map the proposed time line intervals (1870, 1900, 1920, 1950, 1970, and 1990).
1830 The city of Chicago was incorporated as the eastern terminus of the Illinois & Michigan Canal.1871 1892 The elevated rail was built and the population in Chicago was approximately one half million people. 1900 Rail cars became electric and rail lines opened up areas in all directions for business and residential development. Today Chicago's rail system carries a half million passengers per day on more than 247 miles of track. Much of this infrastructure was built 100 years ago. 1920 There was an increased use of the automobile during this decade. Highways became ubiquitous diminishing the importance of the canal towns west of Chicago. 1950 Population expansion post WWII. 1970 Plan 21, revitalization of the central Chicago area known as the Loop
Current development patterns for the greater Chicago-Milwaukee area are shown in figure 2. The nested images present a view of the changes in the extent of urban or built-up land for 1975 and 1995. Centered on Lake Michigan, the synoptic view shows the large urban cores around Chicago and Milwaukee. Many smaller cities (Grand Rapids, Lansing, South Bend, Fort Wayne, Peoria, Rockford, Madison) are also clearly visible. From this image growth trends and urban corridors become evident around the Lakes southwestern shore and in the developing corridor between Chicago and Milwaukee.
A review of local environmental issues and user organizations is ongoing. One organization, Chicago Wilderness, a collaboration of 34 groups and agencies, focus their attention on the protection of open land within Chicago's urban corridor. Research projects within the six counties in Illinois which make up the bulk of Chicago Wilderness' influence (Cook, DuPage, Kane, Lake, McHenry, and Will) study rare habitat types, such as tall grass prairie and oak savannas, pollution of waterways and abandoned lands, and biodiversity throughout the region.
Another regional environmental concern is lake shore erosion and impending floods. These concerns stem from man influenced in-fill of the lake throughout the last century. The impact of land-filling the lake shore to increase the amount of developed land was common from Milwaukee, Wisconsin around the lake to Gary, Indiana. Taking into account Chicago Wilderness research projects and growing lake shore concerns, selection of a geographic extent which encompasses the six counties and the urban corridor which stretches around the west and south end of the lake will create a database useful to many different clients.
Final geographic limits for Chicago should extend from Milwaukee in the north to Joliet in the southwest and around the lake to include Gary, Fort Wayne, and Grand Rapids, Michigan. A look at this region in relation to the resources needed to map the area provides a reality check to insure a timely mapping effort. The area described encompasses about four 1:250,000-scale quadrangle maps which experience has proven to be a fairly manageable size for a study area.
Defining Portland's temporal and geographic limits.
The Portland-Vancouver metropolitan region, although much smaller than Chicago-Milwaukee, has grown just as dramatically since it was first settled. With a population of less than 200, 000 in 1900, the area continues to increase, and was over 1.5 million in 1990. Plots of population (Clackamas, Multnomah, Washington, Clark, and Cowlitz) for the region showed significant upward swings in 1910, 1940, 1960 and 1980.
The ongoing compilation of the land use history for the region supports the need to document development at the turn of the century and just before the development boom of the 1980's. Local users confirm the need to document pre/post WWII development patterns, as well as how growth patterns developed with respect to transportation systems (Metro, 1996).
Reviewing the regional growth trends over the last 20 years for the Portland-Vancouver region (Figure 3), a rapidly developing urban corridor is seen extending between Longview, Washington in the north, and Salem, Oregon in the South. This area has a much smaller geographic extent than the Chicago-Milwaukee metropolitan region, with the study area easily encompassed by two USGS 1:250,000-scale quadrangles.
Air and water quality, and regional planning top the list of issues and local concerns important to this area as population increases the demand on limited resources. Local environmental organizations include Metro, a directly elected regional government, established in 1979 to manage urban development by creating regional land use goals and objectives. These growth management efforts have been encouraging compact growth near existing transportation routes, reducing land consumption and limiting impacts to air quality. The geographic extent for the Metro Program is the extent of Multnomah, Washington and Clackamas counties. A viable temporal urban mapping study must encompass this area in order to provide data applicable to these issues. Another example of local environmental issues is the effort to curb degradation of water resources in the Lower Columbia River Estuary. The Lower Columbia River Estuary Program includes the lower 146 miles of the river, and receives funding from the States of Oregon and Washington and the EPA. The Columbia River is a major shipping route which supports a diverse and productive economy. The Columbia River also drains a large and complex watershed. A temporal urban mapping study in this region could not effectively cover this entire watershed, but data generated for the metropolitan regions bordering the Estuary Program boundary would contribute significantly to the body of information needed to protect this river system.
After evaluating the criteria proposed for selecting temporal and geographic limits, a study area extending from the Pacific Ocean east to The Dalles and north of Longview, Washington, to south of Salem, Oregon, was deemed most appropriate for studying the Portland-Vancouver metropolitan region. This area corresponds to land area equivalent of two USGS 1:250,000-scale maps, an area well within the practical constraints of a temporal mapping effort.
The process of defining the temporal periods and the geographic extent for mapping urban land use changes involves several factors that must be considered individually and together to determine these limits. The procedures defined involve evaluating the availability of source materials. Gathering the information necessary to understand population trends, historic events, growth patterns, and development trends. Identifying the local environmental issues, public concerns, and data needs of local organizations. And as a final check it is necessary to consider the practical constraints involved in mapping a large area.
Acevedo, W., Foresman, T.W., and Buchanan, J.T., 1996, Origins and philosophy of building a temporal database to examine human transformation processes: Proc. ASPRS/ACSM Annual Convention and Exhibition, Baltimore, MD, v. 1, p. 148-161.
Bell, C., Acevedo, W., and Buchanan, J.T., 1995, Dynamic mapping of urban regions: growth of the San Francisco Sacramento region: Proc. Urban and Regional Information Systems Association, San Antonio, TX, v. 1, p. 723-734.
Clark, S.C., Starr, J., Foresman, T.W., Hinzman, D., Wiggins, H., Acevedo, W., and Solomon, C., 1996, Development of the temporal transportation database for the analysis of urban development in the Baltimore-Washington region: Proc. ASPRS/ACSM Annual Convention and Exhibition, Baltimore, MD, v. 3, p. 77-88.
Crawford, J., Acevedo, W., Foresman, T.W., Buchanan, J.T., and Prince, W., 1996, Developing a temporal database of urban development for the Baltimore/Washington region: Proc. ASPRS/ACSM Annual Convention and Exhibition, Baltimore, MD, v. 3, p. 101-110.
Elvidge, C.D., Baugh, K.E., Kihn, E.A., Kroehl, H.W., and Davis, E.R., 1996, Mapping City Lights with Nighttime Data from the DMSP Operational Linescan System, Photogrammetric Engineering and Remote Sensing, (in press).
Fegeas, R.G., Claire, R.W., Guptill, S.C., Anderson, K.E., and Hallam, C.A., Land Use and Land Cover Digital Data, 1983, U.S. Geological Survey Circular 895-E, 21 pp.
Gardner, G., 1996, Shrinking Fields: Cropland Loss in a World of Eight Billion, World Watch Paper #131, World Watch Institute, Washington D.C., 56 pp.
Loelkes, G.L., 1977, Specifications for Land Use and Land Cover and Associated Maps, U.S. Geological Survey Open File Report 77-555, 103 pp.
Metro, 1996, Urban Growth Report, Growth Management Services Department, Portland, Oregon, 36 pp.
Thompson, M.M., 1979, Maps for America: Cartographic Products of the U.S. Geological Survey and Others, U.S. Geological Survey, Reston, VA.