«Urban Problems and sPatial methods VolUme 17, nUmber 1 • 2015 U.S. Department of Housing and Urban Development | Office of Policy Development and ...»
Kim, Sunwoong. 2000. “Race and Home Price Appreciation in Urban Neighborhoods: Evidence from Milwaukee, Wisconsin,” The Review of Black Political Economy 28 (2): 9–28.
Lieberson, Stanley. 1981. “An Asymmetrical Approach to Segregation.” In Ethnic Segregation in
Cities, edited by Ceri Peach, Vaughan Robinson, and Susan Smith. London, United Kingdom:
Massey, Douglas S., and Nancy A. Denton. 1988. “The Dimensions of Residential Segregation,” Social Forces 67 (2): 281–315.
Massey, Douglas S., and Mary J. Fischer. 2000. “How Segregation Concentrates Poverty,” Ethnic and Racial Studies 23 (4): 670–691.
Massey, Douglas S., Michael J. White, and Voon-Chin Phua. 1996. “The Dimensions of Segregation Revisited,” Sociological Methods Research 25 (2): 172–206.
Morrill, Richard L. 1991. “On the Measure of Geographic Segregation,” Geography Research Forum 11: 25–36.
R Core Team. 2014. R: A Language and Environment for Statistical Computing. Version 3.1.0. Vienna, Austria: R Foundation for Statistical Computing.
Reardon, Sean F., and Glenn Firebaugh. 2002. “Measures of Multigroup Segregation,” Sociological Methodology 32 (1): 33–67.
Reardon, Sean F., and David O’Sullivan. 2004. “Measures of Spatial Segregation,” Sociological Methodology 34 (1): 121–162.
Reibel, Michael, and Moira Regelson. 2007. “Quantifying Neighborhood Racial and Ethnic Transition Clusters in Multiethnic Cities,” Urban Geography 28 (4): 361–376.
Reiner, Thomas A. 1972. “Racial Segregation: A Comment,” Journal of Regional Science 12 (1): 137–148.
Shannon, Claude E. 1948a. “A Mathematical Theory of Communication,” The Bell System Technical Journal 27 (3): 379–423.
———. 1948b. “A Mathematical Theory of Communication,” The Bell System Technical Journal 27 (4): 623–656.
Taeuber, Karl E. 1968. “The Problem of Residential Segregation,” Proceedings of the Academy of Political Science 29 (1): 101–110.
Theil, Henri. 1972. Statistical Decomposition Analysis: With Applications in the Social and Administrative Sciences. New York: American Elsevier Publishing Company.
U.S. Census Bureau. 2014. “Geographic Areas—Definitions.” American Community Survey Office.
Vinikoor, Lisa C., Jay S. Kaufman, Richard F. MacLehose, and Barbara A. Laraia. 2008. “Effects of Racial Density and Income Incongruity on Pregnancy Outcomes in Less Segregated Communities,” Social Science & Medicine 66 (2): 255–259.
White, Michael J. 1986. “Segregation and Diversity Measures in Population Distribution,” Population Index 52 (2): 198–221.
———. 1983. “The Measurement of Spatial Segregation,” American Journal of Sociology 88 (5):
Williams, David R. 1999. “Race, Socioeconomic Status, and Health: The Added Effects of Racism and Discrimination,” Annals of the New York Academy of Sciences 896: 173–188.
Williams, David R., and Chiquita Collins. 2001. “Racial Residential Segregation: A Fundamental Cause of Racial Disparities in Health,” Public Health Reports 116 (5): 404–416.
Winship, Christopher. 1978. “The Desirability of Using the Index of Dissimilarity or Any Adjustment of It for Measuring Segregation: Reply to Falk, Cortese, and Cohen,” Social Forces 57 (2):
Wong, David W.S. 2008. “A Local Multi-Dimensional Approach To Evaluate Changes in Segregation,” Urban Geography 29 (5): 455–472.
