"SMALL AREA POPULATIONS FOR THE UNITED STATES"
      (shorter version of the paper initially presented)

                        Robert Leddy
                 Geographic Studies Branch
             Center for International Research
                U.S. Bureau of the Census

                For a presentation at the
           Association of American Geographers
       Annual Meeting San Francisco, 31 March 1994




       Map images can be accessed by clicking on high-
       ligthed or underlined links within this document.

Introduction

The Geographic Studies Branch of the Center for International Research (CIR) of the U.S. Bureau of the Census distributes the population of countries at their last census into small areas, and projects the small area populations annually for 12 years.

Distribution and projection are two-fold, the small areas fitting into an urban-rural dichotomy. All urban agglomerations of at least 25,000 population are covered by one or more circles which encompass at least 95% of the population. Agglomerations consist of contiguous built-up residential areas. Because they do not correspond to administrative units or officially designated urban areas, agglomerations frequently cross state or province boundaries. Circles, which are identified by the latitude and longitude of their centers, are drawn with radii ranging from 0.3 to 2.0 nautical miles. The residential area of a circle is not necessarily of uniform density.

The country's rural residual population is distributed in a gridwork of cells of 20 minutes of latitude by 30 minutes of longitude. The population of each cell is the sum of the populated places -- towns, villages, farms -- that are not included in agglomeration circles. For some countries, CIR has distributed the rural population in smaller "mini" cells which measure 5 minutes of latitude by 7? minutes of longitude.

Projections of the total country population are prepared by CIR using the components-of-population change methodology, which takes into account births, deaths, and migration. The population of each circle and cell is projected to the current year and 11 additional years by a variation of the ratio/proportion method. For example, if an agglomeration contained 10% of a country's total population in 1980 and 15% in 1990, it would be projected to have 20% in 2000 and 25% in 2010.

Large agglomerations typically have numerous groups of circles with different rates of growth. These can consist of a static or declining downtown area, nearby suburbs of low growth, and outer suburbs with relatively high growth. A large agglomeration can have as many as 100 different growth rates.

Whenever possible, the staff uses the most recent census results along with a variety of large and medium-scale maps. Up-to-date city plans are especially useful for reference. For the results of the U.S. census in 1990, the TIGER Line Census Block data were aggregated on ARC/INFO with a series of in-house ARC Macro Language (AML) programs.

U.S. Small Areas -- An Overview

[Map: The 526 U.S. Agglomerations]There were 526 agglomerations with a minimum of 25,000 inhabitants in the United States at the 1990 Census. The TIGER Line Census Block populations of these urban centers were aggregated into 11,811 circles.

The remaining Census Block population became the rural residual population. It was aggregated into 7,434 of the "big" cells measuring 20 minutes by 30 minutes and 111,655 "mini" cells.

The areas of cells are largest at the equator, and then diminish toward either pole. The area of a 20-minute by 30-minute cell in Washington, D.C. (immediately south of 39?N latitude) is almost 600 square miles; that of the 5-minute by 7?-minute mini cell is approximately 35 square miles. There are 6 big cells to every "square" degree, and 16 mini cells to every big cell.

These numbers for U.S. cells are too voluminous to show legibly on one map; instead the agglomerations and cells of Spain are shown. [Map: Spain's Small Areas] In Spain there are 428 big cells and 5,558 mini cells, along with 135 agglomerations with 837 circles. (The map does not show the Canary Islands.)

Sample Agglomerations

[Map: The Agglomerations in the Baltimore-Washington Area]This map shows the circle coverage of the eight agglomerations in the Baltimore-Washington area, which are distinguished here by color.

Each circle must be placed so that it contains a minimum of 5,000 inhabitants. Industrial concentrations, parks, bodies of water, and other minimally-populated areas are avoided as much as possible. At least 80 percent of the area covered by circles with radii of 0.6 nautical mile or greater must be residential build- up. The growing places on the fringes of agglomerations are usually covered by circles of 0.5 nautical miles to provide space for the anticipated areal residential growth during the 12 projection years.

Each agglomeration is named for the most populated city within it. Circle 1 contains the city hall or administrative equivalent.

The Washington, D.C. Agglomeration includes all of the Census Block population of the District of Columbia, and additional census blocks from the surrounding areas in Maryland and Virginia. It has a total of 237 circles, and a 1990 population of 3,093,000. The Baltimore Agglomeration has 160 circles and a population of 1,723,000.

