.|  Baltimore Ecosystem Study
Demographic and Socioeconomic Research Projects
Utility of Historical Census Data for Long Term Ecological Studies
  • Christopher Boone, School of Human Evolution & Social Change/School of Sustainability, Arizona State University
  • Geoff Buckley, Dept. of Geography, Ohio University
  • Morgan Grove, Northeastern Research Station, USDA Forest Service
Long term ecological change cannot be understood without an assessment of human activities over time. One of the richest sources of historical data on human activities is the decennial census. We assess and summarize the utility of historical census data, published and unpublished, in conjunction with other data sets to evaluate the relationships between human and ecological dynamics in the Gwynn's Falls watershed and Baltimore City from 1870 to 2010.
Because of the comprehensive and detailed information available in the census over a long period of time, it is critical that researchers know how this resource can be used to better understand the impact of human beings on local ecology and how ecological factors may influence patterns of human decision-making. The purpose of this research is to examine how published and manuscript census information can be used for:
  1. Longitudinal studies at sample locations within the Gwynn's Falls watershed and Baltimore City from 1870 to 2010;
  2. Cross-sectional studies at sample locations within the prescribed study area;
  3. Linking demographic data with ecological data, including geology, soils (fertility, permeability), vegetation (witness trees, surveys, aerial and satellite imagery), water, slope, aspect;
  4. Linking demographic data to other social data, such as health records and consumption information (market clusters);

    Longitudinal Studies
    Since humans are responsible for considerable ecological change, particularly in urban areas, we propose to assess how historical census data can demonstrate changing human patterns over time and space. We need to assess which methods allow us to fairly compare demographics from one period to another. Since variables are not constant over time (e.g. ethnic or race designations can differ considerably), methods must be developed that permit reasonable comparisons between decennial censuses.
    The need to integrate and manage information spanning 200 or more years has presented a number of challenges in developing a geographic base suitable to analyze the relationships between population, land use change, and changes to the physical environment at a variety of geographic scales. Changing census geographies over time will require careful consideration. In the case of Baltimore City, present-day census tracts have been derived largely from historical ward boundaries. Outside city limits, there has been less consistency of boundaries over time. The census geography for the time periods will be digitized allowing a spatial comparison from one decade to the next. For Baltimore City, aggregation and dis-aggregation techniques will be developed to allow statistical comparison of census areal units for longitudinal studies.
    Cross-sectional Studies
    Historical census manuscripts offer an opportunity to reconstruct the human geography of the Baltimore area on a fine scale. Our purpose is to assess which sampling techniques of the manuscript data over time and space would best allow us to map efficiently the human geography of the study area while maintaining a reasonable degree of accuracy. Interpolation of the data using GIS methods will be assessed for accuracy and applicability.
    Linking Census and Ecological Data
    A variety of historical data sets are available that allow the reconstruction of past ecologies. One imperative is to understand how land use and vegetation cover have changed in response to demographic changes. Land use and land cover information can be collected from a variety of sources:
    • forest cover maps, including the State Board of Forestry maps (1906-1916) of the five county area;
    • aerial photographs (1937-present);
    • satellite imagery (1973-present);
    • Sanborn fire insurance atlases (1890, 1901, 1912, 1929, 1953, 1970).

      Linking Census and Other Social Data
      While census data, especially the manuscript records, reveal much about human patterns, to obtain a fuller picture of society in the past we need to link these data to other data sets. We intend to investigate how to link census data to a number of other social indicators that were collected using spatial and temporal methods that differ from those used by the census. Mortality and morbidity data, for example, are often direct links between culture and environment. Poor environmental conditions and in turn public health conditions may be a product of both cultural practices and ecology. Once we assess how to link the data on public health to the census data, we can get a better sense of the cultural factors that interacted with ecological factors to generate a propensity for poor (or good) health. Similarly, once we are able to link the census data to consumption patterns, and other social data sets, it will enrich the census data and give us a more complete understanding of the historical-cultural geography of the Baltimore area.



      [Enlarge This Image]


      Boone, Christopher G., Fragkias, Michail, Buckley, Geoffrey L., Grove, J. Morgan. In press. A long view of polluting industry and environmental justice in Baltimore. Cities (accepted September 2013).
      Boone, Christopher G., Cadenasso, Mary L., Grove, J. Morgan, Schwartz, Kirsten, Buckley, Geoffrey L. 2010. Landscape, vegetation characteristics, and group identity in an urban and suburban watershed: why the 60s matter. Urban Ecosystems. 13, 3: 255-271. doi: 10.1007/s11252-009-0118-7
      Boone, Christopher G., Buckley, Geoffrey L., Grove, J. Morgan, and Chona Sister. 2009. Parks and People: An Environmental Justice Inquiry in Baltimore, Maryland. Annals of the Association of American Geographers 99(4): 1-21. doi: 10.1080/00045600903102949
      Boone, Christopher G. 2008. Improving resolution of census data in metropolitan areas using a dasymetric approach: applications for the Baltimore Ecosystem Study. Cities and the Environment 1, 1: article 3. Online: http://digitalcommons.lmu.edu/cate/vol1/iss1/3
      Boone, Christopher G. 2002. An Assessment and Explanation of Environmental Inequity in Baltimore. Urban Geography 23, 6: 581-595.
      Boone, Christopher G., Buckley, Geoff, Dyer, James. "Utility of Historical Census Data for Long Term Ecological Studies: A Case Study of the Baltimore Metropolitan Region through the Baltimore Ecosystem Study. " Invited Seminar, United States Census Bureau, Suitland, Maryland (July 2000).
      Data and Data Collection
      • A historical GIS (1940-2010) of census data for Baltimore City and sections of surrounding counties at the census tract level is complete. These data area available for download from the National Historic GIS database (http://nhgis.org).
      • A 5 percent sample of manuscript census data for 1910 for Baltimore City is about 60 percent complete. Enumeration districts have been digitized from an enumeration map obtained at the National Archives II. Manuscript census data on microfilm for Baltimore City was purchased from the National Archives. It is housed at the Long Term Social and Ecological Laboratory at Ohio University.
      • Aerial photographs of the Gwynn's Falls watershed. The 1937, 1952, and 1957 air photos have been scanned at the Baltimore County NRCS office. They have been georeferenced and used for a variety of projects in the BES.
      • Forest cover maps of the study area from the early 20th century have been scanned and georeferenced. These were found at the Maryland State Archives.
      • Sanborn Fire Insurance Atlases. The 1890 atlas for Baltimore City has been scanned and georeferenced. Other years (until 1957) will be scanned and georeferenced to support the NSF project examining the longitudinal dynamics of environmental equity in Baltimore (1880-2000).
      • Consumption data. We have used current market cluster classifications developed by Claritas (PRIZM) to reconstruct historical analogues for 1960 and 1970. Boone, Cadenasso, and Grove used these data to examine the role of past demographics on current vegetation patterns (HERCULES) in the Gwynn's Falls Watershed.
      • Satellite imagery. The Baltimore Ecosystem Study has acquired satellite imagery for the Gwynn's Falls region for 1971 through the present.

      Research Topics to be addressed through this Project
      • Urbanization, particularly indices of urbanization;
      • Health, particularly as it relates to environmental equity;
      • Economic structure and land use;
      • Residential differentiation and segregation: race and class; and
      • Migration.
      • Demographic legacies on current vegetation patterns.