———. 2004. “Comparing Traditional and Spatial Segregation Measures: A Spatial Scale Perspective,” Urban Geography 25 (1): 66–82.
———. 2002. “Modeling Local Segregation: A Spatial Interaction Approach,” Geographical & Environmental Modelling 6 (1): 81–97.
———. 1998. “Measuring Multiethnic Spatial Segregation,” Urban Geography 19 (1): 77–87.
———. 1996. “Enhancing Segregation Studies Using GIS,” Computers, Environment and Urban Systems 20 (2): 99–109.
———. 1993. “Spatial Indices of Segregation,” Urban Studies 30 (3): 559–572.
Wright, Kevin. 2012. Corrgram: Plot a Correlogram. R package version 1.4.
Zelder, Raymond E. 1972. “Racial Segregation: A Reply,” Journal of Regional Science 12 (1): 149–153.
Abstract Many types of urban policy analyses, particularly those relating to exposure to hazards or accessibility to resources, rely on accurate and precise spatial population data, although such data are not always available. Dasymetric mapping is a technique for disaggregating population data from one set of source spatial units to a finer resolution set of target spatial units through the use of an ancillary dataset, typically land use, zoning, or similar nominal datasets related to population distribution. Dasymetric mapping operates by employing weights that capture both the relative areas of the target spatial units and the relative population densities of the different nominal ancillary classes, and it is typically implemented in Geographic Information System, or GIS, software. An example application demonstrates the efficacy of the dasymetric approach by comparing census tract-level and dasymetric data in an assessment of the population living in proximity to hazardous air pollutant releases in Philadelphia, Pennsylvania, using block-level data as a validation dataset.
Introduction Many types of urban policy analyses rely on accurate and precise spatial population data. Of particular note are analyses of exposure and accessibility, where one must assess the population in proximity to, or overlapping with, some geographic feature. Examples of such analyses include the estimation of population exposed to natural and technological hazards, such as flooding or air pollution.
Other relevant research applications concern access to amenities and resources, such as recreation facilities, health centers, nutritious food, or employment opportunities.
Although the U.S. Census Bureau provides high-resolution demographic data for the United States, certain variables may be available only over coarser spatial units, such as census tracts.
Other population-related datasets, such as disease incidence data, may be limited to distribution at a coarse spatial resolution for purposes of privacy protection. In many developing nations, population data at a fine resolution are not available at all, because many countries do not have the resources to invest in census infrastructure. In addition, in all these cases, population data are likely to be available aggregated to spatial units that are derived by convenience of enumeration or are a reflection of administrative or political jurisdiction boundaries and, consequently, are unlikely to capture the nature of the actual population distribution. Thus, the development of small-area estimates for urban population data remains a challenge in both developed and developing nations.
Dasymetric mapping is a technique for estimating population in small areas in situations where one has access to population data aggregated only at a relatively coarser scale (Mennis, 2009). It uses ancillary data, an additional dataset related to the distribution of population but distinct from it, to disaggregate population data from one set of spatial units to another set of smaller spatial units. The formal principles of dasymetric mapping were initially developed for a Russian mapping project in the early 20th century (cf. Petrov, 2012) and were introduced to English-speaking audiences in a series of articles appearing in the 1920s and 1930s, most notably in an article by Wright (1936). The dasymetric mapping technique, however, was little known outside cartographic circles until the widespread availability of Geographic Information System (GIS) software and digital data products that could serve as ancillary data, such as those derived from remotely sensed imagery, spurred the growth of dasymetric mapping algorithms and applications beginning in the 1990s through the present.
Dasymetric mapping more recently has been employed for a wide variety of applications that benefit from high spatial resolution population data, including environmental justice (Mennis, 2002), public health (Maantay, Maroko, and Porter-Morgan, 2008), crime (Poulsen and Kennedy, 2004), and historical population estimation (Gregory and Ell, 2005). It has also been used to create national-level, high-resolution population datasets (Bhaduri et al., 2007).