[Map: The New York Agglomeration]The New York Agglomeration is by far the largest in the U.S., both in population and area. There are 704 circles: 363 in New York (state), 329 in New Jersey, and 12 in Connecticut.

This map depicts population densities of the circles in 1990.

Parts of Manhattan have population densities comparable to large Third World cities. Three of the circles, all bordering Central Park, have densities that exceed 100,000 people per square mile (or 40,000 people per square kilometer), and one of these has more than 132,000 people per square mile.

The 704 circles encompassed a total of 14,783,000 people in 1990, and covered an area of 1,291 square miles, which is slightly larger than Rhode Island.

[Map: The Phoenix Agglomeration]The following two examples are agglomerations that grew rapidly between the 1980 and 1990 censuses, both in population and area. Both the 1980 circles and the 1990 circles are overlaid on the Digital Chart of the World so that the impact of urban growth on the land, especially in areas short of water, can be illustrated cartographically.

The map of Phoenix shows that the agglomeration of 1980 already extended into areas that had been desert shrub not too many years previously. That agglomeration had 115 circles covering 241 square miles and 1,282,000 people.

Ten years later, the expansion of the Phoenix population into the desert shrub was so enormous as to encompass 39% more area and 46% more population. The 1990 agglomeration comprised 181 circles covering 337 square miles and 1,872,000 people.

[Map: The Orlando Agglomeration]The second example of rapid growth is Orlando: the 1980 agglomeration had 53 circles covering 71 square miles and 354,000 people; the 1990 agglomeration had 108 circles covering 133 square miles (87% more area) and 720,000 people (103% more population). To the north, Sanford, an agglomeration with 27,000 in 1980, was subsumed into the 1990 Orlando Agglomeration.

An Example of Mini Cell Data--Connecticut

[Map: Connecticut--Mini Cell Rural Residual Populations]The TIGER Line Census Blocks that were not aggregated into circles were aggregated into big cells and mini cells by their centroids. Connecticut has 165 mini cells, including those that cross boundaries and shorelines.

Of Connecticut's total of 3,287,116 people in the 1990 Census, 1,528,000 (47%) were aggregated into cells. Connecticut's urban agglomerations were covered with 178 circles. Only densities of the rural residual populations of the mini cells in 1990 are shown on the map. Even in mini cells with numerous circles, such as those for New Haven, the residual population is significantly large.

[Map: Connecticut--Mini Cell Total Populations]This map shows the density of Connecticut's minicells when the population of the circles is combined with the population of the rural residual to equal the state's total. Not surprisingly, densities are highest by far in the urban areas, and then tend to decrease with greater distances from the urban areas.

Projections by Growth Rates--the Philadelphia Agglomeration

[Map: The Philadelphia Agglomeration]The Philadelphia Agglomeration is a good example of how circles are projected by varying growth rates. This agglomeration, whose 276 circles contained 3,968,000 people in 1990, is typical of large agglomerations in the Northeast and Midwest. Its central city's population has been declining since after World War II. Its aging inner suburbs' populations have been declining since at least the 1970 Census. However, the decline of many years of the populations of Wilmington and Camden was reversed in the 1980s.

Philadelphia's newer outer suburbs have been growing, both in population and in area.

In projecting, circles are numbered sequentially and grouped by similar intercensal growth (or declining) rates. As the map shows, a city the size of Philadelphia has various positive and negative growth rates.

The Value of CIR's Small Area Population Data

Much of the planning and setting of policies for the future is shaped by population and population change, especially as competition for space and the demand for areal resources increase. In a large, growing U.S. urban area, knowledge about small area population change is crucial in order to assess the impact of adding a highway or airport. In a Third World city whose population is tripling in a quarter-century, planning with small area data can help policy makers find the best way to preserve the nearby limited and fragile arable land.

Land use including agricultural, water allocation, infrastructure, industry and commerce, migration, the environment, and disaster relief are among the components that must play. The quality of life of the area's inhabitants is the ultimate focus. Because important planning and policy-making cannot be constrained by artificial administrative boundaries, CIR's geographic distribution of population without regard to administrative units makes it especially useful.

[Map: Small Area Data in CIR Database--Status of Data by Country] Since 1965, CIR has done nearly every country in the world at least once. Although many data have expired, small area data for over 150 countries currently exist in the database. The uniformity of circles and cells in CIR's small area population data is ideal for shaping the scenarios necessary for planning on local and larger scales.