The purpose of the present article is to describe dasymetric mapping, its theoretical basis, and its implementation using GIS software. As an illustration of dasymetric mapping and its application to urban analysis, an example is presented for Philadelphia, Pennsylvania, where tract-level population data are disaggregated to sub-tract-level spatial units. These data are then used for an analysis of population residing in close proximity to facilities releasing hazardous pollutants to the atmosphere. The tract and the dasymetric data are then compared with an analogous analysis using census block-level data for accuracy assessment.
The Dasymetric Mapping Technique Dasymetric mapping can be considered a form of areal interpolation, the transformation of data from one set of spatial units to another set of spatial units; for example, the assignment of population originally encoded in U.S. counties to a set of watershed boundaries. The original set of spatial units is referred to as source zones and the set of destination spatial units is referred to as target zones. The simplest approach to areal interpolation is areal weighting, which assumes a homogeneous distribution of the data within the source zones. Thus, data are apportioned to the target zones based on the proportional area that each source zone contributes to each target zone.
A particular case of areal interpolation occurs when the target zones are formed by the geometric intersection of the source zones with another—ancillary—polygon data layer, so that the target zones spatially nest perfectly within the source zones, and each source zone can be disaggregated into one or more target zones. Areal weighting in this case implies that given a target zone f nested within a source zone g, such that f ∈g, then the population of the source zone can be distributed to where A is area. The target zone population can then be estimated as ŷf = yg ARf, where ŷf is the its constituent target zones based on the area ratio (AR) of each target zone, where ARf = Af ⁄Ag, and estimated count of the target zone and yg is the population of the host source zone (Goodchild and Lam, 1980).
Dasymetric mapping can be viewed as an extension of areal weighting in which the ancillary dataset overlaid with the source layer is typically an area-class map, which exhaustively tessellates a region into nominal classes that are related to the distribution of the variable being mapped.
Thus, dasymetric mapping incorporates not only the relative proportion of the contributing area of each target zone but also its ancillary class to redistribute data from the source zone to its constituent target zones. As such, dasymetric mapping employs not only the area ratio but also a density class c associated with target zone f, the density ratio (DR) can be defined as DRc = (D̂ c )⁄ ( ∑D̂ c ), ratio among the ancillary classes to make target-level estimates. If we formally consider an ancillary where D̂ c is the estimated density of the ancillary class c. The total fraction (TF) integrates the area ratio and density fraction into a single term, where
ŷf = yg (TFfc).
The target zone population can then be estimated as (2) Note that the sum of the population of each source zone is maintained in the dasymetric output The value of D̂ c can be set by the analyst through his or her own expert knowledge (Eicher and (Tobler, 1979), because the area ratio and density ratio both sum to 1 for each source zone.
Brewer, 2001) or it can be estimated by sampling the variable values of source layer zones that are set directly by the analyst without setting the values of D̂ c in cases in which one is knowledgeable spatially coincident with different ancillary data classes (Mennis, 2003). Values of DRc can also be about only the relative densities among the classes.
By far the most common ancillary dataset used in dasymetric mapping of population is land use or land cover data, often derived from classified remotely sensed imagery. The most basic dasymetric mapping implementation involves the use of such ancillary data to simply distinguish between inhabited and uninhabited land area; for instance, by distinguishing between developed regions and those occupied by water or barren land. In this case, all population in a source zone bisected by uninhabited and inhabited land would be allocated to the land classified as inhabited, leaving the remaining portion of the source zone with zero population. Exhibit 1 illustrates this principle of dasymetric mapping using a schematic diagram. A set of source zones with observed population densities is shown on the left, with an ancillary land cover data layer used in the dasymetric
km2 = square kilometers.
mapping shown in the middle. The resulting dasymetric map is shown on the right, where all the population of a source zone with both inhabited and uninhabited regions is apportioned to the inhabited region. Thus, the population density of the inhabited regions of a source zone increases and is conversely held to zero in the uninhabited regions.